Macy’s was sued in 2017 over bedsheets it sold as 900 thread count. The lawsuit alleged that the true count was only 249. Macy's ultimately settled the case for $10.5M. The court estimated the class at “around 2.5 million individuals,” everyone who bought one supplier’s cotton-poly sheets from Macy’s over nearly 10 years.

If the lawsuit was right, then Macy's number was simply false. But a true number can mislead just as well. In almost every industry I've looked at, manufacturers print numbers that sound precise but tell you very little about the product on their own.

Thread count, down fill power, horsepower and hardness ratings are a few examples of common measurements that manufacturers love to quote. Manufacturers print them not to inform, but to drive sales.

For today’s article I picked five numbers I know well. For each of them, I'll explain what the number actually measures, what it leaves out, and what to check instead.

Many of you will know the numbers in your own areas of expertise better than I do. Please do share that knowledge in the comments, or in the community, if you’re a member.

Thread count

Thread count is the number of threads in a square inch of fabric, counted in both directions. What is often left out is ply. Ply counts how many individual yarns are twisted together to create one piece of thread. By counting each ply as a separate thread, manufacturers can arbitrarily double or triple their thread count number.

In 2005, the FTC issued guidance to the National Textile Association on how sheets should be labeled. James Kohm, the FTC's associate director for enforcement, wrote that sellers should "state count followed by ply, as in: 300 thread count, two-ply yarn." Calling that same sheet a 600 "would likely mislead consumers about the quality of the product being purchased."

Bear in mind, this is just staff guidance, so it isn’t legally binding.

Even an honest count that follows the FTC guidance doesn’t tell you much. After washing sheets repeatedly in its tests, Consumer Reports "could not find any correlation between manufacturer claims of materials or thread count and sheet performance." One of its top-rated sets is a 350. A 550-count Macy's set sat near the bottom after it shrank too much to fit a 10-inch mattress.

What to check instead: look for "single-ply" on the package. Then read the material content, because the fiber itself tells you more than any thread count number about how the sheet will behave. In Consumer Reports' tests, cotton sheets "shrank up to 6 percent." The polyester sheet "hardly shrank at all," but it "doesn't absorb water," so some people find it sweaty. Bamboo viscose shrank by more than 15 percent.

Down fill power

Fill power measures how much volume down takes up. REI defines it as "how many cubic inches of loft one ounce of that down fill produces." The testing lab IDFL calls 700 and above "extremely high quality." The fill power on a label is allowed to be off by 5% in either direction.

What fill power leaves out is fill weight: the amount of down actually in a jacket or sleeping bag. The point of down is warmth, and to compare the warmth of two jackets you need both numbers. Fill power alone tells you how light and packable the down is. It can't tell you how warm the jacket is.

REI sells a lot of down jackets, and its own guide admits "Very few jacket makers report fill weights." Patagonia's Down Sweater page, for example, lists "800-fill-power" down and a total weight of 13 ounces, with no figure for the down itself.

Here's how the two numbers interact, using a couple hypothetical jackets. Jacket A holds 3 ounces of 800-fill down, giving it 3 × 800 = 2,400 cubic inches of loft. Jacket B holds 4 ounces of 650-fill down, for 4 × 650 = 2,600 cubic inches. Jacket B is the warmer of the two, even with the smaller fill power number.

Higher fill power buys you less weight and bulk for the same warmth. In REI's words, "In theory, 50 g of 900-fill-power down would be the same warmth as 100 g of 450-fill-power down; it would also compress to about half the space."

As a backpacker and general outdoor tomfoolerer, I think high fill power is worth paying for. My frustration is how rarely brands quote both numbers, which makes clean comparisons between options almost impossible.

The FTC's guidance covers what's in the fill. A product labeled "down" should be "more than 70%" down, and a manufacturer "may not 'shoot for the tolerance.'" It says nothing about fill weight.

What to check instead: ask for the fill weight in ounces or grams. If the brand won't give it, be skeptical, since the fill power alone doesn’t tell you much.

Leather grades

Leather grades describe which layer of the hide was used and how its surface was treated. Each grade suits different things. Not everything needs to be full grain, and full grain does not mean high quality.

Full grain keeps the outer layer intact. It's the most durable grade and develops a patina, so it's the one for boots, belts and bags you plan to keep for years.

Top grain is the outer layer sanded and refinished for an even look. It gives up some durability and patina for suppleness, and works well for jackets, handbags and furniture.

Corrected grain is treated more heavily to hide flaws. It's cheaper and holds up to wear, so it shows up in budget furniture, car seats and shoes.

Split leather and suede come from the lower layers of the hide. They're softer and weaker than grain leather, but split cowhide is common in work gloves because it resists abrasion and costs less.

Bonded leather is shredded leather fiber mixed with a binder. It's the cheapest and least durable grade, and prone to cracking. Avoid.

None of these categories have a legal definition. In 2008, the leather industry's trade group asked the FTC to define the grade words, starting with "top grain or full grain leather." The FTC declined:

The FTC’s Leather Guides have not changed since 1996. They say the bare word “leather” should only describe a product “composed in all substantial parts of leather,” and that bonded material must state its share of leather fiber.

So the “genuine leather” stamp you see everywhere tells you only that the material is leather. Every grade outside of bonded leather qualifies. It tells you nothing of the quality of the product.

What to check instead: the grade on the tag is only a starting point. There's good full grain and bad full grain. Telling them apart takes a trained eye. If you don't have one, borrow an expert's. Bedo's Leatherworks, a third-generation cobbler, and Rose Anvil, who cuts boots in half to show how they're built, are two of the best places to start.

Peak horsepower on shop vacuums

Wet/dry shop vacuums are sold on peak horsepower. Often the number is printed right there in the product name. Ridgid sells a "6 Gal. 3.5 Peak HP" model at Home Depot. Shop-Vac sells a "5 Gallon 6.0 Peak HP" model. Its product page explains what the number means:

That disclosure came out of a 2016 settlement. The lawsuit claimed an independent lab found the vacuums generated "only a small fraction of the advertised 'Peak Horsepower.'"

You can check the number yourself directly from the spec sheet. The 6.0 peak horsepower Shop-Vac is rated at 11.8 amps on a 120-volt circuit, and 120 × 11.8 = 1,416 watts. One horsepower equals 746 watts, so the most this motor can draw from the wall is about 1.9 horsepower. It delivers less than it draws. Producing 6.0 horsepower would take 4,476 watts, or about 37 amps.

The numbers that matter are airflow, measured in cubic feet per minute (CFM), and sealed suction, measured in inches of water lift. Air watts combine the two. According to Pro Tool Reviews, airflow keeps debris moving, and water lift is what gets it moving in the first place.

The Ridgid, rated at 3.5 peak horsepower, lists 64 CFM, 67 inches of water lift and 186 air watts. The Shop-Vac, rated at 6.0, lists 90 CFM, 50 inches of lift and 180 "peak air watts." The Shop-Vac moves more air, the Ridgid pulls harder, and their air power is about the same. Peak horsepower alone tells you none of that.

What to check instead: look for air watts, CFM and water lift on the spec sheet. The amp rating tells you how much power the motor really draws.

Knife steel and hardness

A knife's performance depends on three things: the steel, its hardness and the angle of its edge.

Hardness is measured on the Rockwell C scale, written as HRC. Wüsthof prints 58 for its Classic line. A higher number means the steel resists wear for longer, but that comes at a cost. Larrin Thomas, who runs Knife Steel Nerds, explains it: "Higher hardness/strength reduces the toughness of steel. Toughness is the resistance to fracture or chipping." 

The type of steel used lets you look up its mix of toughness, edge retention and rust resistance. Those properties trade off against each other, so the right steel depends on the job. A hard-use outdoor knife needs toughness. A kitchen knife needs a thin edge that won't roll or chip.

The thing that matters most, though, has nothing to do with materials or manufacturing. A knife’s edge angle is set by the owner every time they sharpen their blade. How well a knife’s edge has been maintained is more important than its steel type or hardness.

Thomas puts it plainly: "sharpening in terms of grit finish and edge angle have a greater effect on performance than steel choice." He tested 48 steels and built a model from the results. In it, four extra points of hardness added about 63 mm of cardstock cut. Sharpening the edge 5 degrees thinner per side added about 89.

What to check instead: a named alloy, and hardness as a tradeoff. Then put your effort into keeping the edge sharp.

That's why I'm comfortable using Victorinox knives exclusively, even though Victorinox publishes no steel grade or hardness. America's Test Kitchen, which tests sharpness with an industrial machine, calls the Fibrox Pro 8" chef's knife its "top choice for decades."

The case for standardized testing

Three of the five numbers discussed here have no outside standard behind them. Sellers choose how to count threads, and the FTC's guidance on ply isn't binding. Leather grades have no legal definition. Peak horsepower is a figure from the maker's own lab tests. The numbers that hold up better are measured with a standardized test that someone other than the seller wrote. Ridgid, for instance, says its air watts are "tested to ASTM standard."

One great example of standardized testing is the grade stamp on a piece of lumber. A stamped 2x4 shows the grading agency, the mill number, the grade, the species and the moisture level when it was graded.

The grading rules sit under PS 20, a voluntary federal product standard. The American Lumber Standard Committee's Board of Review "accredits agencies to grade and inspect under those rules and monitors the agencies' performance." The agencies then send inspectors to the mills.

A standardized test does two things. It makes products comparable, because every brand is measured the same way. It also holds companies’ marketing claims to account. The Macy's lawsuit could set 900 against 249 because a thread can be counted. There is no equivalent number to dispute when a wallet says "genuine leather."

In Between Craft and Quality, I argued that products have grown more complex while each of us has become less able to detect quality at the point of sale. I wrote then that "a quality nobody can detect is a quality nobody pays for, and a quality nobody pays for stops being made." 

A shared, independent test is one way to make quality detectable again. When buyers can compare real numbers, the better product is more likely to win, and a company that cuts corners shows up in the results.

Perhaps it's time to bring standards like these to more consumer industries. Exactly how is an open question. My instinct is that independent private bodies, like the lumber committee, would do it best. The government could do it too.

What numbers do you trust?

I have shared only a few examples of misleading numbers from the categories I am most familiar with. I am certain that there are many more. If you work with fabric, wood, steel, engines or electronics, you'll know which numbers matter in your trade, and which ones a seller leaves off.

Tell me yours in the comments. If you're a member, post it in the community, where readers compare notes on exactly this. I’ll follow up with a more complete list once we’ve built it.

Thanks as always for reading.

In March 2022, a federal judge in Pennsylvania summarized testimony from a former salesman at Horizon Services, a prominent East Coast heating and cooling contractor. In one case, a technician reported a failed compressor. On that diagnosis, the salesman sold the customer a brand new system for $22,000. The actual problem all along was a “$15 breaker switch”.

The salesman, Philip DelPalazzo, was acting as a whistleblower in this case. His son Vincent, a maintenance technician at the same company, had told him that other technicians "were told that if it is a ten-year old unit and you don't set a lead you are going to be punished." DelPalazzo himself described a 2018 sales call during which the customer said a Horizon technician “sprayed oil by her furnace to make it look like it needed to be replaced when it did not."

Horizon is just one such example.

In 2012, Arizona’s attorney general ordered ARS Rescue Rooter, a contractor that operates nationwide, to “ensure that technicians don’t invent dangerous conditions in the customer’s plumbing or ventilation equipment in order to make a sale”.

In Utah, Action Plumbing is accused of charging a woman with dementia $22,120 to replace fifty light switches.

In Pennsylvania, Curtis Total Service told an 85-year-old her furnace could “blow up at any second” and had her sign a blank contract.

Earlier this month, Illinois sued ABC Plumbing, Heating, Cooling & Electric for the same play Arizona outlawed fourteen years prior. The state alleges that technicians manufactured "emergencies or life-threatening hazards" and had customers sign tablets they “were not allowed to read”.

Those are a selection of cases that we are aware of only because they reached a courtroom. In this month’s member investigation ballot, this topic received more votes than the next two options combined. The comments the vote sparked were even more indicative that something is going deeply wrong in this market.

One comment: “my HVAC guy keeps pushing me to replace my 20 year old functional AC Compressor... which still works fine." Another: HVAC problems, once "a nuisance," are "now viewed with genuine terror at the likely cost of repair or replacement."

Three things happened to this market at once. First, private-equity funds bought the largest regional groups, implemented aggressive sales tactics and tied technicians’ pay to what they sell. Second, the manufacturers producing the hardware have been consolidated to the point that the thirty-four most common household names are produced by only six companies. Hardware prices are up 53% since 2020. Third, refrigerant rules, warranty terms and state licensing laws drive up maintenance and installation costs because a licensed contractor is the only person allowed to buy the gas or install the unit.

In what follows, I’ll uncover who owns your local contractor (and how to tell for yourself). I’ll map the consolidation on the hardware side and share which units are made in the same factories so you can price the cheaper ones. And I’ll share my playbook on what to ask any contractor when they show up, along with the simple tactics you can use to check their word.

Over the past two months, six readers have written in about sewing machines. I have never once mentioned sewing machines in this newsletter. All the complaints arrived independently, through the tip line.

I must admit that I am not a sewist. The extent of my capability is limited to ugly patches and button repairs, so I was hesitant to investigate the category at all. But given the sheer number and conformity of the reader messages, I knew there must be an interesting story here.

The general shape of all of the complaints is that Singer is riding on a reputation it earned a long time ago. Readers describe machines that are now plastic, sealed, and generally not worth repairing.

Several found that brands they assumed were competitors turn out to have the same owner. One reader bought a new Singer that broke in under a year. They also bought a Singer from the ‘50s, for a fraction of the price, that still runs.

The pattern is one we’ve covered many times before. A market that most buyers cannot navigate, where reputations carry little meaning and honest information is hard to come by.

What makes this investigation worth your time, whether or not you sew, is how effectively it demonstrates what happens when a reputation stops being something a company protects and instead becomes something it extracts.

What follows is the story of a once-great American company. I’ll also cover who actually makes the machines you see for sale, and what I’m planning on buying.

What Singer did when it started losing

Singer’s decline is often told as a private equity story. That isn’t really true. The financiers arrived in 1988, eight years after Singer announced the end of American household machine production. The true story starts 30 years before any Wall Street suits showed up.

Between 1951 and 1957, Singer’s share of the American market fell from about two-thirds to about a third. Japan built 134,000 machines in 1947 and more than a million by 1951. Its makers were aiming at America almost from the start. Brother's own history records an executive, Masayoshi Yasui, arriving in the United States in 1950 to assess the market.

Japanese machines used a multicam zigzag mechanism that enabled users to swing the needle sideways to complete a zigzag stitch.

By 1959, Japanese imports accounted for 22.6% of US sales of multicam zigzag machines, against Singer's 61.4%.

We know exactly how Singer decided they’d handle this new Japanese competition because the United States sued over their approach and the Supreme Court quoted sections of Singer’s internal mail over the course of the lawsuit.

Singer wanted a patent broad enough to cover the Japanese machines. It did not have one. Its own application was stuck in a Patent Office dispute with the Swiss firm Gegauf. Gegauf's application carried a priority date nine days earlier. Singer realized that if this dispute was fought to a finish, one side would lose outright and the winner's patent claims would be narrowed. A narrow patent could not be used to block anyone’s imports.

So Singer stopped fighting its rivals and recruited them. It cross-licensed with the Italian maker Vigorelli, royalty-free and worldwide, after Singer’s President had threatened to litigate Vigorelli's patents into invalidity. Then it settled with Gegauf, and the two agreed to help each other get claims "as broad as possible."

M.L. Waterman, a Singer VP, wrote to Singer’s president on 13 April 1956 explaining why an agreement with Gegauf was worth having:

Three weeks later, Singer’s patent attorney wrote to his own department expressing his optimism about how Gegauf was being handled.

Ultimately, Gegauf sold Singer the American patent application for a meager $90,000.

In January 1959, Singer petitioned the US Tariff Commission to have the President exclude every imported machine covered by that patent, naming Japanese and European makers alike. In that filing Singer put its own minimum price at $300 and said the Japanese machines sold for an average of $100 less, "often far below that figure."

The filing went in under Section 337 of the Tariff Act of 1930. That provision lets the Commission order Customs to stop infringing imports at the border.

Singer won at trial. In May 1962, Chief Judge Ryan found that "Singer's sole interest was to protect its '401' machine against all infringing machines." 

In 1963, the Supreme Court reversed him and held the whole arrangement an unlawful conspiracy.

The patent system is, in my view, a wonderful thing. It exists to reward a company or individual for building something better. Singer coopted the patent system to keep someone else’s better thing out of the country and out of the hands of consumers. They then asked the government to seal the borders behind it.

At Singer, legalism replaced innovation as their primary competitive strategy.

When no one is responsible for the product

Singer was a century old by 1956, but the family legacy was already long-gone. No member of the founding Singer or Clark families had run the company since Edward Clark died in 1882. What followed was professional management of the purest form.

The two men with the greatest impact on the company had no manufacturing or textiles background at all.

Donald Kircher was a lawyer that Singer hired in 1948 to look after its legal affairs. Kircher became President of the company in 1958 and immediately went on a spending spree. He bought 22 manufacturing companies, in audio, aerospace, defense electronics and calculators. Sewing fell from about 90% of revenue to about 35%. In 1963 the company dropped "Manufacturing" from its name, to better reflect the nature of its business. By 1974, the company owed $1.1 billion.

Kircher was replaced the following year by Joseph Flavin, a finance executive from Xerox, who wrote off $411 million of the losers.

Singer’s 1977 annual report described their sewing business as experiencing “dramatic growth”, representing an “excellent opportunity”.

Despite the optimism, their 1978 report outlined a new plan for their Clydebank manufacturing facility that would cut the labor force by almost half and make the plant "a cost-effective operation" again.

On 12 October 1979 Singer announced that it had been forced to close Clydebank and was establishing a $130 million reserve. The plant had been built in 1882 as the largest sewing machine factory in the world, and it shut the following June.

On 6 October 1980 Singer announced the end of household sewing machine production at Elizabeth, New Jersey. UPI reported it the same day:

The same story notes that Singer "is the sole remaining American maker of consumer household sewing machines and has been since about 1950." Elizabeth’s closure in 1982 marked the end of the entire American sewing machine manufacturing industry.

Enter, the bottom-feeders

The rest happened quickly. Paul Bilzerian took the company over in 1988 and, between July and October, sold eight of Singer's twelve divisions for about $2 billion. He was then convicted in 1989 for securities fraud in other companies' takeovers, not for anything to do with Singer.

The sewing division and the name then passed to a Hong Kong conglomerate, which put it into Chapter 11 in September 1999. Forbes counted $23 million in cash against $1.25 billion of liabilities, and 18,000 employees were "left to wonder what would become of them." 

Kohlberg bought the sewing business and the Singer trademark in 2004 for total consideration of $145 million, merged it with Husqvarna Viking and Pfaff in 2006, and sold to Ares in 2018. Ares sold to Platinum Equity in 2021.

There was no US factory on the asset book in any of those transactions. American production had been over for twenty-two years by the time Kohlberg wrote its cheque. What changed hands was a trademark, a dealer network and 80% brand recognition. The machines behind the name were already being built in Brazil and China to whatever spec the margin-target allowed.

Since then the Husqvarna factory in Sweden made its last machine on 18 May 2010 and production moved to Shanghai. The Swedish company now employs 58 people, doing design. The Husqvarna Viking page still says the brand "still designs high-quality sewing machines in Sweden" [my emphasis].

In 2022 the group's EBITDA fell 76%, Platinum put in $50 million of its own money as notes paying 20%, and Moody's cut the rating to Caa2 on leverage above 13 times earnings.

Singer's own history page runs 1851, 1855, 1858, 1865, 1870, 1889, 1890, 1921, 1929, 1933, the war years, 1951, 1952, 1975, 1978, 1985, and then 2001. There is obviously nothing good to speak of in those gaps.

Who actually makes the machines

Singer is the clearest case, but it is not the only one. If you go looking for a sewing machine today you will meet about eleven names. The reasonable assumption is that each belongs to a company that designs and builds the machine with its name on the front.

That is true of five of them.

Janome, Brother, Juki and Bernina design and build their own machines. The first three are independent and publicly listed in Tokyo. Bernina is private and family-owned, and in February 2026 it decided to move assembly of its flagship 990 and part of its mechanical production from Steckborn, Switzerland to its plant in Thailand, starting in May. Development, prototyping and spare parts stay in Switzerland.

Singer, Husqvarna Viking and Pfaff are one company. All three belong to SVP Worldwide. An American brand, a Swedish one and a German one, sold as three separate heritages, with an effective price ladder running between them (this reminds me of Tempur Sealy’s strategy in the mattress world). Dean Brindle, SVP's CMO, explained to Forbes in 2022:

The three brand sites all run on singer.com. SVP does build its own machines. The plants are in Brazil, Vietnam and China, and Shanghai is the most advanced of them. None of them are in America, Sweden or Germany.

Baby Lock, Kenmore and Necchi have no factory behind them at all. Baby Lock builds nothing themselves. Below about $300 its machines are made by Janome, above about $500 by Brother, and its sergers by Suzuki Machinery in Yamagata. Kenmore never built anything either. Sears contracted the work out, to White until 1958 and then to Japanese makers, and the 385-series machines were Janome's until Sears ended the relationship in 2013. Elna is a Janome subsidiary. Necchi is an Italian name now held by a Singapore company, distributed in America from a town in Texas, with no factory disclosed anywhere.

The machines got cheaper anyway

You would assume a story like that ends with people paying more for less. That is not strictly true.

Sears sold three Kenmores in 1956. The basic 58 at $89.95, the aluminum 71 in the middle at $134.95, and at the top the 84, a zigzag machine, at $239.95. Median family income that year was $4,800. So the middle machine took 2.8% of a year's income, about a week and a half of earnings. The top one took 5%, about two and a half weeks.

Median family income in 2025 was $110,500. Brother's cheapest machine, the GX37, is $149.99 and sews zigzag with a one-step buttonhole, which is what the top-of-the-line Kenmore did. That is 0.14% of a year's income, about half a day. Match the middle of the range instead and the 70-stitch CS7000X at $279.99 is 0.25% against the 71's 2.8%, a factor of eleven.

Put it the other way round. A week and a half of a 2026 family's income is about $3,190. That is what an ordinary 1956 sewing machine purchase is worth now, and it buys a genuinely superb machine.

Good machines do still exist. The Juki TL-2010Q, the Janome HD9, the Janome MC6700. None of them are cheap, but they never were.

What I'm planning on buying

I am not the person to tell you which machine sews best, so this section leans on experts who work with them for a living. Cale Schoenberg is a technician at his family's dealership, in Walnut Creek, California. He sells machines as well as repairing them. His newsletter is the best I found on this subject - I highly recommend it.

If you want a modern machine with all the features, the key factor is construction. Schoenberg describes a machine worth owning as one where "all the critical components of the machine are attached to a one-piece all-metal housing." On the cheap ones, he writes, "all the internal components are screwed in with self-tapping screws to the back cover of the machine, which is simply a one-piece plastic cover." The plastic shell is the structure, so when it flexes the needle bar moves with it and the stitch goes wrong.

On brands, technicians put Janome first for electronic reliability and Brother first for ease of use. Schoenberg writes that "it is rare to see electronic issues on any of these machines, but it is the most rare to see such issues with Janome." A Brother NS80e is $699. A Janome Skyline S3 is $1,199.

If you want something stout, Cale’s checklist for genuinely heavy duty is metal housing, a strong motor, a long enough take-up stroke, a vertical hook, a flatbed and a powder-coated frame. The machines that pass are old in design and mostly do one thing. A Juki TL-2010Q is straight stitch only, aluminum die-cast, 25 pounds, 1,500 stitches a minute, and he prices it at $1,099. A Janome HD9 does 1,600 stitches a minute in an all-metal cast aluminum body, $1,599 at his shop against a $1,999 list. Neither has a screen. At the cheap end he names a $200 Janome MyStyle 100 as a machine that passes the housing test.

One warning. The published price is often not the price. Dealers advertise at a manufacturer's minimum and sell below it in the shop, which is why a dealer's own site will tell you that a machine listed without an add-to-cart button is in-store only. Call and ask.

What I'm actually going to buy is an old one. Restoration shops start at $125 for a full service, and a serviced Singer 201-2, 301 or 15-91 generally runs a few hundred dollars. A full tune-up on a seventy-year-old Singer 221 is $99 at a shop in Wisconsin that charges a $159 minimum on a modern Husqvarna Viking. Singer published Adjuster's Manuals, full factory service documentation, and ISMACS hosts them free. There is a public directory of technicians at Sewing Doc Academy.

I want to learn to sew properly, and I would rather learn on something I can also take apart and fix myself.

What do you own?

Singer spent fifty years competing on everything except the sewing machine itself. What is left of the company is a hundred and seventy years of goodwill, and a business that consists only of spending it.

If you have a modern machine you are genuinely happy with, put the make and model in the comments or the community thread. I’ll build out the list and add a section in The Ledger.

You will be familiar with the tactics that many companies use to get out of honoring a warranty.

The sticker over the screw that reads “warranty void if removed”. The clause in the manual that warns “any modification voids the warranty”. The FAQ page that says “all repairs must be done by an authorized dealer”.

If you own a car or a lawn mower or anything with a circuit board, you’ve seen these warnings. I work on my own cars and motorcycles. For years I worried a claim would be denied because I changed my own oil or fitted an aftermarket part. That worry only ended when my Toyota’s warranty ran out in August.

So this week I looked into where the law actually stands. It turns out I never needed to worry. Most of those warnings claim more than the law allows. They stay in place because many consumers have their claim denied and stop there. You should know that you have recourse beyond that.

Today’s article outlines what the law says, what 111 current warranty policies say, and what to do if your claim is denied for the wrong reason.

What the law says

The Magnuson-Moss Warranty Act is a US Federal law passed in 1975. It bans any manufacturer from making its warranty depend on you using a particular brand for parts or a particular company for service.

The only exception is if the manufacturer provides those parts or that service for free. A company can also ask the FTC for a waiver. In 51 years, three have applied, and all three were refused.

The FTC’s rule on the Act has been on the books since 1977 and gives an example of the language it forbids:

"This warranty is void if service is performed by anyone other than an authorized 'ABC' dealer and all replacement parts must be genuine 'ABC' parts."

The rule is also clear about what warranty exclusions are permitted to manufacturers. They may deny claims for "defects or damage caused by 'unauthorized' articles or service." They may deny a claim "where the warrantor can demonstrate that the defect or damage was so caused."

So the obvious question is: whose responsibility is it to prove whether the damage was caused by my shoddy repair work, or their shoddy design?

The FTC is clear: "Before they can deny warranty coverage, the manufacturer or dealer must first prove the aftermarket or recycled part caused the needed repairs."

The law places the burden of proof on manufacturers, regardless of any exemption clauses they may include in their policy.

Repair, replace, modify

Before getting onto how to handle a denied warranty claim, we should cover an important distinction. The warranty warnings you see on stickers and in manuals lump three different acts together. The law separates them clearly.

1. Repair - Fixing it yourself, or paying a shop the manufacturer didn't bless. 

Protected. This is the first half of the FTC’s example policy above: "void if service is performed by anyone other than an authorized dealer." The bare fact of an outside repair cannot legally end your coverage.

2. Replace - Fitting another company's part.

Protected. This is the second half: "all replacement parts must be genuine 'ABC' parts." They can only deny if they can prove that the part caused the failure.

3. Modify - Changing the thing from how it was sold. 

The Magnuson-Moss Act never mentions modification, so this is the least settled of the three. The one time the FTC has drawn the line was in its 2022 order against Harley-Davidson. The agency barred Harley from saying a warranty "will or may be void" because of an unauthorized modification, but allowed Harley to exclude damage caused by one.

The FTC ordered that Harley’s policies be rewritten. The 2026 Softail warranty now reads:

"Taking your motorcycle to be serviced by a repair shop that is not an authorized Harley-Davidson dealer will not void this warranty, and using non-genuine Harley-Davidson parts alone will not void this warranty."

That is repair and replacement, in the FTC's mandated words. A few paragraphs later the warranty applies the causation test to parts, excluding damage from any part, "even those installed by an authorized Harley-Davidson dealership, that cause a genuine Harley-Davidson part to fail."

The order did allow Harley one exclusion with no causation test. A bike tuned with software lacking a California Air Resources Board order or EPA approval loses every powertrain claim "if Harley-Davidson or any authorized Harley-Davidson dealer has any information showing" it.

On repair and parts you have rights. On a modification you have the manufacturer's terms, so if you do tune a vehicle under warranty, read those terms first.

111 warranties, analyzed

A warranty’s legal text and the warnings written around it are not the same thing. The legal text is written by lawyers and is the true legal basis of the policy. The warning stickers, FAQ pages and customer support lines have no real legal basis, yet these are what most consumers go by.

Samsung's appliance warranty, for example, excludes only damage caused by outside service. Its appliance FAQ page, on the other hand, lists "unauthorized repairs or parts not supplied by Samsung" under what is not covered.

So I looked into the current US warranties of 107 brands across 13 categories, from cars to blenders, to see what each one says about outside repair and non-branded parts and whether their policies align with the FTC’s rules.

That came to 111 documents, because a few companies publish separate warranties for different product lines. Every document, the clause it uses, and a link to check it are in the appendix. They fall into five groups:

Of the eleven that say the warranty is void if anyone but the company or its dealer repairs the product, Vitamix sells a "10 Year Full Warranty." "Full" is a federal label, and a full warranty can't ask the owner to do anything beyond reporting the defect. Inside it states that the warranty "is also void if repairs ... are performed by someone other than either Vitamix or an authorized Vitamix Service Provider." Bose: "Unauthorized repair, modification or customization of your product voids this warranty." Sena: coverage is "terminated prematurely" if "an unauthorized person has tried to disassemble, repair, or modify the product." Just an attempt at a repair yourself is enough to void.

So 43 of 111 make coverage depend on outside repair or parts, or on any modification, without a causation test.

Only 9 tell the owner their actual, federally protected rights.

Whether that is carelessness or strategy, the effect is the same. The clauses work on every owner who reads them as binding and stops there. I’m certain that describes most people. The next section explains why companies are able to print those clauses, with little consequence.

By category, cars, motorcycles and major appliances are the cleanest: only 1 of 6 car warranties, 1 of 8 motorcycle warranties and 3 of 15 appliance warranties suggest that outside repair or aftermarket parts would void warranty. Home batteries and EV chargers (3 of 4), grills and kitchen gear (5 of 7), bicycles (5 of 7), and mowers and generators (7 of 11) are the worst. Bear in mind, the sample size is not very large, so conclusions on sector performance should be taken with a grain of salt. The full breakdown is in the appendix.

Another interesting data-point: in 2021 PIRG called and asked 43 appliance makers whether independent repair voids the warranty. All 43 said yes. Nobody has re-run the calls since. Perhaps that would be a worthwhile thing for us to do.

Why nobody enforces it

The Act gives the FTC no power to fine for a first violation. A fine can only be issued for breaching an order once one exists. The FTC admitted in 2021 that "current law does not provide for civil penalties or redress."

In 51 years the legislation has only produced two batches of warning letters, in 2018 and 2024, and three consent orders in 2022 with no fine.

The FTC hasn't sent a warranty warning letter or brought a warranty case since July 2024.

Owners don't sue over the clauses alone either. A sentence in a warranty does no harm until a claim is denied, and a denied claim on a blender or a mower is rarely worth a lawyer's time. On top of that, many warranties carry an arbitration clause that blocks a class action.

So given that there is no pushback from the government, plaintiff lawyers, or consumers themselves, companies will keep pushing those exclusion clauses.

Having said that, consumers have more power than many realize. What the law offers is clear recourse when a claim is denied. Under the Act consumers can sue for the cost of the repair, in small claims court or in arbitration. Once a legal dispute is initiated, manufacturers must prove that your part or outside repair caused the failure. If you hire a lawyer and win, the court can order the company to pay your fees. That is the only viable enforcement mechanism I see.

What to do when your claim is denied

Ignore the sticker. To repair or even inspect most things you have to open the case, yet the sticker says opening voids the warranty. It doesn't. The FTC's business guide explicitly lists a seal that voids coverage when broken as prohibited. A seal can lawfully be placed on a product and it does show that it was opened. They can raise that when arguing over what caused the failure. On its own it is not a reason to deny.

Keep the invoice. When a claim is denied after outside work, the argument is about cause. They will say the shop or the part must have caused the damage. Without records you can't answer that, and the denial may stand. With an invoice that names the part, the work and the date, you can show what was touched and when. Keep every receipt for the life of the warranty. If you do the work yourself, keep your own detailed records and keep all receipts for any parts/fluids you purchase.

Write the letter. When a claim is denied over outside repair or parts, write to the seller and then the manufacturer. Send it certified mail with a return receipt, so there's proof it arrived. The letter also clears a legal hurdle. Before you can sue under the Act, the company must be given “a reasonable opportunity to cure.” This letter is that opportunity. Quote their stated reason back. Cite 15 U.S.C. 2302(c) and 16 CFR 700.10(c). Ask them to name the defect they say your part caused, and the evidence for it. Give a date. Say what you want as restitution.

Most denials come from a customer-service tier that is applying a standard operating procedure from corporate. A warranty denial is your cue to ask for proof. A written request for proof, with the citations, is what you need to formally invoke the law. Before you do that, you are just accepting the company’s defensive, cost-saving procedure.

Here is a template you can use (you may need to adjust slightly for your specific case):

Kia denied a Pennsylvania couple’s engine claim last winter because their oil-change receipts lacked a VIN. Jamie and Debbie Rekasie had fourteen documented oil changes on a 2019 Optima that failed at 80,000 miles, inside Kia’s 100,000-mile powertrain warranty. Kia approved the replacement only after a Pittsburgh television station called. Michael Brooks of the Center for Auto Safety described the pattern:

"Often, the first tactic is to deny the claim and hope the consumer goes away."

Now you know how not to go away.

What it costs them to keep saying no

For illustration, here is an invented case, though I’ve applied real numbers:

The car is at 45,000 miles, inside its 5-year, 60,000-mile powertrain warranty. Oil changes were all done at an independent shop. An aftermarket cold-air intake was fitted. The transmission fails, and the service writer denies the claim "due to the aftermarket parts on the vehicle," an argument a GM dealer has used in real life. A replacement transmission runs $6,238 to $6,764. Call it $6,500. The law asks them to show precisely how an aftermarket intake or an oil change caused this particular transmission to fail. That burden is theirs from the start. In practice, nobody makes them honor it until you dispute the denial, in writing or in court.

To begin proceedings, your outlay is $10 to send the letter and $55 to file in Colorado small claims or $75 in California, if it comes to that. Depending on the state, you’re looking at a range of $30-$379 to file.

Granted, many of these warranties carry an arbitration clause. Arbitration sends disputes to a private arbitrator instead of a judge, and a class-action waiver says you must make your claim individually. But the consumer's fee in arbitration is capped at $225 and the company pays the rest, which can run to thousands before its lawyer bills an hour.

If you'd rather hire a lawyer, a court can make the company pay a winning owner's legal fees by the hour. In Waldock, an Alaska owner won $17,686 and the court awarded $36,526 in fees on top.

So you have $90 at risk on a $6,500 claim. Their cheapest move, once a letter citing the law and a request for proof arrives, is often to just fix the transmission.

Have you been denied recently?

If a company has denied a warranty claim of yours because you used an outside shop or fitted another company's part, reply to this email with the product and what they told you. Tell me whether you'd be willing to send the letter and, if it comes to it, file in small claims. It would be interesting to follow one or two of these from letter to outcome and report back.

Thank you, as always, for your readership and support.

“Where do you go for product reviews?”

That is one of the more common questions in my inbox. Thank you to those who have asked. This piece is for you.

In ‘On Independence’, I touched on the reasons why it has become nearly impossible to find trustworthy reviews of consumer products online.

Today, I want to share the sources that I have bookmarked and return to whenever I make a purchase.

It is by no means a complete list. It is just an account of the resources I have personally come to value over the years. My hope is that we might build a more complete list together in The Community.

Before the list, I'll outline the three mechanisms that corrupt almost every review you read online, then the three questions I ask of any reviewer. Together, they'll tell you in about five minutes whether a source is worth your time.

How the machine works

The degradation of trust in online product reviews is caused by three primary mechanisms.

1. Seeding - Get free products in return for reviews.

Brands and marketing agencies send product to reviewers on a schedule, and the units keep coming as long as the coverage stays positive. ToolGuyd exposed the terms of this sort of arrangement:

"We require you to post on social media before the 30th," "one pre-determined item from each brand every other month," and "In return...you keep the product."

Very little is disclosed. Often, all we see is a “Paid Partnership” tag, with no mention that product was free, or that there was a deadline, or that future freebies depend on the content of the review.

The FTC's own guidance counts a free product as a material connection and says a platform's built-in tag is "not necessarily" adequate disclosure.

But it’s not surprising that with this level of vagueness in the rules, many creators choose the minimum-viable disclosure.

And what happens when brands don’t get the ratings they want from influencers?

In December 2020 NVIDIA told Hardware Unboxed it would stop sending review cards because its reviews "continued to focus singularly on rasterization performance," and would revisit the decision "should your editorial direction change." The email was published, the backlash was immediate, and NVIDIA reversed within two days.

Motorcycle review channel FortNine's Ryan Kluftinger described his experience: "we get a press bike, we make a video, and they don't like some negative things we've said, so they never give us anything else."

2. Sponsorship - Get paid in return for videos, posts or sales.

Furniture maker Anne Briggs explains the lexicon well in an interview on Christopher Schwarz’s substack:

"Ambassador" means free product, "affiliate" means paid per sale, and "sponsorship" means paid for the affiliation.

Agencies offer "a three- to six-video deal," and the contracts "typically require you to give the brand way too much control over what you say about their products."

Here, the sponsor relationship is at least disclosed in the ad read. However, what isn’t disclosed is that the same brand’s products get reviewed elsewhere on the channel or blog, or that the discount code in the description pays the reviewer.

The FTC has settled with companies over exactly this. Warner Bros. paid influencers "from hundreds to tens of thousands of dollars" to cover a video game, "required the influencers to promote the game in a positive way and not to disclose any bugs or glitches they found," and had them put the disclosure where it was "visible only if consumers clicked on the 'Show More' button." The videos were viewed 5.5 million times.

Global influencer spend is now about $33 billion a year.

3. The costume of testing - Adopt the language of objectivity, without the methodology.

When you search “best air purifier”, Google’s algorithm decides which page you see first. Ranking really matters, because the first results get all the clicks. Today, nearly every page linked to a query carries shopping links that pay the publisher a commission when you buy.

So a publisher’s ranking is its revenue. A page that sits at the top of the “best air purifier” query is worth a lot of money every month.

In 2022 Google announced it would move pages up if they showed evidence of actual testing. Its guidance still asks for "quantitative measurements," and "the number of products that were tested, what the test results were, and how the tests were conducted."

Testing is expensive. Claiming you test rigorously is free.

Real Simple's best air purifiers page has a section called "Our Testing Process." It describes unboxing, app setup, filter changes and appearance. On performance, it says this:

"Finally, we looked at the performance, considering any substantial difference we felt in the overall air quality."

No test instrument is named. No robust methodology. No measurement appears anywhere on the page. The only numbers they provide are the manufacturer’s own specs.

Its sister title Better Homes & Gardens says "we tested 144 options in our lab and at home" with "an air quality monitor that measures PM2.5s and TVOCs" and "a decibel meter." Some numbers do appear: a noise range on each product, and on a few, a line like "decreased PM2.5 numbers by 90% and TVOCs by 80% when tested in an enclosed environment."

None of it is checkable by the reader. There is no standardized methodology, no data table, no charts, and no way to tell which figures came from their meter and which came from the manufacturer. There is no consistency in how they report numbers at all.

Both titles belong to Dotdash Meredith, which reported $1.1 billion in digital revenue last year, with affiliate commerce up 21 percent.

HouseFresh is an independent outfit that actually tests air purifiers, publishes their methodologies and provides the raw test data. They documented in 2024 that BH&G's testing claim first appeared in the twenty days before Google's update took effect.

To see the difference, let’s look at the same product reviewed by both outlets. BH&G's page names the PuroAir 240 "Best for Large Rooms". HouseFresh bought the same unit and gave it one star.

Do yourself a favor and actually look at both reviews, then judge them for yourself.

BH&G is just one offender. Every industry has dozens of publishers playing this game.

A real test produces numbers, and numbers produce losers. A costume lab produces ranked lists of only winners in spurious categories with buy buttons on every row.

Sadly, the pantomime wins. In a study of 186,000 Google searches where a content site ranked first, 86 percent of those first places belonged to a large media brand. HouseFresh's own traffic from Google fell from 4,000 visitors a day to 200.

What trust is made of

There are three simple checks that make identifying trustworthy sources a whole lot easier.

1. Who bought the unit? Retail purchase is the single strongest signal of honesty there is. A manufacturer that supplies units can provide hand-picked ones. It can also withhold the next one if a review goes badly.

Consumer Reports states their standard plainly: “CR buys all the products and services it rates at retail, just as consumers do”.

Solereview literally posts an image of each receipt for every running shoe it reviews.

That is trustworthy.

2. Who pays for the verdict? Entirely reader funded is best, but that is a tough standard to uphold.

Consumer Reports spends about $30 million a year of its membership revenue buying and testing products. Almost nobody else can achieve that kind of scale. Readers rarely subscribe based on a review they saw. They search for one, read it once, and leave. So most reviewers get paid one of two ways: a commission when you buy, or money from the brand.

Those are wildly different incentives. A commission pays the same whichever brand a buyer picks. Brand money pays for a specific verdict. That’s why affiliate revenue can be ok, assuming the reviewer is objective and discloses the relationship. Sponsorships are never ok, because they necessarily undermine editorial independence.

3. Show me a negative verdict. A real test produces losers. If a tester’s results page only names the 10 best in various spurious categories, that is not a good sign. Thankfully, it is an easy-to-spot format.

So look for the loser. Does the reviewer tell you what not to buy? Vacuum Wars keeps a list of robot vacuums to never buy.

Independence is expensive. RTINGS spent $714,000 buying 618 products last year. I already mentioned Consumer Reports’ $30 million a year commitment. Torque Test Channel is a one-man operation. It sells a $15 rankings spreadsheet that has moved 13,800 copies. That spreadsheet is easily the best money I’ve spent on tools in years. Providing true value like that is how publishers can reject brand money and become incorruptible.

That is no small feat. Every source below has found a way to get paid that doesn’t depend on the brand liking the verdict.

Testers vs. practitioners

The sources below come in two kinds. Both are useful for different reasons.

Testers aim to provide an objective number. They conduct the same test, across every competing product, side by side, with a method anyone could repeat. This enables them to produce objective rankings of the products they are testing, in the tasks they were designed to do.

Practitioners provide something more: judgment. They did the work for a living. A cobbler has seen a thousand soles come unglued and knows which construction fails and why. A master technician has had the same part in his hands hundreds of times and knows which model years to avoid. That knowledge is demonstrated in how they handle the object and in what they notice that you wouldn't. When you watch them work, you borrow a competence you never had time to build.

This is the second of four essays on the American automotive industry.

The first described the effects that financialization has had on the market. It showed how automaker profits transitioned from vehicle sales to loans, and outlined the impact that this trend has had for consumers.

Today’s article examines the impact that our regulations have had on both consumer choice and long-term cost of ownership.

The next essays in the series will cover the global automotive supply chain and the erosion of our right-to-repair.

The curse of complexity

A modern car is a complicated web of systems. Each contains three components: a sensor that measures things, an actuator that physically alters things, and a computer that reads sensor states and determines which actuators to actuate.

Those computers are called electronic control units (ECUs). Counting their prevalence over time gives us a good indication of how complicated cars have become.

Manfred Broy, a computer scientist at the Technical University of Munich, found that premium cars contained “not less than 70” ECUs in 2006. A current Volvo XC90 carries upwards of 110, and a BMW 7-series can run to 150 or more.

Beyond electronics and computers, mechanical complications are also on the rise.

The EPA publishes the fleet-wide technology content of every new vehicle sold in America, every year:

The table above contains just a few examples of systems that hardly existed a generation ago and are now ubiquitous. Each comes with its own array of sensors, actuators and ECUs. Each adds to the underlying complexity of the vehicle.

Fitting all of it is cheap. The marvel of modern manufacturing means automakers are able to add all of these systems without massively increasing their costs of production.

The problem is maintaining them. Every sensor is a part that fails. Every controller needs accurate information about the state of the car. When it does not get that information it protects itself by shutting down. A single failed sensor puts the car into limp mode, capped at low speed and reduced power, until somebody diagnoses it.

Every system now has multiple expensive points of failure. Every ordinary repair is now complex.

Consumers have no option to opt-out. There is no bare-bones car on the market.

The twenty-dollar mandate

A lot of this complexity is driven by regulations. Many of them are effective. Let’s take a look at one positive example.

Automatic emergency braking uses a forward-facing camera to watch the road ahead. Some systems add a radar sensor. If the car in front stops and the driver does not react, your car brakes itself. NHTSA made it mandatory in a rule finalized in May 2024, with compliance required by September 2029.

It is cheap. NHTSA costed the whole rule at $354 million a year. Of that, $282 million is software applied to every vehicle. Only $72 million is hardware, because most new cars already have a front-facing camera fitted. Across roughly 17 million new vehicles a year, that comes to about $21 a car.

AEB is also effective. The Insurance Institute for Highway Safety found that automatic braking cuts front-to-rear striking crashes by 50 percent. Crashes causing injuries are reduced by 56 percent.

NHTSA puts the cost per equivalent life saved at $550,000 against the $11.6 million the DOT values a life at. By their own accounting, the rule returns roughly $20 of benefit for every dollar it costs.

This regulation is a no-brainer, given the number of lives it saves on the road.

That being said, there are significant hidden costs that aren’t captured in the agency’s analysis. The crash-detecting camera typically sits behind a windshield and looks through it. Its aim needs to be calibrated to within a fraction of a degree. New glass has a marginally different thickness and curvature, so every windshield replacement now requires a complex calibration: a technician, a target board, a level floor and a scan tool.

The Highway Loss Data Institute compared trim levels of the same models, where the system had been optional. Across more than 270,000 glass claims:

“Hondas with front crash prevention were over 12 times more likely to have a glass claim of $1,000 or more. Subarus with front crash prevention were over 6 times more likely.”

AAA had estimators price the same job. In its 2023 report, the average windshield replacement ran $1,439, of which $360 was attributable to the driver-assistance system.

You cannot decline the calibration to save the money. Under the American national standard for auto glass, a shop must calibrate or refuse the job outright.

Americans replace 11 to 15 million windshields a year. Calibration runs $300 to $400. If most of those cars now need it, the recurring national bill is somewhere between $3 billion and $5 billion annually.

NHTSA costed the rule that created the bill at $354 million.

That is a cost-tail that is roughly 10x the mandate. For an emergency braking system that is shown to be effective in saving lives, that gap may be understandable. For many of our regulations, it is not.

Telling the good from the bad

NHTSA publishes a cost-benefit analysis for every rule it issues. The key metric is cost per equivalent life saved. It allows us to compare the costs and benefits of one rule against another, and set those against the value the department itself puts on a human life.

Top to bottom, that is a variance of about a hundred to one. The seat belt figure was derived by dividing the per-vehicle cost by the 15,485 lives NHTSA credits belts with saving in 2012. Every other number is the agency's own.

Three of the rules in the table above failed on the government’s own accounting.

Tire pressure monitoring. The regulatory analysis found net costs rather than net benefits across nearly every scenario it modeled. The rule went in anyway, in 2007. To satisfy it, manufacturers put a pressure sensor and a radio transmitter inside all four wheels. Each has a sealed battery that cannot be replaced on its own. When it dies after five to ten years you buy a whole new sensor, and the dashboard light stays lit until you do. The rule prevents an estimated 120 deaths a year.

Roof crush resistance. In 2009 the agency tightened the standard and extended it to vehicles weighing up to 10,000 pounds. Its own analysis of that extension is blunt:

“Nearly all alternatives covering vehicles from 2,723 and 4,536 kilograms (6,001 and 10,000 pounds) GVWR yield net losses rather than net savings to society.”

For that class of vehicle the agency expected to prevent two deaths a year, at somewhere between $18.8 million and $140 million each.

Backup cameras. The final rule says it outright:

“While the costs of the rule exceed its quantifiable benefits, Executive Orders 12866 and 13563 call upon us to assess the costs and benefits of a rulemaking, including those costs and benefits that are difficult to quantify.”

The cameras prevent thirteen to fifteen deaths a year, at $15.9 to $26.3 million each. In fairness to the agency, it did not choose to implement this rule itself. Congress ordered it in the Cameron Gulbransen Kids Transportation Safety Act of 2007.

We pay for the rule at the bottom of that table exactly the way we pay for the one at the top. Both are fitted to every car, and no version of any car exists without either.

Incentivizing complexity

Safety rules are simple. They mandate that certain equipment be installed on new vehicles. Environmental rules are different.

The EPA set each manufacturer a fleet-average carbon dioxide target, measured in grams per mile across everything it sold that year. For model year 2026 the industry-wide target was 161 grams per mile.

A manufacturer that came in above its own target ran a deficit, and it had three model years to erase that deficit or buy credits from a competitor who had spare ones. If a manufacturer failed at both, then its cars were not covered by a certificate of conformity, which exposed it to a civil penalty of $59,114 per vehicle sold.

So grams were money and therefore credits were money, so the manufacturers went hunting for credits.

A credit does not reduce what comes out of the tailpipe. It reduces the emissions number the manufacturer reports. Finding a real gram of efficiency requires difficult engineering: a better combustion chamber, a lighter body, a smarter transmission. That is expensive and slow. Fitting a credit-earning system is cheap and quick. Both count the same against the target.

The EPA published a list of technologies that earned credits automatically. They are called off-cycle credits, because the claimed benefit does not show up on the standard fuel economy test. They were applied after the true emission testing was done.

Some of the list:

Grams of carbon dioxide per mile, credited on every vehicle fitted. A separate schedule covered exterior lighting, running down to 0.08 grams for a more efficient license plate lamp.

Claiming from that list took no work at all. The manufacturer fitted the technology, looked up the published number, and subtracted it from the fleet emission average it reported to the EPA. There was no test to run, no application to file and nothing to approve. The agency's only power of review was that it "may request" supporting data afterwards.

A manufacturer could take up to 15 grams per mile this way, close to a tenth of that 161-gram obligation, without conducting a single test.

The list just outlines the default credits for each system. A manufacturer that wanted more credits could apply, submit a methodology, justify why ordinary testing would not capture the benefit, and wait for the EPA to publish the application for public comment. Most of the large manufacturers did. Ford applied for 1.2 grams per mile for an improved window anti-fogging strategy. Jaguar Land Rover requested 1.6 grams for a 48-volt motor-generator. Mitsubishi requested 1.1 for an air conditioning compressor.

The whole exercise became a hunt for grams, a fraction of a gram at a time.

This is why the credit structure is so ineffective. It creates perverse incentives. When the government mandates something it has to defend the requirement. The industry then has every motivation to attack the numbers. When it offers a credit instead, manufacturers are incentivized to claim every gram on offer, regardless of how effective they truly are at reducing emissions.

Ventilated seats are a good illustration. The list paid 1.0 gram per mile on a car, on the theory that a cooled seat lets you tolerate a warmer cabin, so the air conditioning runs less. Then manufacturers applied for more.

Fiat Chrysler asked for 2.3 grams on cars and 2.9 on trucks, more than double the published value, citing a National Renewable Energy Laboratory study from 2005 and later work by the seat supplier. A separate request from General Motors rested on a trial in which four men sat in two Ford Focuses in Golden, Colorado for twelve days in September 2016 and reported feeling equally comfortable at a cabin temperature 2.6 degrees Celsius higher. That trial was not blinded, and the heat thrown off by the seat fan itself was left out of the model. That is the evidentiary standard we are talking about.

In model year 2023 the industry averaged 8.6 grams per mile of off-cycle credits, and twelve of the fourteen largest manufacturers finished the year emitting more than their standard allowed. They complied on credits.

Stop-start earned 1.5 grams on a car, and 2.5 grams if the system included an electric heater circulation pump. So manufacturers added the pump. On conventional gasoline engines, stop-start went from 0.6 percent of production in model year 2012 to 57.6 percent in 2024. That is what happens when you incentivize complexity.

Interestingly, in February 2026 the EPA repealed the entire greenhouse gas program. Its own press release described the stop-start credit as "nothing more than a regulatory loophole that allowed automakers to claim GHG credits without delivering real-world emission reductions or benefits to human health."

The credits survived though. The same technologies at the same values now count against fuel economy standards instead, through model year 2032. Stop-start still pays 2.5 grams on a car and 4.4 on a truck.

So despite the EPA's own finding that the credit delivered no real-world emission reductions, nothing comes out of the car. Three-quarters of model year 2026 models still ship with stop-start as standard equipment. It stays there for the life of the vehicle, and so does the bill for maintaining it.

The bill arrives a decade later

CarMD compiles what American shops find when customers drive in with an illuminated check engine light. It logged more than 39 million vehicle failures and recommended repairs in 2025. The most common one in the country was replacing the catalytic converter, at an average of $1,511. Second was the oxygen sensor. Sixth was the evaporative emissions purge valve. Those three come to 19 percent of the list. Nearly a fifth of what an American check engine light reports is a part fitted to control emissions.

The evap system is a great example of the issue. It is a charcoal canister, a purge valve, a vent valve and a pressure sensor. Its entire job is catching fuel vapor so it does not escape while you refuel. Gas stations in smoggy states were separately required to catch the same vapor at the nozzle. In 2012 the EPA waived the gas station requirement, finding the two systems "redundant" and noting that they sometimes interfered with one another. It let states drop the half that belonged to the gas station. The half bolted to your car stayed. It is still throwing codes. A lit check engine light is an automatic fail on an emissions test, so in the states that require one you cannot renew the registration until somebody has fixed it.

The catalytic converter is another such example. Tightening nitrogen oxide limits worldwide raised the amount of rhodium each converter needs, and rhodium peaked above $29,000 an ounce. That turned a bolt-on part hanging under every car into something worth stealing with a battery saw. Theft claims went from 16,660 in 2020 to 64,701 in 2022. State Farm alone paid $115 million on 45,000 claims in 2022, averaging $2,500 each.

The same pattern runs through the rest of the car.

A stop-start system cycles the starter tens of thousands of times more than a conventional one. That of course wears starters out faster, so manufacturers must include an over-engineered and expensive version to compensate. Worse than that though, an ordinary lead-acid battery will not survive so many cycles, so the car needs an absorbed glass mat battery instead, at 40 to 100 percent more. On several brands, including BMW, Mini, Audi and Volkswagen, fitting one requires a scan tool to register the new battery with the car. Skip that step and the charging system keeps treating it as the old, tired battery it replaced, and the new one dies early. All of that for a job that used to take five minutes in any AutoZone parking lot.

There is no opt-out

A manufacturer could build a car that meets only the mandatory minimum. Nothing stops it. What they can’t do is sell you one with any of the mandatory items taken out.

The law that governs this is 49 U.S.C. § 30122:

“A manufacturer, distributor, dealer, rental company, or motor vehicle repair business may not knowingly make inoperative any part of a device or element of design installed on or in a motor vehicle or motor vehicle equipment in compliance with an applicable motor vehicle safety standard.”

This law binds every party you could possibly transact with, and all of them are forbidden to hand you the car without the mandated safety features.

The statute in no way limits the owner. You may disable equipment on your own car in your own garage. You are permitted to want it gone. You are not permitted to buy it that way.

Slate Auto is attempting to demonstrate where the simplicity floor sits. It is a new American manufacturer whose entire proposition is deleting unnecessary components: no paint, no infotainment screen, no speakers, crank windows, $24,950 before delivery. Its head of design says the interior is "not corrupted with a bunch of screens that are useless to people."

It ships with a screen anyway, behind the steering wheel, because FMVSS 111 requires the rear camera image to appear within two seconds of shifting into reverse.

What the motorcycle market proves

Two machines, both on sale new in America today.

A Suzuki DR650S costs $7,299. Air-cooled, single cylinder, carburetor, no ABS, a design essentially unchanged since 1996. Its owner can fix it in the field, as I have done many times.

A BMW R 1300 GS starts at $20,395 and can be specced to $30,000+. Four riding modes, cornering ABS, traction control governed by an inertial measurement unit, radar cruise control and an electromechanical clutch.

People ride both on the same trails. One buyer wants the most capable machine available and will pay for it. The other wants something agricultural that they can repair themselves and that will still run in thirty years. The motorcycle market serves both of them. The car market serves only the first.

The regulatory difference is the reason. Cars must meet dozens of federal safety standards. Motorcycles are subject to only a handful of restrictions affecting brakes, lights, mirrors, tires and glazing. None of them govern what happens to a rider in a crash.

The DR650 shows what regulation does to consumer choice across different markets. Sales were halted in Europe around 2001, because a carbureted engine cannot pass European emissions standards. It left Australia in November 2021 when an exemption to that country's anti-lock brake mandate ran out.

There is a fair objection here. A motorcycle crash usually kills only the rider, and motorcyclists die at about 28 times the rate of car occupants per mile. Riding is closer to a risk you take on yourself, so it makes sense to regulate it differently.

But if that principle governed car regulation, we would expect the heaviest rules to fall on the harm a car does to other people.

They do not. Washington has spent sixty years regulating how a car protects the people inside it, and rather less on how it affects the people outside. It has never written a rule about what the front of the car does to a pedestrian it actually strikes. Nothing covers the hood, the general size or the shape of the nose. NHTSA conceded as much when it proposed one in 2024, saying it would add pedestrian crashworthiness requirements to the federal standards "for the first time."

Where this leaves us

Technology itself is not the problem. Automatic braking prevents half of all front-to-rear collisions for about twenty dollars a car. Stability control saves more than a thousand lives a year. Anti-lock brakes are close to free and they work.

Nor is this an argument against regulating cars. Driving imposes costs on people who never agreed to bear them. That is what regulation is for.

The problem is that rules go in easily and only come out with great difficulty. Tire pressure monitoring went in with negative net benefits and has never been revisited. NHTSA published a finding that the backup camera rule cost more than it saved and issued it anyway because Congress ordered it.

I do not know what the fix looks like. Sunset clauses expire good rules along with the bad. Periodic review could work, but also adds to the already inflated bureaucratic load. If you have seen an efficient policy reform process that works, in any domain, then send me a note. There must be a better way.

Until we figure it out, consumers can expect to pay a steadily increasing amount for repairs that were once routine, or not necessary at all. Windshields, wheel sensors, converters, batteries that need a computer to install.

Jobs that used to take an hour, a wrench, and a little elbow grease.

Join the discussion, share product recommendations and suggest topics for me to investigate next in the Worse on Purpose community.

Every category I have covered has had what you might call ‘a golden age’. A period when buyers could truly judge quality. In every case, I published a story about how the factors we could once judge were somehow removed or concealed from consumers.

Supplements never had that golden age. No one has ever been able to judge the quality of a capsule by inspecting it. Even so, the category has grown even more challenging for consumers to navigate over time.

Our regulations are where this story starts. Since 1994, no supplement maker has been required to demonstrate to the FDA that its claims about its products are true, or even that they contain what the label claims, before selling them. US laws place the burden of proof entirely on the government itself. The statute reads: “the United States shall bear the burden of proof on each element to show that a dietary supplement is adulterated”.

To pull a product, the government must test it itself. That is not possible in today’s market.

In 1994, when Congress wrote the rule, the supplement industry sold roughly 4,000 products and did about $4 billion that year. The FDA now says there may be as many as 100,000 supplement products on the market. Americans spend nearly $73 billion a year, and three in five adults take a supplement in any given month.

The FDA is unable to provide firmer figures because it is hamstrung. Cara Welch, who heads the agency’s Office of Dietary Supplement Programs, explained its predicament plainly: “Under our current authorities, FDA has no systematic way of knowing what dietary supplements are on the market, when new products are introduced, or what they contain.”

The agency responsible for policing this entire industry has no mechanism for even listing the products within it.

This essay is about how that happened, and what you can do to ensure you are getting the supplements you paid for.

A few factories and a deluge of brands

Walk into a CVS or Walgreens vitamin aisle and you will be confronted by a sea of gummies. Hundreds of jars covering dozens of independent health claims. Sleep gummies, energy gummies, stress gummies, focus gummies, and so on… Many of the brands you see today did not exist ten years ago. A large share of them are manufactured in the same buildings.

Catalent tells prospective clients it serves "21 of the top 25 self-care companies."

Its competitor, Vitaquest, boasts that it manufactures for "more than 500 brands around the world", launched "over 800 new products in the past year," and proudly lists where its products end up: Amazon, Costco, CVS, GNC, Sam's Club, Target, Vitamin Shoppe, Walmart, Walgreens and Whole Foods.

Robinson Pharma runs a plant producing 20 billion softgels and 7 billion gummies a year.

International Vitamin Corporation calls itself "the leading private label vitamin and supplement manufacturer in the US," and says it is "trusted by thousands of retailers and brands worldwide."

Not one of them will tell you which specific brands they manufacture for.

The prevalence of contract manufacturing practices only becomes visible to the public when something goes wrong.

In January 2020 the FDA published a recall covering every dietary supplement one company had made between January 2013 and November 2019. Seven years of production, pulled in its entirety, at once.

The company in question was ABH Nature’s Products out of Edgewood, New York. Inspectors found “significant violations of manufacturing practice”.

It happens on smaller scales too. In November 2021 a single manufacturer in Norcross, Georgia recalled 77 products in one go, sold under more than seventy different brand names across seventeen states. Alpha, Chemix, Muscle Rage, Wycked Naturals and dozens more you have never heard of. The labels had failed to declare that the products contained milk.

All of this raises the obvious question of how a market in goods so intimately tied to our health can allow so many new participants, with such little oversight.

Enter, the gummy

The supplement game used to be difficult to get into. Making tablets and capsules at scale required something akin to pharmaceutical grade production processes. This created steep barriers to entry.

Making gummies, on the other hand, requires a manufacturing process akin to making Twizzlers.

Once the format transitioned, starting a supplement brand moved from being a manufacturing problem to a marketing one. You do not even need a formula to get started. The same Catalent About page offers brands "250+ Formulations available for Self-Care brands to get a head start."

The cost of launching a supplement brand fell to a logo, an Amazon listing and one production run. Triton Nutra Group in Chandler, Arizona, states that its "minimum order quantity for private label is 1,000 units per SKU". Nutricraft quotes minimum order sizes for gummies at 1,000 bottles and says a first run starts at $8,000.

That explains how a market goes from 4,000 products to over 100,000 without any new brand having to build anything. The plants still had to be built. Somebody else was buying them.

Where the money went

What we are left with is a market that has grown some 20x over the past 30 years, while production concentrated into fewer hands. Following the money reveals some important facts for consumers.

In 2017 a private equity firm called Highlander Partners started buying gummy plants. It bought four. Its president, Jeff L. Hull, described the plan: "In 2017, we started with a simple investment thesis of creating manufacturing capabilities around fortified/functional products in consumer-preferred formats." In August 2021 Highlander sold the assembled business to Catalent for $1 billion.

Lonza's capsule division, which produces 260 billion capsules a year across ten sites, is going to the PE firm Lone Star Funds for about $3 billion. Lonza keeps 40%.

IVC bought Perrigo's entire US vitamins and supplements business in 2016, then bought GNC's factory in 2019 for $176 million.

And Vitaquest, the plant producing supplements for more than 500 brands, is wholly owned by CK Life Sciences, the Hong Kong company in Li Ka-shing's group.

More than four billion dollars in disclosed prices, most of it inside five years, and every dollar of it paid for plants and production lines.

So what was happening to the brand names over the same period?

What happened to the brands?

Church & Dwight paid $650 million in 2012 for Vitafusion and L'il Critters, then the biggest gummy vitamin brands in America. It called the purchase "a new growth platform in one of the fastest-growing segments of the attractive vitamin/mineral/supplement category." In December 2025 it sold both. It did not say for how much, though it did say that the two brands had shrunk to under 5% of its expected sales for the year.

GNC sold its factory in March 2019 and filed for bankruptcy fifteen months later.

Bayer switched off the personalized vitamin service Care/of in July 2024.

The Vitamin Shoppe changed hands twice in six years, with a $2.6 billion buyout and a Chapter 11 in between.

This can’t be explained as a category going out of fashion. The plants were being bought at record prices over the same years. What lost its value was the brand itself.

A supplement brand is now cheap to start, cheap to abandon and cheap to replace. Names like that do not accumulate reputations worth protecting. The value in this market sits in the production plants instead. The plants do not sell to you. They do not appear on the label and do not compete for your trust. They are completely isolated from reputational damage.

What ends up in the jar

The gummy format is the hardest to produce honestly. The people who make them say so in their own trade press.

Pierre Albert Thomas of Rousselot, which supplies the gelatin, says "the real complexity comes with the introduction of active ingredients," and that brands historically had to "reduce the active-ingredient content of gummies to mitigate interactions with the base, or switch to a more conventional—a.k.a. less-exciting—supplement format."

Doug Brown of Sirio Pharma describes a process running on "complex temperature, humidity, and pH conditions that require careful stewardship of ingredients."

Some direct testing has also been done.

Pieter Cohen's team bought 25 melatonin gummy products and found 22 of them inaccurately labeled, running anywhere from 74% to 347% of the dose printed on the jar. One contained no melatonin at all.

NOW Foods, which makes and sells supplements itself, bought 25 vitamin C gummy brands off Amazon and paid an accredited lab to test them. They published the results last week. Twelve had almost no detectable vitamin C. Five claimed 1,000mg or more per gummy and contained under 2mg.

When we buy supplements, we are often not getting the dose we paid for. Sometimes that is frustrating, but benign. In other cases, like the Norcross recall outlined above, the label is wrong in more dangerous ways: those seventy-seven products contained milk and did not say so.

The one brand worth $3.8 billion

That is not to say there are no supplement makers that hold brand value. Earlier this month, Procter & Gamble agreed to buy Thorne for $3.8 billion in cash. The private equity firm L Catterton had taken Thorne private in October 2023 for about $680 million. They made roughly five and a half times their money in thirty-four months.

What P&G is buying, according to Thorne's own announcement is a vertically integrated plant in South Carolina, in-house doctors and researchers, products certified under NSF Certified for Sport, and more than a hundred professional sports teams as customers. Paul Gama, who runs P&G's health care business, said Thorne "strengthens our position in premium wellness with a trusted, science-backed brand."

Look at what the two ends of this market have in common. The gummy brands that got written off were selling a name attached to a formula from a catalogue. The brand that appreciated is selling verifiability and reputation; the fact that somebody else checks its work. In a market where consumers can verify nothing for themselves, being the brand that employs an independent third-party to check its products turned out to be worth $3.8 billion.

Which makes what happens next incredibly telling.

P&G sells to everybody, through Target and Walmart and other such mass-outlets. Thorne sells premium, through practitioners, professional teams and direct to top-of-market consumers.

The commercially obvious move is to take the brand into the mass channel. NSF certifications screen every production lot rather than certifying a formula once, so the cost of the certification scales with volume. The most profitable way for P&G to sell Thorne at Walmart is to stop paying for the testing.

To be clear, that may not happen. P&G has done nothing untoward yet.

But just so we have them, here are the three things that would tell us the extraction has started:

• NSF Certified for Sport quietly dropped from any Thorne product

• Production moved off the South Carolina facility

• An uncertified mass-channel line launched under the Thorne name

I’ll be checking in periodically. I’ll let you know if any of these come to fruition.

A trustworthy supplement company needs to own the plants it manufactures in so that it has a reputation attached when something goes wrong. It must submit its production lots to a tester outside the company, whose verdict it does not control. And it should publish that testing publicly, even when results may be negative.

Almost no one clears all three, though Thorne comes close.

What to do about it

Start with the format. The people who make gummies say themselves that it is the hardest format to get actives into correctly, and both retail tests above found gummies failing badly. Tablets and capsules are made on pharmaceutical lines by firms that have done it for decades. If both are available to you, buy the tablet version over the gummy.

But that is no silver bullet. You should also look for a third-party test mark. NSF Certified for Sport is the strictest one available, because it screens every production lot rather than certifying a formula once. You can check a specific bottle against its database using the lot number printed on the label. Bear in mind though that the same directory carries Gatorade, Red Bull and 265 separate listings for Celsius energy drinks. All it tells you is that there are no banned substances present, and that the label matches what is actually in the product.

USP Verified is the most thorough formula-level program, and it pulls products off retail shelves once a year to re-test them.

That USP mark is rare though. USP's directory lists 159 verified products across sixteen brands, against 80,000 to 100,000 products on the American market. 99 of those are Nature Made, which is owned by the Japanese pharmaceutical company Otsuka. Looking for the USP seal narrows you to a short list of products from a small number of large companies. They are the ones who pay for the testing.

And be clear about what a seal does. The Pentagon, which has spent years trying to keep service members away from bad supplements, puts it well: "Third-party certification does not mean the product has been evaluated for safety or effectiveness. It means the product content and label match."

Other important forms of testing

You should also know about a small number of organizations that buy supplements at retail and test them without being paid by the companies that made them. They are the closest thing this market has to an independent auditor.

ConsumerLab has tested more than 7,000 products from over 1,000 brands since 1999. It buys everything itself: "CL does not accept product samples from manufacturers or others for Product Reviews." It is funded by member subscriptions rather than advertising. In its March review of 22 multivitamins, seven failed.

The Pentagon runs Operation Supplement Safety. It is free and open to anyone. Its scorecard walks you through seven questions about a product and tells you to walk away if it scores under four. Granted, it’s not the most robust test, but it’s better than nothing.

Before blindly trusting any tester, check how they make their money. Some take a commission on the products they rank, or even charge a fee for accreditation. As we’ve established, a recommendation that pays the recommender is nothing more than an ad.

What I’d actually buy

Given how quickly names change hands in this market, it’s not easy for me to give you an unqualified recommendation to buy from any particular company. What I can share are the small handful of companies that stood out for good reason during the course of this research.

Only three of the forty-eight supplement brands now in the Ledger got the Approved rating. They all have one thing in common: they maintain strict control over their manufacturing process.

Standard Process grows its own ingredients on a certified organic farm in Palmyra, Wisconsin, and makes the supplements 1.4 miles from the fields. Its organic certifier, not its marketing department, states that at least 75% of the raw plant material starts on that farm. It carries no third-party seal of any kind.

Nature Made owns four plants in California and Alabama, and holds nearly two thirds of the products in USP's verified directory. It belongs to Otsuka, a Japanese pharmaceutical company.

NOW Foods owns its plants in Illinois and Nevada and runs its own analytical labs inside them. In January it became roughly 30% employee-owned. It also buys competitors' products at retail and publishes the results, which is where the vitamin C testing above came from.

You will notice that nothing is rated Avoid in the Ledger for now. To rate a brand Avoid I need a documented act by a named company. This industry makes that standard nearly impossible to uphold. When the FDA shut down seven years of production in Edgewood, the recall named ABH Nature's Products; a company no consumer has ever bought from. I could not find a clean list of the brands affected anywhere.

Every brand named here is on The Brand Ledger. When one changes hands, or quietly drops a verification mark, that is where I'll record it.

In sum

The gummy was a manufacturing innovation. It made a supplement brand cost almost nothing to start, which explains the flood of new products to the aisles.

That deluge resulted in a market a consumer cannot navigate alone. The verification marks and the independent testers help, and for the products that carry them you can at least confirm that what is on the label is in the jar. For the other ninety-nine thousand, you are taking the brand's word for it.

Working out which compounds are worth ingesting in the first place is a different problem, and a story for another time.

Something you owned for years finally gives out. You feel like you got your money’s worth. It served you well. So you go online and buy the same thing again, the same model from the same brand. Only when it arrives, you notice immediately that something is different.

Nothing is broken. No one thing is obviously cheaper. You turn it over in your hands looking for whatever changed but cannot find it. After a while, you start questioning yourself rather than the item. You have gotten older. You are fussier now. Maybe the first one was never as good as you remember.

I have seen some version of that story dozens of times through Worse on Purpose’s tip line. A perception of some reduction in quality, impossible to pin down, followed by expressions of self-doubt.

In ‘On Quality’, I argued that quality exists in the connection between an object and the person interacting with it, rather than as some feature of the object itself. That essay focused on the implications of this view for companies. Firms now measure performance more closely than they ever have. The things that make a product good rarely show up in those numbers, so quality quietly drains away while executives optimize only what they can measure.

This essay is about our relationship to quality as individuals. Companies stopped seeing quality because they narrowed their attention only to what they could measure. What happened to us is different.

Over the past two centuries the things we buy were quietly rebuilt so that most of what determines their quality exists where no buyer can see it. At the same time, we lost the skills that enabled us to judge quality on the parts that still show.

Both losses have a single cause.

On the maker’s side, production withdrew into factories and distant trades. When things were made in the household or the town, the manufacturing processes sat in plain sight. We knew what went into the objects we owned because we saw them being produced. Now the things that determine whether a product will last are completely hidden from view.

Meanwhile, the same forces pushed each of us deeper into narrower fields of work. We do one thing for a living and outsource the rest. But the ability to judge how well a thing is made comes from having made things. The fewer kinds of things we make, the fewer kinds of things we can judge.

What follows is an account of how that happened, and what it cost us.

Inspection through intuition

Hand an object to someone who has spent a lifetime around examples of its kind. A jacket to a tailor, a stool to a carpenter or a knife to a chef. They’ll turn it over once or twice and deliver their verdict. “It’s good” or “not something I’d use.” The whole examination takes only a handful of seconds. Those short moments deserve close attention.

The expert-examiner is performing an inspection. They are holding the object against a standard and determining whether the object meets it. The verdict is entirely intuitive. That is expertise.

But where did the standard that the examiner is applying come from?

It came from having lived with various examples of the object in question for years. It came from a loose familiarity with how such objects should go together. It’s a familiarity that can only be acquired through decades of use.

We are all carrying similar capabilities for judgement. In the domains where we have put our hands to work, we can tell when things are not what they were.

That raises something to settle before going further. If the examiner is detecting something real, where is that thing? Inside the object itself? Or somewhere else?

The competent judge

Suppose quality is a property that exists within an object. If that were true, then quality must be findable. We must be able to create some instrument to detect it. Imagine the finest jacket ever made. Take it to a laboratory and try to design a test for its ‘quality’. You can’t. You can measure the tensile strength of the thread or the abrasion resistance of the fabric, but you cannot design an objective test measuring the overall ‘goodness’ of the jacket.

Suppose instead that quality exists only in the perceiver’s mind. Then it is nothing more than a simple preference. Quality just means whatever you happen to like. This can’t be right either. We agree on quality, even if we can’t express why. We have also all had the experience of buying something on the perception of quality, only to find we were wrong in use. If quality is just a feature of our minds, then the gap between our quality perception and the true quality of the object makes little sense.

So if quality is not a property within an object, nor just a perception in our minds, then what is it?

Robert Pirsig provides, in my view, the most satisfying answer in Zen and the Art of Motorcycle Maintenance. Quality, he says, is not a thing, but an event. Quality arises when a subject meets an object in use. It exists only in the connection between the object and its user. It does not exist independently in the jacket or in you. Quality is what happens when you pick the jacket up, put it on and step into the wind and rain.

This raises a problem: if quality happens in the meeting between an object and its user, then any verdict on the quality of the object will depend on who is delivering it. Different users will reach different verdicts. Something must separate a good judge from a bad one, otherwise we are back to simple preference.

Two centuries before Pirsig, David Hume laid out his criteria for the competent judge:

This is essentially a job description. It describes the preconditions of expertise. Pirsig tells us where quality happens; in the meeting between an object and its user. Hume tells us the meeting of those two things only produces a reliable verdict if the user brings a trained faculty to it, built out of lifelong practice and comparison.

So quality itself has a precondition. Somebody must be skilled enough in their domain to detect it.

A quality nobody can detect is a quality nobody pays for. A quality nobody pays for eventually stops being made. Everything, then, turns on what a buyer can detect.

What we could always see, and what needed policing

There are some properties of an object that you can check yourself, standing in a store, with no special training whatsoever: how the fabric hangs, whether there is any play in the joint, whether a chair will support your weight. These are observable traits.

Others, you have never been able to check at any time, no matter how skilled you were: the purity of the silver, whether the flour has chalk in it, the true provenance of the grapes in the wine. These are concealed traits.

Societies have taken the problem of concealed traits incredibly seriously in the past. They built protective institutions, and ran them for centuries.

English cloth aulnage began in 1197 and did not stop until 1699. Five hundred years where a salaried official’s entire job was to measure every piece of woolen cloth offered for sale, stamp it with a seal if it passed, seize it if it did not, and collect fourpence for the service. The reason is purely practical. Cloth was sold folded. You cannot measure twenty-four yards of folded broadcloth across a market stall, so the state sent a man with a yardstick to do it on your behalf.

Hallmarking began in London in 1300 and is still the law. If you sell something as silver in Britain today, it must carry a mark from an assay office.

The Trial of the Pyx has been testing the coin of the realm since 1282. It tested 7,960 coins in 2024. It is still operational.

Notice the things no institution was created for. Nobody was ever appointed to certify that a chair was sturdy. There is no medieval office of the Sitter. You could just sit on it yourself.

The institutions cluster on the concealed side and leave the observable side entirely alone. This tells you that past societies were not confused about what a buyer could and could not judge at the point of sale. They knew exactly where the line ran, so they spent resources policing it.

The line marks the ratio of observable traits to concealed traits in the marketplace. For most of history it was stable. Most traits were observable, and the few that were not could be policed.

That ratio has since been overturned, from both ends at once. Products grew complex. More materials, more processes, more of their fate decided in places no buyer will ever see. The concealed share of traits within ordinary products exploded. At the same time, we grew narrow. Each of us commands one trade and buys the output of every other, so even the traits that remain visible are harder for us to judge competently.

Most of what determines quality is now concealed. Most of what still shows, we can no longer judge. A quality that cannot be verified cannot be charged for, and what cannot be charged for will cease to exist.

After the paywall: the cause. The concealment of quality and the loss of our ability to judge are two consequences of a single mechanism. What follows names that mechanism, measures it and shows it at work in the production of a simple shirt across four centuries.

Right up top, I must admit that I am a huge car guy. I love old cars. I want to feel connected to the machine. I miss rowing through the gears on a manual transmission. In that sense, modern cars are unequivocally worse than their older counterparts.

So I went into this expecting to write an article bashing the industry about the myriad ways that cars themselves had gotten worse. I figured I’d find that components had gotten flimsier. That vehicles now don’t last as long as they used to. But after looking at the data, that just simply isn’t true. By almost every measure, the new car in your driveway is better than any equivalent in history.

A lot did get worse. Just not the car itself.

The machine underneath is better

First of all, price. Adjusted for inflation, new vehicle prices are down roughly 45% since the early 1980s, according to BLS data. Before you type up a reply asking how that can possibly be true, know that the BLS index tracks a like-for-like car. It simply asks what a 1985 car’s worth of metal and capability costs today. That bundle of components costs a little over half of what it cost in 1985, adjusted for inflation. Cars feel more expensive because it is no longer possible to buy a simple car with limited features. I’ll expand on this idea in a moment.

Cars last longer. In 1970 the average American car on the road was 5.6 years old. By 1999 that number had extended to 8.9 years, according to FHWA tables built from Polk registration data. Today S&P puts the average passenger car at 14.5 years. Americans keep cars roughly two and a half times longer than they did in 1970.

They kill fewer people. In 1980 there were 3.35 deaths per hundred million miles driven. By 1999, 1.57. In 2024, 1.19. Of course, better roads, safety laws and emergency medicine all contributed, so the car is one input among several. But it is a large one.

They use less fuel. EPA puts model year 2024 economy at a record 27.2 mpg, up from about 19.2 twenty years earlier. A 41% improvement, achieved while vehicles got heavier, faster and larger at the same time.

And they get recalled less. NHTSA recalled 31.3 million vehicles in 2025, the lowest total since 2013, down every year since 2020.

Cheaper (with a caveat), longer lived, safer, more efficient and recalled less. The machine itself is better for the average consumer.

What actually got worse

There is a lot to cover here, so I will be publishing four essays on the automotive sector over the coming months:

This first essay covers the transaction. The vehicle itself is now the least profitable part of the automotive business. Dealers now make nearly as much money arranging your financing as they do selling you the car. The way they make it is by marking up the interest rate the lender already quoted them. The average loan has doubled in length since 1971. On top of that, the cheap, no-frills car has been discontinued outright.

The second covers the software. Basic functions of the car that used to be physical parts you owned as soon as you bought the vehicle are now integrated into the software suite. You access these through the infotainment system, which has become the worst-scoring category in the industry's own reliability study. Phone pairing is the single most reported problem in American cars. The same failure-prone infotainment is also a revenue line. Hardware already fitted to your car, switched off until you subscribe. Not to mention the added perk of being able to sell your driving data.

The third covers the supply chain. Cars have always been assembled from other companies' parts; nothing has changed there. The issue is how concentrated the supplier network has become. When a single supplier fits the same component to nineteen manufacturers, a single defect becomes a market-wide event. Takata's airbag inflators did exactly that. Roughly 67 million have been recalled, in 34 brands from Ferrari to Mazda to Tesla. The NHTSA has confirmed 28 related deaths in the US. That is just one of many similar stories.

The fourth covers our right-to-repair (or lack thereof). Service and parts are now the most profitable department in the dealership. This has created an incentive for manufacturers to make repair harder to get outside of their dealer network. A minor collision on a car with driver assistance sensors costs roughly twice what the same crash cost before them. AAA's guidance is that returning those systems to service "requires special training, tools and information." Meanwhile, Ford's CEO, Jim Farley tells investors that parts and service is a $15 billion business growing 8% a year, and that its great virtue is being anti-cyclical: "when the car business goes down, people tend to repair their vehicles." Sure, but Jim… the difference is we used to be able to repair our cars ourselves.

We start with the financialization of the transaction, which is arguably the most societally damaging trend I have covered to date.

Nobody sells the like-for-like car

Back to that 45% price-reduction figure I mentioned in the intro.

The BLS price index measures a constant car. It asks what a fixed bundle of metal and capability costs this year against last year, and adjusts for all of the new features that now come standard with vehicles. Backup cameras have been federally required on every American vehicle since 2018, so the index treats them as a quality improvement and subtracts their cost before comparing. On that basis, yes, cars are cheaper. But they feel more expensive because you can no longer buy the no-frills version.

If you measure instead what people actually pay, against what people actually earn, you can clearly see the decline in affordability. Cox Automotive counts how many weeks of median household income it takes to buy the average new vehicle, financing included. In December 2020 it was 32.2 weeks. This June it was 35.3.

Comerica Bank ran a similar index for decades before Cox launched theirs. It used median family income rather than household income, so the two series should not really be compared directly. But between 2007 and 2012, Comerica's index never once went above about 25 weeks, and it bottomed out at 21.8 weeks in 2009.

However you handle the methodology gap, the direction is not in question. A new car costs materially more of your working life than it did fifteen years ago.

A like-for-like car got cheaper. The car available for sale got much more expensive.

A couple of factors explain what is going on.

The first is what the manufacturers decide to build. In model year 2010, sedans and wagons accounted for 54.5% of American vehicle production. By 2024, they had dropped to 23.7%. Truck-based SUVs went from a fifth of the market to almost half of it over the same period. J.D. Power has trucks and SUVs at 83.1% of retail sales as of January 2026. The industry moved the median buyer out of the cheaper category and into the more expensive one.

The second is what stopped being built. Nissan made the last cheap car in America. The Versa started at $18,585, and production ended in December 2025. Nissan's explanation: "In line with Nissan's product strategy, the Nissan Versa ended production in December 2025 for the U.S. market."

For the 2026 model year, nothing sold in America starts under $20,000. The cheapest new car you can buy is a Hyundai Venue at $22,200, after destination charges.

The industry discontinued the no-frills compact car, and sold you the larger one with more equipment in it. It then went looking for a way to make the payment feel survivable.

The loan doubled in length

The Federal Reserve has tracked the average length of a new car loan since 1971.

In June 1971 the average new car loan ran 35 months. By 1980 it was 44. By 1985, 50.7. By January 2011, when that particular series ended, the average loan was 62.3 months long. The successor series put it at 66.1 months in December 2025.

Today it is about 70.

Two firms measure car loans independently today, using different methods and different data. Experian reads credit files and puts the first quarter of 2026 at 69.48 months and $770 a month. Edmunds reads dealer transactions and puts it at 70.3 months and $773. They agree to within a percent.

Experian also found 35.55% of new loans that quarter had terms longer than six years, up from 30.83% a year earlier.

Edmunds' second quarter 2026 came in higher again. The average buyer financed $44,156 at 7.0% and paid $777 a month, a record for the third consecutive quarter. Over the life of that loan they will pay $9,811 in interest.

This did not happen because borrowing got more expensive. It got much cheaper. A new car loan at a commercial bank cost 13.37% in 1985. The same loan costs 7.47% today.

Hold the loan amount steady and change only the length of the term. Financing $44,156 over the 36 months customary in 1985, at 1985's rate, costs about $1,500 a month and $9,688 in interest. Today's buyer at today’s rate over a 70-month term pays $777 a month, and $9,811 in interest.

The rate fell by nearly half. The interest did not fall at all. The length of the loan ate the entire benefit, and what the buyer got instead was a payment that looks half the size.

The cheapest car on the lot got more expensive, and the loan got long enough to keep the monthly payment roughly where buyers expect it. That is what the extra years are for. Not to save you money, but to move a price you could not afford into a payment you can.

The strain shows in who is signing. Subprime has gone from 14.40% of vehicle financing to 15.75% in a year. Nearly a quarter of buyers now commit to seven years or longer. And a buyer who trades in a car they still owe money on will pay an average of $16,270 in interest.

This is why I say the financialization of the automotive industry may be the most societally damaging trend I have covered to date. A bad frying pan costs you eighty dollars once. Americans now owe $1.685 trillion on their cars. They added $182 billion in new debt last quarter. In most of this country, owning a car is not optional. Demand is inelastic. A car is the thing that gets you to the job that pays for the car. Auto-makers understand this. They know that opting-out is rarely an option.

Where the money is made

GM's 2025 results put GM Financial at $2.8 billion in earnings before tax on $17 billion of revenue. A margin of 16.4%. The same release puts GM as a whole at $12.7 billion on $185 billion. That is only a 6.9% margin for the whole business.

Ford's own release tells the same story, but worse. Ford Credit earned $2.6 billion before tax on $13.8 billion of financing revenue, a margin of 18.8%. Ford as a whole earned $6.8 billion of adjusted EBIT on $187.3 billion of revenue. That is only a 3.6% margin.

Lending is more than twice as profitable as manufacturing at GM, per dollar of revenue. At Ford it is more than five times as profitable. Ford Credit accounted for 38% of Ford's adjusted profit last year.

Some of that gap is Ford having a bad year. It lost $8.2 billion on a GAAP basis in 2025 and its adjusted profit was barely half of GM's. But that is exactly what a finance arm is for. It earns when the car business does not.

The paperwork out-earns the car

Presidio-NCM benchmarks more than 4,000 American franchised dealerships and reports their profit by department. In the first quarter of 2026, selling a new car produced $1,781 in gross profit per unit, down 11.2% from a year earlier.

Arranging the financing and insurance on that same car produced $1,727 per unit, up 7.1%.

Service and parts, meanwhile, generated 53.8% of all dealership gross profit, the highest share in the benchmark's history.

At the biggest dealer group in the country, the story is the same.

AutoNation made $2,769 per unit on finance and insurance in 2025 against $2,564 on the vehicle itself, and by the first half of 2026 the gap had widened to $2,826 against $2,444. Its CEO Michael Manley told investors that "approximately 80% of our profits come from CFS and after sales."

So financing is a bigger share of this business every year, which should strike you as strange. The dealership is not lending you anything. A bank or a credit union is. The dealership introduces you to them.

Which raises the question: what is the dealership finance office actually being paid for?

In 2013 the CFPB and the Justice Department ordered Ally Financial to pay $98 million. Explaining how the arrangement worked, they wrote this:

"Ally sets a risk-based interest rate, or 'buy rate,' and then allows auto dealers to charge a higher interest rate when they finalize the deal with the consumer. This is typically called 'dealer markup.' Ally then shares some or all of the revenue from that increased interest rate with the dealer."

The lender decides your rate based on your credit history. The dealer then decides how many additional points they can get you to agree to. The resulting difference is split between the dealer and the lender.

Dealer profits are falling, by the way. Presidio has average net pretax profit per dealership down 11.2% in the first quarter of 2026 and 11.8% in the second. Selling cars is a hard business and it is getting harder. I have sympathy for that.

What I have less sympathy for is where the shortfall went. A business under pressure is entitled to raise its prices. That is what prices are for, and a dealer who told you plainly that the car costs more this year would be doing nothing wrong at all.

The finance office recovers the money a different way. It seeks only to obfuscate. It arranges the transaction so that you cannot see what you are being charged. The markup on your rate is never quoted, and the products arrive presented as conditions of the sale.

The add-ons and up-sells

At AutoNation, three quarters of cars leave with a finance contract. Michael Manley says buyers there purchase "on average more than 2 products per vehicle with extended service contracts."

Nobody publishes the margin on those. It sits with the administrators, JM&A and Zurich and Assurant, and is not in any filing I could find.

What is public is what happens when regulators take an interest. In November 2023 the CFPB fined Toyota Motor Credit $60 million. Bundled products had added between $700 and $2,500 to individual loans. More than 118,000 calls from customers trying to cancel were routed to a dedicated retention line, where representatives were instructed to make them ask three times before accepting a written request. Refunds, when they came, were applied as principal payments rather than issued as checks.

In December 2025 the FTC mailed $9.6 million to 168,179 people who had bought vehicle service contracts from CarShield. They had paid up to $120 a month for coverage that did not cover their repairs.

The rule that lasted 13 months

In December 2023 the FTC finalized the CARS Rule on a 3-0 vote. It banned bait and switch pricing on vehicles and financing terms. It banned charging for add-ons that provide no benefit. It required dealers to disclose the actual offering price, to say plainly that add-ons are optional, and to obtain express informed consent before charging for one.

The Commission estimated the rule would save Americans more than $3.4 billion a year and about 72 million hours of their time.

Thirteen months later the Fifth Circuit vacated it. They held: "The FTC violated its own regulations when it failed to issue an ANPRM for the CARS Rule."

An ANPRM is an advance notice of proposed rulemaking. It is a procedural step that comes before the proposal that comes before the rule. The court did not vacate the CARS Rule because it was not an effective policy. It vacated because it ruled that the FTC did not file the right document.

Nothing replaced it. The FTC now pursues the same conduct one dealer at a time.

Who pays after month 36

The standard bumper to bumper warranty in America is three years or 36,000 miles. Chevrolet, Ford, Toyota, Honda, Nissan, Subaru, Mazda, Jeep and Chrysler all sit there.

More than a third of new car loans now run past six years, which means the warranty expires with three years of payments still left.

Covering warranty liabilities costs big money, and manufacturers have to set it aside in advance, before they know what will fail. Ford put $6.71 billion into its warranty reserve last year. GM put in $5.07 billion.

Both figures are climbing. Measured against sales, the American industry ran warranty stockpiles at about 1.26% between 2015 and 2019. Ford is now at 4.0% and GM at 3.0%, two-to-three times the old rate.

Some of that can be explained by the price of a repair going up rather than more repairs happening. Some is manufacturers revising estimates on cars they sold years ago. The direction of the trend is the important thing, and this is one area where the claim that the cars are becoming less reliable has some evidence behind it. It also varies enormously by company: Toyota set aside $561 per vehicle in 2024, Stellantis $1,230.

It is money the manufacturer wants back. Kumar Galhotra, Ford's chief operating officer, calls warranty "the largest component of our competitive cost gap" and "a major cost opportunity for us."

There are two ways to reduce a cost like that. Either build cars that break less often, or have them break later, once the warranty window has closed and the bill has become the owner’s.

Whether anyone engineered this transition is a question I won’t answer here. It requires evidence about how cars are actually built. That is the second essay.

What to do about it

Arrange your financing before you go to the dealership. A credit union or bank approval in hand turns your time in the finance office from a negotiation into a comparison. This single step removes the markup described above, because there is no buy rate to mark up.

If you do finance through the dealer, ask what the buy rate is. Ask what rate the lender approved you at, as distinct from the rate on the contract in front of you. You are entitled to know both, and the question itself tells the person across the desk that you know the difference exists.

Judge the loan by total interest, not the monthly payment. Every lengthening of the term I described above was sold as a lower payment, and each one cost more money. $9,811 is the current average (I recommend you aim well below that). Ask for that number, in dollars, before you sign.

Decline the add-ons in writing. Extended service contracts, gap insurance, tire and wheel protection, paint sealant. If you want a service contract, buy it later from someone who is not also selling you the car, and know that you can almost always cancel one for a prorated refund. Toyota Motor Credit built a phone system to make that hard. Cancelling is still your right.

Check the term against the warranty. If the loan outlives the coverage by three or four years, you have agreed to keep paying through a period where every repair is yours. That may still be the right deal, but it should be a decision rather than a surprise.

And never roll negative equity into the next loan.

Above all, if you can, avoid the loan altogether and pay cash. Nowadays, a 10-year-old used car that was well looked after will still be in fantastic shape, with all of the features you could want.

If that is the difference between you having a loan and avoiding one, I’d go used over new.

Where the money went

The car, in many ways, is the best it has ever been. The transaction wrapped around it is the worst it has ever been.

The cars and the production processes around them got so good that the industry stopped making its money on the vehicles themselves. It discontinued the cheap version, moved you into a bigger one, stretched the loan until the payment looked survivable, and took its margin on the lending, the paperwork and the service bay instead.

I still miss the manual gearbox… but that is a complaint for another day.

Next in this series: where the car actually got worse. Cars fail more often now, and less catastrophically. The drivetrain lasts; the screen, the sensors and the electronics do not.

It has never been more difficult to find accurate information about product quality online. A few researchers have explored why that is.

Start with search. Glen Allsopp is an SEO researcher who spent years mapping who actually owns the first page of Google’s search results. His findings were shocking. In his most recent analysis, he found 16 media companies (operating over 580 brands between them) ranking on the first page for 85% of the 10,000 affiliate-heavy searches he tracked. When a content site held the #1 result for a specific product search, it belonged to one of those 16 conglomerates 86% of the time. Collectively, they receive an estimated 3.5 billion clicks from Google a month.

Source: Detailed.com

For the media companies listed above, their ranking on Google is their competitive moat. The “reviews” that you see these companies and their subsidiaries publish are now their core commercial operation.

In a 2023 Digiday & Impact.com survey, 81% of publishers said commerce content plays an important role in their media portfolio. 29% said they churn out 26 or more new commerce pieces a month (more than one every business day), with another 14% publishing at least 16.

Nobody is actually testing products at that pace. There are so many egregious examples of this. For instance, Money.com, a personal finance website, inexplicably publishes buying recommendations for paint sprayers, garage door openers, air purifiers and various other household goods.

At the same time, HouseFresh, an independent outfit that buys and lab-tests every air purifier it covers, watched its Google traffic fall from 4,000 readers a day to 200 after a 2024 algorithm update. A 95% collapse, despite doing the real testing that all the others skip.

The mechanism driving all of this is the affiliate link: a tracking URL that pays the site a commission whenever a reader clicks through and buys. US advertisers passed $10 billion in affiliate marketing spend in 2024, on a forecast track to nearly $16 billion by 2028.

The outcome of this new incentive for publishers is obvious: recommend as many products in as many categories as possible, and collect a commission on every sale that clicks through.

Real product testing is expensive. Instead, volume is now the business. That is how a personal-finance site ends up ranking for paint sprayers and why the "best" lists you read always name ten products instead of one.

Affiliate practices are rarely disclosed too. When Princeton researchers analyzed affiliate content across 500,000 YouTube videos and 2.1 million Pinterest pins, only around one in ten affiliate posts carried any disclosure at all.

The last bastion of truth was supposed to be anecdotal reviews from individual purchasers online, but even those are now compromised. The Transparency Company ran 73 million reviews through fraud detection and flagged nearly 14% as likely fake, estimating the damage at $300 billion a year, or about $2,385 per US household.

Tripadvisor's own transparency report admits that 4.4% of the 30 million reviews submitted in a year are fraudulent. That’s especially notable since Tripadvisor is grading its own efficacy. The FTC considered all of this serious enough to ban fake reviews outright in October 2024.

The problem is caused by perverse incentives.

A recommendation system that pays the recommender will drift toward whatever pays best. Once the tracking link arrived, the drift was inevitable. The “best” product in any given review turned into the best-paying. The reviewer’s customer moved from the reader to the advertiser.

The information ecosystem that surrounds consumer products has been eroded by the same mechanisms that eroded the products themselves. The incentive moved from serving the buyer to extracting from the buyer.

Outside of regulation, there is only one way I see to fix the information ecosystem: publications must commit to a new, entirely resilient incentive structure. They must establish rules that make capture impossible and write them down in public, in such a way that breaking them would cost the publication immediate trust.

So Worse on Purpose is going first. What follows are the rules this publication will run on for as long as it exists, written in public, for readers to enforce.

Goals

Worse on Purpose exists to be the most accurate, up-to-date and readable source on consumer product quality on the internet.

Three core goals follow from that:

Document the decline. When a great product gets hollowed out, prove it. Show who bought the brand, what changed, when, and who profited.

Map what's still built right. Tell the stories of brands that still prioritize their customers over shareholders and deserve your money.

Answer to readers alone. No outside influence. Accurate information survives only where nobody can pay to influence it.

Every policy below exists to protect these three core goals. These are the rules I can be held to.

Commitments & Transparency

Money

Worse on Purpose will never accept money from a brand. Not advertising, not sponsorships, not "partnerships," not paid placements, not press trips. The only way for a publication to avoid any ulterior motives when promoting brands is to commit to never benefiting financially from any promotion. So that is what we’re doing.

The Ledger is not for sale. No brand can pay to get on The Brand Ledger, off it, up it, or down it. No Approved brand has ever paid me a cent, and none ever will. The ledger is also not a blacklist. Every entry carries a last-reviewed date, statuses move in both directions, and a brand that earns its way back up gets written back up.

There are no affiliate links here. The moment a recommendation pays the recommender, it stops being a recommendation. It is just an ad. Any links you see to brand or product pages, in the ledger or newsletter, will in no way benefit me if you purchase through them.

Nothing I cover arrives free. Whenever I test products directly, they get bought at retail, with my own money, the way any consumer would buy them. If I have not tested them directly, you will know. If a brand ships me something unsolicited, it gets returned or donated.

And there are no investors. I have spent every essay in this archive documenting what happens to good things when outside capital buys in. I will not be running that experiment on this one. There will never be assets to sell. Worse on Purpose dies the moment I stop personally publishing it.

Where the work comes from

The work is built on the documented record: SEC filings, court dockets, import and customs data, FDA complaint databases, earnings calls, patent filings. Sources are linked to every claim so you can check me.

I don't do access journalism. I don't ask brands for comment before publishing, I don't take briefings, and I don't attend junkets. There is nothing a company can withhold to obstruct this newsletter, because I have never needed anything from them except their public paper trail.

If a brand (or anyone, really) responds to something I've published with substantive counter-evidence, I'll engage with it seriously and append it to the piece.

The other place the work comes from is you. The best leads so far have arrived in the inbox from readers who watched a product they loved quietly fall apart. If that's you, write to tips@worseonpurpose.com.

One person, on purpose

Worse on Purpose has no staff, and it never will.

Every rule in this policy is about financial incentives that could affect the integrity of the work. A payroll is another form of financial incentive. Yes, employees provide leverage, but that works both ways. Every employee is a real obligation. They are a person you owe, whose rent depends on your revenue. When money gets tight (and money always gets tight) that duty will affect editorial decision-making.

The brand sponsorship deal stops looking like a compromise of values and starts looking like responsibility to your people. The affiliate deal becomes how you make payroll. It may be a more understandable way for a publication to be captured by incentives, but it is capture nonetheless.

Alone, the worst case is that I go unpaid for a few months to protect the integrity of this publication. That's a trade I can make without asking anyone's permission, and one I could never impose on a staff. Because of that, there is no staff.

The catch with working alone is of course the workload, which brings us to the machines.

How I use AI

I use AI in this work. Readers should know exactly where.

One person publishes this newsletter, and its ambitions are lofty: the most accurate, most rigorously fact-checked, most readable reporting on product quality anywhere on the internet. That standard used to take a team of editors, researchers and fact-checkers. With machine intelligence, that is no longer the case.

I use AI to sweep ground no individual could cover alone. Thousands of pages of SEC filings, import manifests, court dockets, patents, decades of forum archives etc...

I also use it in editing in the same way I’d use an intelligent thought-partner. I consult with it about structure, framing and whether there might be a better, more precise way to make a particular point.

I never generate any writing end-to-end with AI. Any time I use AI through the editing process it is to improve the comprehensibility of my original drafts.

Every claim, in every essay, gets verified by me against the primary source before it ships.

The verdicts are mine. The writing is mine. And the mistakes are mine too.

You can always reach me directly via email if you want to talk directly to the human behind the project. I read everything and do my best to reply.

I will get things wrong

A one-person operation working at this scope will make mistakes. Any material errors I make get corrected in the piece itself (whether in The Ledger, or in an investigation), with a dated note. Readers are my best auditors. If you do see any errors, please let me know.

Membership is the model

If you read back through this policy you’ll notice it is essentially a list of money refused. The ads, the sponsorships, the affiliate commissions, the investor's check, and so on...

Rules like that can only hold if an alternative source of revenue can be found.

This work has costs: products bought at retail, records and database fees, the hosting bill behind a few million ledger requests a day, and of course, my livelihood.

The only alternative revenue source that won’t affect the integrity of the work is readers themselves.

So that is the business model. Readers fund the work, and the work stays free. The weekly essays and the Brand Ledger are free for everyone, because if this work matters, it has to reach the person googling "why did my boots fall apart" at midnight, not just the people who pay. There should be no paywall on facts.

When you buy a Worse on Purpose membership, you are funding editorial independence.

The perks below exist as a thank you. The true value is in funding a publication that can investigate any brand, publish any verdict, and correct any mistake, because the only people it answers to are the people it serves.

Membership perks:

The community. The place where readers compare notes on what's still built right, what quietly broke, and what's held up after a decade. These are the notes that feed the ledger and steer the next investigation. It currently lives behind the membership paywall because I'm only one person, and an open community like this would be a recipe for astroturfing. A paid door is our best moderation tool, at least for now.

A member-only investigation, every month. One extra investigation each month, exclusively for members, and members vote on which brand or category gets it.

Votes that set the roadmap. Members vote in the community on what gets investigated next. The research agenda gets decided by the people who fund the work.

A priority inbox. Members' replies land in a priority inbox I can actually clear. You will not disappear into the pile.

And more is coming: member-only features for the Brand Ledger are on the roadmap.

Founding membership is $7 a month or $70 a year, locked for life. On September 1 the price goes to $10 and $100 for new members. Founding members keep their rate forever.

Update, Sep 28, 2026: Founding pricing closed on August 31. Membership is now $10 a month or $100 a year. Founding members keep their rate.

We’ll be back to our usual investigative programming next week.

With gratitude,
Keyana

About one in five of the tips that come into Worse On Purpose’s tip line are about running shoes. I have seen some version of the following statement countless times: “I’ve run in [brand] shoes for ten years, and the last pair I bought fell apart in a couple months”

As an avid runner who has also noticed radical fluctuations in quality between pairs myself, I was excited to dig into this category.

The research revealed a mechanism unlike the ones we typically cover. There is no private equity firm to point a finger at in this story. No debt, no flip, no rollup. It starts with a foam innovation in the mid-1970s that moved the part of the shoe that wears out from the outside, where a cobbler could replace it, to the inside, where it cannot be repaired and often cannot even be seen.

Within a generation, that one change caused an entire industry to abandon control over how its own product is made.

What follows tracks how that happened, what the companies themselves disclose about who builds their shoes, and what it means for you when choosing your next pair.

Shoe Goo

There was a time, long ago, when athletic shoes were repairable. This is evidenced by a product a few of you may know about: Shoe Goo.

Shoe Goo was invented by Lyman Van Vliet, a 45 year old executive at Hughes Aircraft, working out of his garage. He was a tennis player who was “dissatisfied with the durability of the soles of his tennis shoes” and “sought a method to extend their life by repairing them”.

He invented Shoe Goo and founded Eclectic Products to sell it, advertising in tennis magazines and pro shops.

The product found its real market in the rapidly expanding jogging crowd. Joggers were hard on shoes, and Van Vliet had given them a way to fix the part that wore out.

In 1972, the part that wore out was the outsole. It sat on the outside of the shoe where you could watch it start to go and easily complete a DIY repair with a tube of glue. A cobbler could do the same job properly.

In 1974, Brooks introduced the Villanova. It was the first running shoe with an EVA midsole, developed with input from Marty Liquori, an Olympic middle-distance runner and Villanova alumnus. EVA came in as an alternative to heavier, less responsive rubber. It made shoes lighter and softer. Better for running.

The unintended consequence was that it changed how shoes wear out.

EVA does not abrade, it compresses. The failure is structural and it happens in the middle of the shoe, where no glue or cobbler can easily reach it. Worse, it happens invisibly. A dead midsole looks no different to a new one, at least not to a casual runner examining her shoes.

Incidentally, Shoe Goo did fine. It found other markets and is still sold today. I became aware of it during my misspent youth as a skateboarder. Skaters still repair their shoes to this day, long after runners stopped.

Running went the other way. Inside a generation, the category moved from a quality item to a disposable one. For those of us that care about quality and pride on the part of the producer in their work, that transition was stark.

We now replace running shoes on a schedule. We do not argue with the schedule, we do not try to fix them, and for most of us it has never occurred to us that we could.

The schedule is manufactured

Ask a runner when to replace your shoes and you will likely hear a number, rather than a description of condition. Three hundred miles. Five hundred. Somewhere in there. It is one of the most widely known rules in the sport, repeated by shops, coaches, training apps and the brands themselves.

This viewpoint appears to stem from a single study, published in 1985.

Cook, Kester and Brunet, writing in the American Journal of Sports Medicine, put running shoes on a machine and measured how much shock absorption they lost as miles accumulated. They tested at 0, 5, 10, 25, 50, 75, 100, 125, 150, 200, 250, 300 and 500 miles. The shoes held about 75% of their cushioning at 50 miles, about 67% at 100 to 150, and dropped below 60% somewhere between 250 and 500.

Note that five hundred miles is just where the testing stopped, not some magic threshold experimenters discovered. The rule every runner now takes to heart is bracketed, at its upper bound, by how far experimenters decided to run a machine in 1985.

There are two other important findings from that same study did not survive into folklore.

The first is that price and manufacturer had no bearing on how quickly shoes degraded. The cheap ones and the expensive ones died at the same rate.

The second is that when the researchers put the shoes on actual runners instead of the machine, the degradation was "not as great." Not as great is generous. Runners' shoes still held about 70% of their cushioning at 500 miles, a level the machine had them hitting by 150 miles. Real running wore shoes down around three times slower than the test that produced the rule.

Nike's own website is a good example of how brands now apply the number in contradictory and confounding ways. On a page titled "How Many Miles Are Running Shoes Good For?", Nike cites a 2011 review in Footwear Science stating that "high-quality running shoes can last for more than 600 miles."

Just a few beats down on the same page, Nike advises that "experts generally advise that you get new running shoes every 300 to 500 miles."

Then it converts that into a calendar. Casual runners, every 7.5 to 12.5 months. Training for a 5K or 10K, every 5 to 8 months. Half marathon, every 4 to 6 months. Marathon training, every 2 to 3 months.

The same page concedes that there is no set rule.

Clearly, these claims cannot all be true at once. Yet the brands will keep making them, because the replacement cycle benefits them massively.

My advice is to focus less on mileage, and more on your personal experience with the shoe. Examine it, be conscious of how it feels as you run. Base your decision to replace on your personal needs and experience, not a set mileage.

The fitting show

Trusting your own judgement is not easy in a market that sells you the way running brands do, though.

If you have bought running shoes at a specialty store, you know the routine. You run on a treadmill for thirty seconds. Someone films your feet. You watch a slow-motion replay of your ankles while a staff member explains that you overpronate, or underpronate, or are neutral. Then you are walked to the shelf that matches the verdict. Stability. Motion control. Neutral.

It is a good ritual. It feels like being measured for something.

In 2009, three researchers went looking for the evidence behind it. Richards, Magin and Callister, writing in the British Journal of Sports Medicine, searched for any study testing whether prescribing running shoes by pronation type actually reduced injuries.

They found none. Their words: "No original research that met the study criteria was identified." Their conclusion was that "the prescription of this shoe type to distance runners is not evidence-based."

Five years later somebody did the original research. Nielsen and colleagues followed 927 novice runners for a year, all of them in the same neutral shoe regardless of foot type. They published in the same journal, under a title that leaves little room for interpretation: "Foot pronation is not associated with increased injury risk in novice runners wearing a neutral shoe."

People inside the industry have been saying versions of this for years. Spencer White, then a vice president at Saucony, told Outside that "there is no correlation between how much the foot moves and who is getting hurt. It's not a motion thing, it is a stress on the body thing." In the same piece, the physical therapist Jay Dicharry asked whether anyone really believes "an 8 to 13 ounce piece of material is going to stop motion in a 180 to 200 pound skeleton."

The industry agrees, and has been quietly walking away from pronation control for years.

The same Outside article cited above traces Brooks’ removal of the medial post, a firm wedge of foam on the inside of the midsole, which was the entire physical basis of a stability shoe. Brooks started removing it some years before 2020, when the Beast launched without one. By 2021 even the Addiction, its maximum support model, lost its post. The rest of the market followed, replacing posts with guiderails, sidewalls and rocker geometry.

They were right to. The science was never there.

But there was no big announcement. Nobody wrote to the runners who had spent twenty years being told they needed a particular kind of shoe to mention that the reasoning behind it had been quietly dropped.

And the treadmill, along with the bogus gait measurement, is still in the store.

The apparatus of expertise around running shoes is elaborate. The mileage rule, the gait analysis, the categories, the vocabulary of pronation and support and stack height. The brands control their sales and marketing strategy closely.

What they don’t control are the factories producing their shoes. The brand on the box does not own them. It does not employ the people inside it. In several cases it will not commit to buying from it past the current order.

Who actually makes their own shoes?

I wish I could say that last sentence was hyperbolic. It is not. Here’s what the major shoemakers disclose in their filings:

Two brands own factories. The rest rent.

Once a shoe could no longer be fixed and the customer no longer expected longevity from it, owning a factory, and with it control over production quality, stopped being worth the cost. In fact it became a handicap. A brand that owned factories was competing against brands that only owned designs. Within a couple decades, the entire market let go of production control.

And they rent from a small pool. Nike's own filing states that "four footwear contract manufacturers each accounted for greater than 10% of footwear production and in the aggregate accounted for approximately 59% of NIKE Brand footwear production."

Four companies make well over half of all Nike shoes.

One of the largest of those contractors is Yue Yuen, which shipped 252.2 million pairs in 2025 at an average factory price of $21.00 a pair (versus $35-$45 in the US). Its named customers include Asics, Adidas and Nike. Three rivals, all manufacturing in one building.

Of Nike's roughly thirty suppliers, at least 22 are Taiwanese firms, accounting for over 60% of the 550 million pairs Nike sells each year. One industry executive describes Feng Tay, another massive contractor, as "the TSMC of Taiwan's shoe industry."

The variance problem

When a brand does not own the factory, quality is not something it controls. The best it can do is to specify some level of quality, then hope it is achieved. The brand writes the standard. Somebody else, in another country, decides how closely to hold it.

That would be manageable if the arrangements were stable. Often they are not. Deckers, which owns Hoka, tells its investors it buys "on the basis of individual purchase orders, rather than maintaining long-term purchase commitments." In the same 10-K, Deckers also says it "continue[s] to diversify our independent manufacturers and the regions in which they operate."

Think about what that means in practice. The same shoe, sold under the same name at the same price, may be built next year in a different factory, in a different country, by people who have never made it before.

That is the mechanism behind the reader frustration this piece started with. You can’t move production lines searching for fatter margins and expect quality to be maintained.

It is important to note that just because a brand uses contract manufacturers does not necessarily mean that the product is bad. Apple owns no factories. Yue Yuen and Feng Tay are highly capable firms building complex products at enormous scale.

Given that virtually every player in this industry outsources their manufacturing, the simple fact that a brand uses contract manufacturers is not a good indicator of quality. In order to assess likely future quality, we must examine how deep these contractor relationships are, and how they are maintained over time.

What to do about it

As we’ve established, only two major running brands own the factories that make their shoes. If buying from a brand that owns the means of production is a strict criterion for you, I’m afraid your shortlist is very limited:

New Balance is the more available of the two. It runs five company-owned US factories in Maine and Massachusetts employing around 1,300 people, and recently put $65 million into doubling capacity at its Skowhegan plant. It is worth knowing that its Made in USA line carries a disclaimer of "70% or greater" domestic value, and that a class action filed in December 2021 alleged roughly 30% of components in those shoes are made elsewhere. The case was voluntarily dismissed in 2024.

Mizuno is the other, stating in its own sustainability reporting that it "remain[s] committed to manufacturing in Japan" alongside its contracted production.

That is not a silver bullet. It is two brands, and purchasing from neither guarantees a shoe entirely manufactured in-house.

For everyone else with more lenient criteria, the useful question is not whether a brand contracts out, but how it maintains its contractor relationships.

Adidas discloses that it has worked with 65% of its independent manufacturing partners for at least ten years, and 37% for over twenty. That is a company with some skin in the game. If one of their older partners leaves, they lose something truly valuable.

Brooks does not own anything, but it publishes its final-assembly factories by name and location and discloses all four tiers of its supply chain.

Deckers, by contrast, tells investors in its annual filing that it buys "on the basis of individual purchase orders, rather than maintaining long-term purchase commitments," and that it "continue[s] to diversify our independent manufacturers and the regions in which they operate."

None of the companies are hiding this. Their supplier practices are right there in the annual reports.

So, practically:

Judge the shoe, not the odometer. Nike's own page cites research saying more than 600 miles.

Ignore the fitting verdict and buy what feels right on your foot. Two studies, in 2009 and in 2014, found no evidence the prescription prevents injury.

And if you find a model that genuinely works for you, buy a second pair while it exists. Next year's version may be built in a different country by people who have never made it before.

Three brands bucking the trend

A very brief honorable mention for three smaller brands that seem to be pulling in the opposite direction:

NNormal, founded by the mountain runner Kilian Jornet with the Spanish family-owned shoemaker Camper, sells a shoe built around three interchangeable midsoles that you swap yourself. Given that the midsole is the part that dies, this is an interesting approach.

Vivobarefoot runs an actual resoling service in the US, UK and EU through a restoration company called NuShoe. They are upfront that it only covers a handful of models so far. You also have to be ok with running in a barefoot footbed. Not for everyone.

Carson Footwear makes what it calls "American-made running shoes built by hand."

I have not tested any of them. Everything above comes from the companies themselves, and after spending this piece demanding evidence from everyone else, I am not going to recommend shoes on the strength of their own marketing, however cool I think their business models are.

Your mileage may vary.

As always, you can track the status of the brands mentioned in this essay in the ‘Footwear’ section of The Brand Ledger.

In 2011 my grandmother took me to a guitar store for my sixteenth birthday. It was a big day. I had been playing for six or seven years at that point. I was obsessed… and it was time to buy my first ‘real’ guitar.

I spent hours comparing dozens of options while my Nanna patiently read her book in a corner of the store. I had never deliberated over a decision with such intensity. I can still tell you the exact moment the search ended.

A Gibson Les Paul Studio came off the wall and onto my shoulder. Its weight settled into the strap like it was made for my frame. The lacquer was so thick it looked wet. I knew nothing about tone woods or pickup winding at sixteen. I knew this was the one.

Fifteen years later I still have it. It survived almost a decade of sweaty metal shows. Its back carries the belt buckle rash to prove it, a raw patch of bare wood ground through the finish by several hundred nights of playing.

Guitarists will tell you that wood opens up as it ages, that the resonance frees up over decades. I can't prove the physics. I can tell you the guitar sounds better today than it did in the store.

Everyone owns something like this: the old pair of boots, the battered wrenches, grandma’s cast iron pan, the 400,000 mile truck that will not die, etc…

Objects that wear in instead of wearing out. The ones whose quality you can feel the moment you pick them up.

Now try to explain what it is you feel.

Some quality lives in the guitar, the boots and the pan. You are able to detect it intuitively within seconds. But try to put to words precisely what it is you detect and you’ll discover a dilemma.

Your first instinct is to point at the materials. It's mahogany. It has real humbuckers, a set neck, the best nitro lacquer. But that answer collapses the moment you notice the description fits every guitar on that wall. Two instruments leave the same factory in the same month, matching spec for spec down the line, and one sings while the other sounds weak. Every guitarist knows this, which is why nobody buys the model, they buy the individual guitar, only after noodling on it for a while. The spec sheet names everything the guitar is made of, and whatever quality is, it is not on the list of materials.

Fall back on "I just like it" and you sell the knowledge short, because standing in that store I wasn't just voicing an arbitrary preference, I was detecting something. “This is a thing of quality. It is the one I want.” Fifteen years of ownership keep proving my instinct that day right.

That gap, between knowing good on contact and being able to say what good is, is one of the oldest open problems in Western philosophy. It is also the fundamental question this newsletter is attempting to answer.

Every investigation I've published documents the same story: a product stripped of what made it good while everything a shopper can check stayed intact. Same logo, same spec sheet, same four and a half stars, same price or higher. What got swapped out, the steel gauge, the stitch count and the years of service life, sits in the parts you can't check from the aisle.

That gap between what you can verify and what actually matters is the whole game, for them and for you. They use it to slowly diminish the quality of the products you once loved without tripping an alarm. You can use it to watch the theft happen.

So this essay is about what quality is and where it actually lives, which makes it the philosophical foundation under every investigation this publication will run.

In the late 1950s, a rhetoric instructor at Montana State College named Robert Pirsig noticed that his contract required him to teach "quality". He asked around the faculty at the university and discovered that nobody could tell him what the word meant.

Teachers had been passing and failing students on the basis of quality for centuries without a definition. The problem ate at him for fifteen years, ultimately driving him to insanity. Pirsig’s best thinking on the nature of quality was eventually set down in Zen and the Art of Motorcycle Maintenance. It stands alone as the most formative book I have ever read.

To illustrate the problem, Pirsig relays an experiment he ran with his students. He read four student papers aloud and had everyone rank them by quality on slips of paper. He ranked them himself, collected the slips, tallied the results on the blackboard, and set his own ranking next to the class average. His ranking and the students’ matched almost every time, across classes and semesters. A room full of undergraduates who could not define quality independently agreed on where it lived and where it didn't.

So Pirsig landed on this problem statement:

"Quality is a characteristic of thought and statement that is recognized by a nonthinking process. Because definitions are a product of rigid, formal thinking, quality cannot be defined."

Then he added the sentence that should be nailed above the door of every product team in America.

“Even though Quality cannot be defined, you know what Quality is.”

When colleagues demanded proof that an undefinable thing existed at all, he offered this subtraction:

Pull quality out of the world and street noise ranks with symphonies, slop ranks with dinner, and no made thing is worth choosing over any other. A world without quality would still function. You just wouldn’t want to live in it.

Similar threads have appeared across disciplines.

The architect Christopher Alexander encountered the same problem from a different angle. A career spent asking why some buildings feel alive and others feel dead ended in the same non-definition, a quality that in his words "is objective and precise, but cannot be named."

What Alexander did about it is the useful part. If the thing itself could not be written down, the places where it reliably appears could be, so he and his collaborators catalogued 253 of them, pulled from centuries of buildings people love, and handed architects and builders a working method for producing quality in buildings and towns.

That is what every craft tradition is: a transmission system for the unnameable, carried in people rather than paper. It is why apprenticeship survives every technology invented to replace it: the judgment that produces quality transfers only by demonstration and correction, through thousands of supervised repetitions across years, from a person who has it to a person who does not yet.

That fact cuts both ways. The judgment that can only be carried in people is also the one thing no spreadsheet can measure.

Whatever resists definition resists measurement. Whatever resists measurement vanishes from the dashboard. And in a company run from dashboards, what vanishes from the dashboard vanishes altogether.

That blindness is not an accident of modern business. It was designed, it has an inventor, and the tragedy is that it was invented to do the opposite job. During World War II, a General Electric engineer named Lawrence Miles was tasked with scaling turbo-supercharger production for B-24 bombers from 50 a week to 1,000 while steel, copper and nickel were rationed to hell. He hunted substitute materials, and noticed the surprising fact that substitutes often made the part cheaper and better at the same time. In 1947 he formalized the method and called it value analysis. Identify the function a part serves, then find the best possible way to serve that function. Function first, cost second. Through this process, Miles built a machine for producing quality cheaper.

The method worked so well it quickly became universal. The Navy adopted it in the 1950s and renamed it value engineering. The Pentagon eventually wrote it into federal procurement rules. Miles's 1961 handbook was translated into a dozen languages, and within a generation nearly every large manufacturer ran a version of the program. Continuous, itemized, never-ending review of everything a product is made of, scored in dollars saved. That framework is now as ordinary as accounting.

Unfortunately, his descendants now run the machine backward. The modern cost-down program starts from the spec sheet and asks what can be removed without a statistically significant change in buyer perception this quarter. The thinner steel passes the test. The glued joint passes. The plastic gear where the brass one was, the foam that loses a third of its resilience in two years etc…

Each change is approved in isolation, and each is too small for any buyer to notice on its own. That is the trick. No test compares the product to what it was ten years ago. Each product version is measured against last quarter's, comes back as "no detectable difference," and ships. So the degradation compounds beneath the threshold of every individual measurement, invisible at each step and enormous in total.

It is by this mechanism that quality, the unmeasurable property, erodes as a company places a greater insistence on measurement.

When we ask what quality actually is, only two answers exist. Either quality is objective, meaning a physical property located in the object itself. Or quality is subjective, meaning an opinion located in the person judging.

Take the first answer seriously. If quality is a physical property of the object, then instruments should detect it. We can measure a guitar's weight, its neck relief, its fret height, its finish thickness down to the micron. No instrument has ever measured whether it is good. If quality sat inside the object the way mass sits inside the object, quality control would be a solved engineering problem and a factory could certify goodness the same way it certifies tolerances.

Now take the second answer seriously. If quality is only an opinion, then a quality judgment reports a fact about the judge and no fact about the object. Two things follow. First, quality judgments should distribute evenly, because nothing in the object itself would constrain them. Second, no quality judgment could ever be right or wrong, because there would be nothing objective in the world for it to be right about.

Ultimately, both answers fail.

Pirsig’s blackboard example demonstrates the failure of the first: his classes converged on the same rankings, semester after semester, with no criteria handed to them. They agreed independently on some notion of quality, recognized it, but could not explain it.

The second fails on an experience everyone reading this has had: being wrong about quality. Nobody has ever been wrong about liking vanilla, because a preference claims nothing factual about the world. A quality judgment claims plenty. When I decided that Les Paul was good, I was predicting that the neck would stay straight, the frets would survive the abuse, and the guitar would still be worth reaching for in fifteen years. Every one of those predictions could have failed, and with other guitars, for other players, they have. Boots that looked right have come apart in one winter, and every buyer of a bad pair has said the sentence that pure subjectivism cannot explain: “I was wrong about those boots.”

So if quality is not an objective property, since no instruments can detect it, and it is not just a subjective opinion because we are able to make predictions about the quality of an object that are verified in time, then what is it?

Pirsig's answer was that the question itself smuggles in the false assumption that quality must be located in one place or the other. He argued instead that quality is a feature of the relationship between the person and the object. Quality occurs when a person and a thing meet in use: the weight settling onto the shoulder, the wrench loaded to its limit and holding. Before they meet there is only a guitar and a player. Quality exists in the connection.

Apply this idea in the realm of consumer products, and the slow decline of quality starts making sense.

Every measurement a company takes lands on one side of the objective/subjective divide or the other. Spec sheets, tolerances, and materials testing measure the object alone. Surveys, star ratings, and focus groups measure opinions alone. And they sample the opinion at the wrong moment. A star rating gets filed in the first week of ownership, while the surface still shines, and the failures arrive in year three, when almost nobody returns to amend it.

Nothing measures the relationship, because the relationship only exists in use, in the hand, on the road and across years. So when quality drains out of a product, it drains from the one place no instrument points at. Both sets of numbers can hold perfectly steady while the thing between them disappears. That is how a product gets worse without a single metric moving. That is also how the people doing it stay convinced that nothing was lost.

What produced quality in the first place was care. Pirsig again: "Care and Quality are internal and external aspects of the same thing. A person who sees Quality and feels it as he works is a person who cares."

A good object is a fossil record of care. Thousands of small selections made by people who could tell the good facts from the bad ones and picked the good, even when not doing so was cheaper or easier. The extra ounce of brass, the second coat of lacquer and the tolerance held a hair tighter than the drawing demanded. Every one of those selections survives for exactly one reason, which is that somebody with power over the object gave a shit.

That is why extraction works the way it does. When outside capital buys a great brand, nobody issues a decree that the product shall now be garbage. Instead, the people who previously made those thousands of selections either get laid off, retired out, or reorged into irrelevance, and the decisions migrate to a floor in an office building where nobody has ever actually used the product. Care cannot be exercised at that distance. Quality follows care out of the building, and the whole thing happens without any individual ever choosing badness directly. Absence does the job on its own.

There is this scene in Pirsig's book that I often reflect on. His riding partner John owns a new BMW motorcycle, the best machine you could buy in 1968, but its handlebars keep slipping. Pirsig offers to fix them with shims cut from an aluminum beer can. Soft, rust-proof, precisely the right thickness, and in his words the best shim stock in the world. John is horrified. He would rather ride a malfunctioning bike with loose handlebars than let a beer can touch it. He cannot see past the ugly surface to the perfect function underneath.

That was 1968's confusion: good things that looked cheap. We industrialized the reverse. The heritage fonts, the small-batch copy and the unboxing moment engineered to simulate the exact feeling my Les Paul earned honestly. The surface has never been more perfected, wrapped around the glued joints and the plastic gears.

Fifty years ago a man refused a great repair because it looked like trash. Today we pay premiums for trash because it looks like greatness. Same confusion, now commoditized at scale.

There are two ways to make money on quality. You can make a thing so good that people pay for it, keep it, and hand your name to their kids. That is the Miles road: quality found cheaper, profit as the receipt for care. Or you can buy the name after the caring is done, spend down four generations of accumulated trust, and be gone before anyone's memory catches up.

Pirsig opened Zen with a line carried down from Plato's Phaedrus, which makes the question in this essay at least twenty-four centuries old:

"And what is good, Phaedrus, and what is not good. Need we ask anyone to tell us these things?"

A couple of years ago, the small All-Clad pan I fry eggs in cut me as I was pulling it out of the dishwasher. A clean slice on my index finger off the rim. I patched it up. Didn’t think much else of it. I discovered what actually happened while researching this essay, sitting in a settlement archive three years too late to file a claim.

All-Clad had been selling its American-made stainless pans as dishwasher safe. In the dishwasher, the bonded rims corroded until the exposed edge got sharp enough to easily slice through skin. The company settled in 2023, capped at $4 million for a class covering seven and a half years of national sales, a rounding error for its French owner.

Anyone who bought a D3, D5, or LTD pan between January 2015 and July 2022 was in the class. If your pan was damaged, the deal was actually fine: a replacement plus $75, or, my favorite option on the menu, a trade of your damaged American-made pan for a set of nonstick ones made in China.

What never got fixed was the pan. Nothing was re-engineered to survive a dishwasher; the words "dishwasher safe" just came off the box, and the product pages recommend handwashing now. The claims window closed in April 2023 with most of the class, me included, never knowing it existed. Nobody mailed a postcard about the metallurgy.

I still use the pan. It's a good pan. But it turns out the cut on my finger had a paper trail, and so does nearly everything else in the cabinet. The cookware aisle runs on old names and new owners. With all of the mergers and asset handoffs, mapping who owns what in cookware was the most challenging of all the markets I’ve covered so far. Here’s the summary.

Six companies, eighty-one brands

Centre Lane Partners, a private equity firm, owns Pyrex, Corelle, CorningWare, Snapware, Visions, Chicago Cutlery, Instant Pot, Anchor Hocking, Lenox, Oneida, and Reed & Barton. That is almost all of the surviving legacy American tabletop industry under one roof, and we'll come back to it, because that story is a doozy.

Groupe SEB, the French conglomerate behind T-fal, bought All-Clad in 2004, WMF in 2016, Lagostina, the Emeril license, and, as of January 2025, de Buyer, the French carbon steel maker everyone recommends as the antidote to enshittification (let’s see what their new owners do with that reputation). SEB announced up to 2,100 job cuts this February, with German outlets reporting three plants on the block.

Meyer, a family-owned manufacturing giant out of Vallejo and Thailand, makes over 120,000 pans a day that almost never say Meyer on them: Farberware cookware, KitchenAid cookware, Rachael Ray, Circulon, Anolon. The KitchenAid pots at Target have the mixer company's name and Meyer's Thai steel.

Newell Brands bought Calphalon in 1998, promised nothing would change. It moved production to China, and closed the last American Calphalon plant outside Toledo at the end of 2023, 130 jobs. Newell is the company that did the same to Mirro in Manitowoc two decades earlier, and this April Manitowoc sued Newell for $6 million to clean up the mess it left.

American Securities, another private equity firm, bought Conair in 2021, and Cuisinart with it. Cuisinart's food processors were already the subject of the largest kitchen appliance recall in CPSC history, 8 million units whose aging blades cracked and shed metal into food, with 69 reports of fragments and 30 injured mouths.

Lifetime Brands owns or rents nearly everything else with a dead founder's name on it: Farberware kitchenware and knives, Mikasa, Pfaltzgraff, the American license for Sabatier. It manufactures almost nothing. Its one growth business in a money-losing 2025 was Dolly Parton housewares, up 150 percent.

The problem with buying from any of the above is the incentive underneath the label. When a brand changes hands this many times, the name becomes the product and the pan becomes a cost to be managed beneath it: thinner steel here, an offshored handle there, a warranty rewritten by lawyers instead of engineers, and so on…

Nobody has run this playbook with more gusto than Centre Lane. Run it all the way out and the ending looks like this: one firm holding most of the surviving legacy American tableware names, a state monopoly case grinding through discovery, and the plant that made Pyrex for 132 years standing empty in western Pennsylvania.

132 years, gone in an instant

Pyrex was materials science before it was a brand. Corning invented it in 1915: borosilicate glass that shrugged off thermal shock. It is the reason three generations learned to take a dish from the freezer to the oven without a second thought.

One important aside about the product before the money story starts. The American Pyrex on shelves today is tempered soda-lime glass, a different material with a thermal safety margin an American Ceramic Society analysis described as borderline, roughly 55°C of headroom against 183°C for borosilicate. Consumer Reports collected 163 shattering incident reports, 42 of them involving injuries. European Pyrex, made by an unrelated company, is still borosilicate. The switch happened under Corning itself, decades before the buyouts. That is exactly what made the brand such a perfect acquisition: the glass had already been downgraded, the reputation had not.

In 1998 Corning sold the consumer business to an affiliate of Borden, then a KKR portfolio company, in a deal worth about $603 million. The new company borrowed $471.6 million on day one and paid Corning a $472.6 million dividend, which means it was born owing more than it had. It renamed itself World Kitchen, bought more brands with more debt, and went bankrupt by 2002. To dig out, it sold the one thing in the portfolio still growing, OXO, for $273 million.

Cornell Capital, a private equity firm founded by a former Goldman Sachs vice chairman, bought the company in 2017 and renamed it Corelle Brands. In 2019 it merged with the Canadian company behind the Instant Pot, valuing the combination at $615 million. What it actually bought is disputed to this day. More than 98 percent of the price was booked as goodwill and intangibles, and within days of closing a Cornell partner on the new board emailed the CEO to ask, "Is there fraud going on???" 

The suspicion, later spelled out in court filings, was that Instant Pot's sales had been pumped up with discounts and channel stuffing before the sale. The CEO's own verdict was that he was running "the most poorly operated 'successful' business on the planet."

Cornell held on to the asset anyway, and two years later, it got its money out. On April 12, 2021, Instant Brands borrowed $450 million. On April 21, it paid a $345 million dividend, roughly $200 million of that to Cornell Capital and its co-investors. The solvency memo blessing the payout assumed new product sales would grow more than 900 percent. In the end, actual sales missed that year's projections by $157 million. In June 2023 the company filed for Chapter 11, and the lenders that were collectively owed $391 million recovered between seven and nine cents on the dollar.

The bankruptcy is how we know all of this. The court appointed a litigation trustee to claw money back for the people left holding the bag. In November 2024 he sued Cornell Capital and its founder for more than $400 million, laying out the fraud email, the solvency memo, and the dividend math in an 88-page complaint. Cornell calls the suit baseless, and the case is grinding through a New York bankruptcy court right now.

Centre Lane bought the wreckage out of the bankruptcy and folded it into Anchor Hocking, the Ohio glassmaker it already owned. Two names matter from here. Anchor Hocking is the glassmaker: the plant in Lancaster, the name on the bakeware. Centre Lane is the owner: the New York firm that decides what happens to it.

Within six months of the purchase, Anchor Hocking announced that Pyrex production would leave Charleroi, Pennsylvania, the plant that had made it since 1893, and consolidate in Lancaster. Roughly 300 people worked at Charleroi.

Pennsylvania fought the closure and lost. Senator Bob Casey alleged the purchase had been structured to duck antitrust review: two-thirds of Corelle bought for $38.5 million, flipped into Anchor Hocking days later at $79.8 million. The state's attorney general sued to keep the plant open, alleging the combined company controls more than 91 percent of American glass bakeware. The judge declined to block the shutdown, and the monopoly case is still in discovery while the thing it was meant to prevent has already happened.

When offered relocation to the Ohio plant, four of the 300 said yes. The last piece of glass came off 114 kiln on April 11, 2025, and the plant whistle blew for 132 seconds, one for each year it was operational. A French industrial glassmaker tried to buy the empty plant last summer and reportedly couldn't get it past the FTC. Lawmakers say the FTC wants want consumer glass made there or nothing.

Tony Payne, an engineer who spent 36 years inside the plant, told the Mon Valley Independent that "fifty percent of Pyrex is now being made outside of Anchor because they can't make it." That claim is his alone, but there is precedent for what happens to the label when Pyrex gets made elsewhere: in January 2023, under the previous owner, the FTC made the company pay for stamping Made in USA on Pyrex measuring cups that came from China. Even the lawyers had to chase the owner. Jones Day sued Centre Lane for $9.6 million in unpaid legal bills this February, naming Anchor Hocking and Corelle Brands among the co-defendants. They settled confidentially last week.

Anchor Hocking itself has run this gauntlet several times: a hostile takeover by Newell in 1987, Cerberus into bankruptcy by 2006, then Monomoy Capital, which bought it for about $75 million, borrowed against it and charged it fees, before the whole thing went bankrupt again in 2015. Brian Alexander wrote a fantastic book about what two decades of financial engineering did to Lancaster, Ohio, called Glass House. His summary of the Monomoy years holds for the entire industry: they "put a little money in and pulled a wagonload of money out."

Under Centre Lane's ownership, the extraction tactics found their final form. In 2023 Anchor Hocking sold its own factory, the building responsible for nearly all of its production, and leased it back for 25 years. Three different owners have now monetized the same plant. The people of Lancaster still work in it. They just don't own any of it, and neither, anymore, does their employer.

Names for rent

Lancaster at least still has its factory. The purer version of the same move is to close the factory outright and keep nothing but the name. The cleanest example of that strategy is Farberware.

S.W. Farber's plant in the Bronx made Farberware pans from 1900 until 1996, when two buyers carved up the company, shut the plant for good, and moved production to Indonesia. Because those buyers filed their paperwork with the SEC, we know exactly what a beloved American brand is worth once the factory attached to it is gone.

Meyer paid $25.5 million, once, for an exclusive worldwide license to put the Farberware name on cookware for a term of 200 years. Not a typo. The lease runs to the year 2196. Nobody alive will see it expire, and nobody involved will ever make a Farberware pan in the Bronx again.

Once you understand the brand licensing structure, you see it everywhere, and the pattern is always the same: one company owns the name, a different company makes the product, and the factory that earned the name is liquidated. Chicago Cutlery began as a knife shop in Chicago and later made its knives in Wauconda, Illinois. Centre Lane owns the name today, and every knife under it is made in China. Sabatier, the storied French knife name, was never one company to begin with, and today more than 30 unrelated firms have rights to it. The Sabatier on American shelves is rented by Lifetime Brands and made in China. Revere Ware rode along in the same Corning spinoff as Pyrex, lost its Illinois plant in 1999, and was discontinued so quietly that the enthusiasts who sell replacement parts had to break the news. Pfaltzgraff made stoneware in York County, Pennsylvania from the early 1800s on. Lifetime Brands bought the brand for $32.5 million in 2005 and deliberately left the pottery out of the deal. Roughly 250 workers got 60-day notices, and the name went to Asia without them.

The Stanley on the viral tumbler and the Stanley on the tape measure are different companies that have shared a name for decades under negotiated truces. The tumbler side belongs to PMI, a Seattle firm that bought the drinkware brand in 2002. TikTok took its Quencher from $73 million to $750 million in annual sales in four years. In December 2024 the CPSC recalled 2.6 million of its travel mugs after lids came off hot and burned 38 people. Two months later, Stanley Black & Decker sued PMI, arguing the new "Stanley 1913" branding breaks the old agreements that kept the two Stanleys out of each other's lanes. The name is the asset worth suing over. The product is the thing that got recalled.

Tupperware ends the same way. It went bankrupt in 2024 and now belongs to a lender group that includes Alden Global Capital, the fund best known for strip-mining American newspapers. Its last American plant, in Hemingway, South Carolina, closed within months, taking 148 jobs. Production moved to Mexico.

Worse from birth

The old brands took decades to get worse. The new ones started worse and spent the difference on ads. Four brands define the tier: Caraway, Our Place, HexClad, and Great Jones. All of them launched in the late 2010s, all of them manufacture in China, and all of them built their businesses on marketing spend rather than materials.

Caraway and Our Place sell the pastel pans: aluminum bodies under a ceramic "sol-gel" nonstick coating, pitched as the healthy alternative to Teflon. The materials science on that coating is peer-reviewed and brutal. A 2025 study ran five commercial ceramic coatings against ordinary PTFE across 90 simulated cooking cycles, and PTFE held its release performance the whole way while the best ceramics quit somewhere between cycle 30 and 60. Cook eggs daily and that is a pan with a working life that is measured in months. Both companies put that lifespan in the paperwork. Caraway covers its coating for one year. Our Place covered its coating for one year too, until May 2025, when it extended coverage on new purchases to three years after complaints about early coating failure had piled up in public. Let’s see how those claims end up being serviced.

Since the pans can't win on durability, Caraway sells fear of the alternative, and the consequences of that strategy have gotten expensive. Last August the ad industry's self-regulator ordered it to drop claims that ordinary nonstick releases toxins into your food, along with an unsupported claim about a "rise in Teflon flu." Caraway kept pushing, and this February the owners of All-Clad and Farberware sued it in federal court over the toxicity marketing. When the conglomerates from the first half of this essay are the ones calling your health claims false, you have accomplished something notable.

HexClad, the Gordon Ramsay brand, ran the fear play with a different surface and got caught more directly. Its "hybrid" pans were marketed as non-toxic and PFAS-free while containing PTFE, which is a PFAS. The class action settled this February: 209,712 people filed claims and got about six dollars each. Judged purely as a pan, America's Test Kitchen found the laser-etched surface failed at releasing an egg. Durable enough to survive steel wool, unable to do the one job nonstick exists for.

Great Jones compressed the whole VC-funded DTC extraction formula into four years: media-darling launch in 2019, a cofounder war that saw all six full-time employees quit by late 2020, and a quiet 2023 absorption by Meyer, the licensing giant from earlier in this story. The $160 Dutchess dutch oven is made in China with a chipping problem the reviews documented from early on, an issue I can attest to personally.

One DTC brand deserves the exception flag. Made In never raised the money that ruins these companies. Its documented funding totals about $8.3 million, against roughly $70 million at Caraway and the $100 million Gordon Ramsay's studio put behind HexClad. The founders still run it. A company that can't afford to buy its customers has to earn them, so the budget went into the product: family-owned partner factories in France and Italy plus American mills, a win in America's Test Kitchen's stainless rankings, and pans in actual restaurants. Its nonstick warranty even admits the coatings are "consumable by nature", which is perhaps the most honest sentence published by any cookware company this decade. Buy their stainless and carbon steel and treat every nonstick pan from anyone, at any price, as a consumable.

What to buy instead

The counter-list in this category runs deeper than any I've covered previously, because cookware is simple enough that stubborn, family-run businesses can still compete. The pattern in who survived is uniform: nobody outside the family has ever been owed a return.

Cast iron

• Lodge: Owned by about 40 members of the founding families, two foundries in South Pittsburg, Tennessee since 1896, and still under $30 for a skillet your grandchildren will likely inherit.

• Smithey: Founder-owned in Charleston, South Carolina. Smooth-finish heirloom pieces at heirloom prices.

• Field Company: Brothers-owned, cast at US foundries. The lightweight, vintage-style pan.

• Stargazer: Founder-owned in Allentown, Pennsylvania. Machined cooking surface, honest price.

• Lancaster Cast Iron: Founder-owned in Conestoga, Pennsylvania with a fully domestic supply chain.

• Borough Furnace: A husband-and-wife shop in Owego, New York that pours its own recycled iron in-house, which almost nobody in this country still does.

Enameled cast iron

• Le Creuset: Privately held, and the cast iron still comes out of the same French foundry it has since 1925. The stoneware and accessories come from Asia, so buy the iron.

• Staub: Still made in Merville, France under Zwilling, the German family conglomerate that has so far acquired without gutting. One caveat, courtesy of reader Tracy: that applies to the enameled cast iron only. Staub's ceramic and stoneware lines (including the mini cocottes styled to look exactly like the cast iron ones) are made in China, along with the bakeware and accessories.

Stainless and carbon steel

• All-Clad: The bonded lines are still made in Canonsburg, Pennsylvania, sharp-rim settlement and French owner notwithstanding.

• Heritage Steel: Family-owned in Clarksville, Tennessee, building on old Vollrath tooling. The fully American clad alternative.

• Tramontina: Family-controlled in Brazil. The honest value play; their clad line embarrasses pans at twice the price.

• de Buyer: The carbon steel standard, still made in the Vosges. It belongs to Groupe SEB as of January 2025, so it’s one to watch over the next few years.

• Matfer Bourgeat: The French commercial standard, family-held, and what a lot of commercial kitchens use.

Knives

• Victorinox: 90 percent owned by a foundation built to prevent exactly the kind of sale this essay is about. The Fibrox chef's knife is the best $45 in any kitchen. All my knives are Victorinox - I plan to cover their story in an independent article.

• Wüsthof: Seventh-generation family ownership, forged in Solingen.

• Dexter-Russell: The knife in every American commercial kitchen, made in Southbridge, Massachusetts since 1818, still family-held.

Bakeware and glass

• Nordic Ware: The family that invented the Bundt pan still makes it in Minneapolis.

• USA Pan: The retail arm of the Bundy family's commercial bakery-pan operation, which equips actual bakeries.

On glass bakeware, the unfortunate truth is that both American options are soda-lime. If you want borosilicate, buy European Pyrex online or thrift the old stuff.

Dinnerware and flatware

• Corelle: Still made in Corning, New York and still the best cheap dinnerware going. The risk is the owner, not the product.

• Fiesta: The last big American dinnerware factory, still family-owned in Newell, West Virginia.

• Liberty Tabletop: America's only flatware, made in the very Sherrill, New York plant Oneida walked away from, bought by two former Oneida executives.

Special mentions

• Vitamix: Barnard family since 1921, Ohio-made, repairable. The appliance that gets willed to people.

• All American 1930: The pressure canner, cast since 1930 by a family foundry in Manitowoc, Wisconsin, the same town Newell abandoned and got sued by. The town that lost Mirro still makes the best canner on earth. Some receipts write themselves.

Every brand named in this essay, and 89 in total from this research, now has an entry in the kitchen section of The Brand Ledger: who owns it, where it's made, the verdict, and what would change my mind. When an owner changes or a verdict moves, subscribers hear first.

Everything above reduces to four rules you can run in about two minutes, standing in the aisle:

• Look up the owner, not the name. The name is the cheapest part of the company to preserve. If the ownership trail ends at a private equity firm or a licensing group, assume the product and the reputation are no longer aligned.

• Read the warranty before the reviews. Its length is the company's own estimate of how long the product lasts, and its exclusions are a list of the ways they expect it to fail.

• Buy bare metal where you can. Cast iron, carbon steel, plain stainless. No coating means nothing that dies at cycle 40 and nothing to rebuy in eighteen months.

• Distrust a name that's on everything. When the same brand appears on your pan, your knives, your toaster, and a box fan, nobody who owns it makes anything.

On June 11, in the second half of the opening match of the 2026 World Cup, the referee raised his hand and stopped play for a hydration break. Fox cut to commercials. The ad-break ended late. American viewers missed ten seconds of live football in the first game of the tournament. FIFA reviewed the incident and penalized nobody.

I watched it from my sofa. I grew up in the UK but live in America now. The tournament has finally come to the country where I live, and I am watching it on TV anyway, because attending priced out somewhere between a used car and a downpayment on a new house. More on that below.

The water break deserves the first word because it is the cleanest example of FIFA’s extractive influence on the game: a player-safety measure converted into ad inventory.

Cooling breaks entered football at Brazil 2014 as a safety measure, triggered only when the heat index crossed a danger threshold. Last December FIFA made them mandatory in all 104 matches of this tournament. Its tournament chief was explicit: "no matter where the games are played, no matter if there's a roof." That covers the air-conditioned dome in Arlington. It covered a cool, wet evening in Boston where England played Ghana.

In March, FIFA cleared broadcasters to run advertising inside the breaks. Inside the stadiums, the pause is announced as the Powerade Hydration Break, with branded coolers rushed onto the touchline. Call it two minutes of sellable air per break, two breaks per match, and 104 matches: nearly seven hours of new commercial time carved out of live football, a sport that had gone a century and a half without an ad interrupting play.

Ten quid at Loftus Road

I grew up in West London going to QPR matches. Ten pounds got me into Loftus Road, fifteen on a big occasion, and that pricing held right through my teenage years. Football was the cheap thing you did with your mates, and the World Cup sat above it as the free festival, the one month every four years when everyone in the country was collectively enthralled by the action.

Hold that image against the bid book. When the United States, Canada and Mexico won hosting rights, the official bid promised group-stage tickets from $21 and a final topping out at $1,550. In April, FIFA released its own direct-sale final tickets at $10,990. That figure was FIFA's face value, sold by FIFA. (On FIFA's resale platform, one optimist listed four seats behind the goal at $2,299,998.85 apiece, with FIFA positioned to take 15 percent from seller and buyer alike.)

The $60 "Supporter Entry Tier" announced after December's backlash turned out to mean "hundreds, rather than thousands" of seats in stadiums holding 80,000, per the Associated Press. Attorneys general in four states are now investigating, and New Jersey's described the buying process as "a gauntlet of confusion, fake scarcity, and impossibly high prices."

The pricing algorithm punished loyalty with precision. Between October and April, FIFA raised prices on more than 90 of the 104 matches, up an average of 34 percent by The Athletic's count, and then cut them roughly 37 percent in the final ten days when real demand arrived. The supporter who bought early because they cared paid the top of the curve. Football Supporters Europe calculated that following your team from the group stage to the final costs about five times what the same journey cost at Qatar 2022, and filed a complaint with the European Commission over it.

The England Supporters Travel Club's cheapest final ticket this year is £3,119. The equivalent seat at the Euro 2024 final in Berlin cost £83. UEFA has already capped Euro 2028 so that nearly half of all tickets come in under £60.

FIFA's extortionate prices are a choice.

The only World Cup on Earth

FIFA sells the only World Cup there is. No substitute exists. A normal business that treats its buyers like this loses them to a competitor. FIFA has arranged a world with no competitor to lose them to.

Last week I catalogued the mechanisms of enshittification. FIFA runs most of the list simultaneously, from inside a Swiss non-profit that pays no federal tax on tournament revenue and demands blanket tax exemptions from host nations as a condition of bidding. Mexico granted a 100 percent income-tax exemption for this tournament.

The strategy shows up in FIFA's own accounts. Revenue per four-year cycle: $4.19 billion in the 2010 cycle, $7.57 billion in the Qatar cycle, roughly $13 billion budgeted for this one, with reserves that peaked near $4 billion after Qatar.

FIFA also brought its hospitality business onto its own books this cycle and budgeted it to more than triple, from $949 million to $3.1 billion, with roughly a million tickets routed into hospitality tiers. A single-match hospitality package starts around $2,500, against $950 at Qatar, and private suites run from $43,200 past $100,000.

The worse the tournament got for the person in the stand, the better it looked in the boardroom in Zurich.

The dress rehearsal

None of this arrived unannounced. FIFA staged the rehearsal in the same country, in many of the same stadiums, one summer earlier. The 2025 Club World Cup opened with dynamic pricing asking $349 for Inter Miami against Al Ahly and settling at $55 by kickoff. A Chelsea semi-final ticket fell from $473.90 to $13.40 across a single weekend. The Associated Press counted more than a million empty seats over the group stage, with grounds averaging 56.7 percent full. DAZN piloted split-screen commercials during stoppages, a first in any FIFA competition. The final at MetLife stretched halftime to roughly 24 minutes for a stadium show.

Every complaint you have heard this month was visible there in miniature, a full year out. FIFA watched the pilot empty the seats and fatten the revenue line, then shipped the whole system to the World Cup.

The pattern completes on July 19, when the final carries the first halftime show in World Cup history. Madonna, Shakira and BTS with Chris Martin curating, and an interval expected to run well past the 15 minutes the laws of the game allow.

Twenty years, one move

I watched the 2018 semi-final in a pub with friends, England against Croatia. Trippier put a free kick in after five minutes and the whole room sang about football coming home. Then Perišić equalized and Mandžukić scored in extra time and the place fell into a sickening silence. I embraced sobbing strangers outside after the game.

That version of the World Cup, the communal one, already sat on top of this machine. We had all seen FIFA executives arrested on the news three years earlier. We sang anyway.

The playbook has been run for over two decades: FIFA books the revenue while someone else eats the cost. South Africa spent about $4 billion on stadiums and infrastructure for 2010, including a waterfront stadium in Cape Town built at FIFA's insistence over a cheaper upgrade in working-class Athlone. A FIFA delegate reportedly told city officials that "a billion television viewers don't want to see shacks and poverty." That stadium loses millions a year.

Brazil's 2014 stadium bill roughly tripled from the bid book, 78 percent of it public money, and the $550 million national stadium in Brasília ended up parking buses. Brazilian law enforced commercial exclusion zones around venues, inside which a local bar could not print the words "World Cup." FIFA left Brazil with about $4.8 billion, untaxed.

Qatar had the worst outcomes of all. The Guardian counted 6,500 South Asian migrant workers dead in the decade after the award, an all-causes figure that remains contested. Qatar's official count of work-related stadium deaths was three; the tournament's own chief later allowed "between 400 and 500." Whichever number you accept, FIFA's response was to tell federations to "focus on the football" and to steer its $50 million legacy fund toward international agencies rather than worker compensation.

Two days before kickoff, Qatar banned stadium beer sales and blindsided Budweiser, a sponsor paying roughly $75 million a cycle. The corporate boxes kept pouring. Even sponsors are customers here, and no customer is safe.

Underneath it all sat the governance. In May 2015 the US Department of Justice unsealed a 47-count indictment and Swiss police arrested seven FIFA executives at dawn at the Baur au Lac hotel in Zurich, with charges covering roughly $150 million in bribes paid across 24 years.

Chuck Blazer, the CONCACAF official who flipped for the FBI, kept a Trump Tower apartment mostly for his cats. When FIFA hired a former US federal prosecutor to investigate the 2018 and 2022 bids, it published a 42-page summary that he publicly disowned as materially misleading, and he resigned. The full 430-page report surfaced only in 2017, after a German paper began leaking it.

My favorite detail from the whole fiasco: FIFA asked the US government to classify it as a victim of the crimes its own executives had committed. The government agreed, and handed FIFA and its confederations more than $201 million of the money forfeited by the convicted.

FIFA's best case

FIFA's defense deserves a fair hearing, so here it is. FIFA is a lawful non-profit that says it reinvests more than 90 percent of its budgeted spending into football. It sends roughly $8 million per cycle to each of its 211 member associations. The $60 tier exists. Dynamic pricing genuinely cut some late prices. And the water-break ad money really does land with broadcasters this cycle rather than with FIFA themselves. Infantino's exact words were that there is "no additional revenue for FIFA."

Each point is true. None survive much scrutiny.

Take the 90 percent first. Most of what FIFA books as investment in football is the cost of staging its own tournaments: the prize money and the production and the operations behind the product it sells. By that accounting, McDonald's reinvests nearly everything in hamburgers.

The money that actually leaves the building goes to the 211 member federations, and each federation holds one vote in FIFA's presidential elections, so the development budget doubles as a campaign fund. Expanding the World Cup to 48 teams minted 16 more federations that owe their place to the man who expanded it. ESPN's Gabriele Marcotti summarized the cynics' reading of the expansion as "creating more inventory to sell to sponsors and broadcasters and generating more coin." And the most notorious development grant in FIFA's history, $10 million routed toward Jack Warner under the banner of an African diaspora legacy program, appears in the Department of Justice indictment as a bribe.

You have already seen the $60 tier for what it is: a few hundred seats in an 80,000-seat bowl. And the late price cuts arrived only after the loyal had paid the top of the curve.

The Infantino quote is accurate for this cycle but beside the point. Fox is charging up to $750,000 for a single water-break spot, per Wall Street Journal reporting, against a reported $485 million it paid for US rights to the whole tournament. FIFA manufactured that inventory and licensed it, and the American rights for 2030, still unsold, get priced against it.

The empty seats are the product

The thing FIFA actually sells broadcasters is the crowd, the noise, the color and the flags. This tournament opened with visible empty sections and, per a Financial Times analysis, roughly 176,000 unsold tickets across the opening round on the eve of kickoff. Infantino called demand "unprecedented by a factor of 10," and fact-checkers pointed out that the raw ticket numbers are inflated mostly because there are 104 matches now instead of 64. The prices keep out the fans who make the atmosphere, and the atmosphere is what FIFA sells. FIFA is degrading its own product and charging more for it than ever.

Of course, none of this has stopped me loving it. I watched England win at the Azteca on Sunday. It was one of the best matches I have ever seen, and I watched it from my living room. I moved to the country hosting this World Cup. The tournament came to me for the first time in my life, and the sensible choice was still to stay home. Multiply that choice by every priced-out fan who would have filled those quiet sections.

So on July 19 I will watch the final at home, the way I watched the opener. Somewhere around the 22nd minute the referee will raise his hand. Twenty-two players will walk to the touchline to drink from Powerade bottles, and a sales desk inside a broadcast tower will quietly hit its number. I thought money had already reached every conceivable corner of a football match. They found one more, and they sold it.

FIFA's statutes open by stating the organization's first objective: "to improve the game of football constantly and promote it globally." The body that exists to make the game better and more open has spent two decades making the greatest sporting spectacle on Earth worse and harder to reach. It has been rewarded with record revenue at every step.

Readers of this newsletter will recognize the shape. It is a story we have told many times before: when the customer has no alternatives, the product can be sacrificed in the interest of profits.

Sport, sadly, is not immune.

Enshittification is the gradual, deliberate degradation of a product or service by whoever controls it. Cory Doctorow coined the term in 2022 for digital platforms, which decay in a recognizable sequence: good to their users first, then worse for those users to please business customers, then worse for everyone to feed the platform itself, and finally collapse. The word caught on fast, and Paul Krugman even formalized an economic model in a 2025 article.

Doctorow has already chased the rot off the screen and into physical things along one key dimension: products with a chip in them and a lock on the chip. The printer that rejects a cheaper cartridge, the tractor a farmer is barred from repairing, the refrigerator that learns to show ads. A thin layer of software, shielded by law, becomes the lever, and the maker can make the product worse because you can neither escape it nor fix it.

This essay is about the other physical goods, the ones with no chip to lock and no app to spoil. Your backpack, your mattress, the sandwich at the diner. They get worse too, on a slower clock, with nothing digital trapping you. A different mechanism is at work.

That mechanism is ownership. My central claim is that a brand's quality tracks who it answers to, and that this, rather than greed, is what explains decline. Founders answer to the product and to a name that is usually their own. Outside capital answers to a return. Both want money, so greed is the one trait founders and financiers reliably share, which is exactly why it explains nothing about why some brands keep their quality and others rot.

The variable is time horizon and audience. An owner who will be there to collect a reputation's future earnings protects it. One who plans to exit, or who answers only to a margin target, spends it down, because a reputation is a depletable asset and the consequences for depleting it arrive only after they are gone.

The purpose here is to set that process down once, in full, as a reference. What follows defines the lifecycle a brand moves through as it changes hands, catalogs the specific mechanisms by which a reputation gets converted into cash, and explains why consumers usually cannot react fast enough to stop it. It ends with a short field guide for telling, before you buy, which side of the line a brand sits on. The evidence for each claim is the set of investigations this newsletter has already run, linked throughout and gathered at the end.

This is a living document - it will grow as I uncover new reputation extraction tactics through my research.

Reputation as capital

A reputation behaves like capital. Benjamin Klein and Keith Leffler made the point formally in 1981: a good name is, in economic terms, a stream of future earnings, and a company keeps its quality up largely because cheating the people who trust it would forfeit that stream. The reputation has a present value, and it sits on the books as an asset whether or not anyone writes it down as one. An owner decides, in effect, whether to live off the interest or sell off the principal.

Spending the principal either amounts to deliberately lowering input costs (quality) or increasing the price of a given product. When Krugman set enshittification down as a formal economic model, he left quality out of it on purpose and built it on rising prices alone. His model neatly fits a technology platform lifting its take but is less useful for a physical good that gets worse while the price holds.

In either case, the key point to consider is that a shorter time horizon will inevitably increase the aggressiveness with which an owner is willing to sacrifice a brand's reputation in the interest of short-term earnings. A private-equity firm working on a three-to-seven-year clock spends a reputation down fastest, since it will be long gone before the name runs out.

The lifecycle

The 'reputation extraction' pattern reliably moves through four stages.

A founder builds something good, and the trust accrues slowly, over years of the product earning it. Then, at some point, comes the handoff. The moment the company stops answering to a person and starts answering to capital: a sale, an IPO, a death, a hostile takeover. Then the harvest, where the new owner converts that trust into cash, cutting quality while the name still carries the old promise. Then the husk: a name with nothing behind it, kept alive because it can still be licensed out for whatever the logo alone is worth.

Scale is not necessarily the villain here. The power tool industry holds a good example. A Hong Kong conglomerate, TTI, bought Milwaukee in 2005. It poured money in: $206 million of R&D in a single year, with the founding family still holding about a quarter of the company. Milwaukee got better.

For an opposing example in the same industry, an American conglomerate, Stanley Black & Decker, bought Craftsman in 2017, carried billions in debt, and built a failed factory in Texas that could not reliably stamp its own name on a wrench and shut down within three years of opening.

Craftsman got worse. The same move in the same decade and the opposite result, because the owners answered to different things.

There are two ways to run the harvest, and the veterinary industry demonstrates both. Mars, the company behind Snickers and Pedigree, buys clinics and keeps them, milking the margin year after year under names like Banfield and VCA. KKR buys clinics, borrows against them (a single $2.3 billion loan the clinics themselves have to service), folds in more, and aims to sell the whole thing inside a few years. One holds and squeezes, the other flips. Either way the clinic is run as a financial instrument first and a place that treats your dog second.

Whole Foods is the arc end to end. John Mackey spent thirty-seven years building a grocery chain that answered to a mission and to the neighbors who once turned up with mops to bail it out of a flood. He took it public in 1992, which let Wall Street into the room. In 2017 an activist fund bought in and told him, by his own account, that it would replace his board, fire him, and sell to the highest bidder. He sold to Amazon for $13.7 billion. The fund walked off about $300 million richer two months later. Mackey was the founder the entire time. What changed was the audience he answered to, and the suppliers and the workers and the quality of products on the shelves all followed it down.

When researchers studied private equity buying up nursing homes, they found that ownership by a buyout firm raised patient mortality against comparable non-PE homes, differing only in who held it and for how long. The owner type was the variable that killed people. And when the harvest is done, the brand does not really die. The name detaches from anyone who ever made the product and becomes a thing to rent.

The mechanisms

These are the specific ways a reputation gets spent. The list is not complete, and it is not meant to be. New ones surface every time I look into a new category, and I'll add them here as they do.

Reputation laundering

Buy a loved brand out of bankruptcy, fire the designers and shut the factories, then rent the name to whoever will pay to stamp it on their product. Authentic Brands Group has run this playbook on Brooks Brothers, Eddie Bauer, and Sports Illustrated, among fifty-odd others, and admitted the whole thing in its own SEC filing: it does not make the products and has limited control over their quality. The trust the original company earned becomes a royalty check, cashed quarter by quarter until nothing is left, and then they go find the next name.

Phantom competition

Fill the shelf with brands that look like rivals and quietly own all of them. One Italian company, EssilorLuxottica, owns Ray-Ban and Oakley and Persol, the stores you would buy them in, and the vision insurance that pays for them. The three premium names on the mattress wall are a single manufacturer. Whirlpool sells you Amana at the bottom of the ladder and JennAir at the top and builds them in the same plants. The choice is an illusion.

Skimpflation

Swap the materials for worse ones and leave the label crisp. Backpacks lost their denier count and their YKK zippers while the logo stayed sharp. A Pennsylvania diner's all-beef patties turned into beef cut with soy filler across a few quiet contract renewals, and the owner only found out when he finally read the label.

Warranty theater

Keep advertising the guarantee that built the loyalty, then rewrite it so it never pays. JanSport's lifetime warranty now excludes "normal wear and tear" on a bag engineered to wear out. Eddie Bauer killed its lifetime guarantee outright in 2019. A twenty-year mattress warranty that voids over a single stain is a sales tactic wearing a warranty's clothes.

Planned obsolescence

Engineer the product to fail right around the end of the warranty so you sell another one. Smart appliances break roughly 40% more often than the plain versions, and a dead control board can cost a third of a new machine, so the broken one goes to the landfill on cue. John Deere built tractors a farmer can no longer legally fix without crawling back to the dealer, the practice that got the company sued by the FTC in 2025.

The unfixable failure point

When the component that fails is moved from a replaceable, visible, exterior part to an unreplaceable, invisible, interior one, the repair market dies and repurchase becomes the only path, without anyone having to shorten the product's life.

Distinct from planned obsolescence, which governs when a product fails. This governs where it fails.

Debt extraction

Borrow against the company you just bought, pay yourself the loan as a dividend, and leave the debt on the books of the thing you intend to sell. George Akerlof and Paul Romer named this one in 1993: under the right conditions it becomes rational to bankrupt a company for profit, looting it on the way out the door. It is why the Instant Pot, a genuinely great product with a devoted following, ended up bankrupt in 2023 after a private-equity owner loaded it with debt.

Renting the customer

This one is newer, and it belongs to the direct-to-consumer era. When you sell straight to people online, the cost of acquiring each customer can rival the cost of making the product itself. The money that could have gone into developing a better product goes to the ad auction instead. Casper spent 43% of its revenue on sales and marketing in 2017, against a cross-industry norm of 10 to 12%, and earned about three dollars back for every dollar it spent. Allbirds put a quarter of its revenue into marketing and still lost more money every year, $101 million in 2022 alone. Then Apple's 2021 privacy change made acquiring a customer more expensive overnight, costing Meta around $10 billion and hitting the smallest brands hardest. A company that has to rent its customers from an ad platform has less left for the product. When the platform raises the rent, the product is the first thing cut.

The auto-renew trap

Once they have you, leaving gets expensive. Banfield sells a wellness plan that renews on its own and front-loads the value, so cancelling means owing them money, and grieving owners have been billed for plans attached to pets that were already dead.

Sourcing churn

Moving a named product between factories and countries with no long-term purchase commitment, so the thing in the box changes while the name, the price, the colorway and the marketing stay constant.

Why it goes unpunished

None of this would survive if the people paying for it could walk away quickly enough to matter. They usually cannot, for three reasons.

The lag. A brand's reputation outlives its quality, sometimes by years, because you keep buying on the strength of what the company used to be. How long the lag runs comes down to how often you buy the thing and how easily you can judge it once you do. A mattress is the perfect target: you buy one maybe twice a decade, and you cannot tell a good one from a bad one in a showroom. A t-shirt is a terrible target, since you buy them constantly and you know inside three washes. So the extraction runs longest and richest in the durable, hard-to-judge categories: mattresses and tools and appliances and eyewear, which is not a coincidence and is most of what I end up writing about here.

The fog. You cannot judge what you cannot see, and George Akerlof won a Nobel for working out where that leads. His 1970 paper showed that when buyers cannot tell quality before they pay, the good stuff gets driven off the shelf by the cheap stuff that looks identical. The brands lean into the fog and thicken it on purpose. The mattress industry sells one bed under ten different names so that no two stores carry a comparable model and the price-match guarantee never has to pay out. Half the time you cannot tell the well-made version of a licensed brand from the gutted one, because they wear the same label.

Nowhere to run. The surest way to stop you leaving is to own everywhere you would go. Tim Wu, the antitrust scholar, put the eyewear monopoly plainly: it is as if Hermès and Louis Vuitton were secretly the same company, mimicking competition while pocketing both sales. In food distribution, the one place an independent restaurant could check a fair wholesale price, Restaurant Depot, is now being bought by Sysco, the giant they were checking it against. This is the physical-world version of the lock-in Doctorow described in software.

What to do about it

The defense is knowing who you are buying from before you pay, and a handful of pointed questions will help you make an informed decision before your next big purchase.

Find out who owns the brand now, and whether the founder is still there and still in control, because that one fact predicts more than any review. Check whether the company went through bankruptcy in the last decade, and who bought the name out of it. If the brand is public, check what share of its revenue goes to marketing. Notice when the same brand turns up at a flagship store and at Costco and at TJ Maxx all at once, a sign the name has been split across tiers and licensees. Ask how the person selling to you gets paid, because commission turns every recommendation into a sale. And run the math that marketing is built to keep you from running: divide the price by the years the thing will genuinely last, and compare across fifteen years instead of one weekend, because the cheap option you replace twice over is usually the expensive one. If you cannot find out who actually makes a product, that is your answer.

I keep the running tally on The Brand Ledger, a searchable record of every brand I have taken apart: who owns it now, what it used to be, and whether it is still worth your money. It changes as the brands change hands, because they will. This essay works the same way. It is one page meant to stay current, and it will be updated every time I discover a new mechanism.

A dollar spent with a company that still makes its own things is a dollar the roll-up did not get, and enough of those dollars is the only language this machine understands.

The machine

Strip it to the bone and the story is simple. The decline you keep noticing is a decision. Somebody chose it, on a clock you could not see, to collect a return that arrives before the bill does, and then they moved on to the next name on the list. The product got worse because the people who own it stopped answering to it.

That is the machine. It runs in tools and groceries and eyewear and your dog's vet, and I’m certain it will run in whatever category I pull apart next.

The case files

Every claim above traces back to one of these. Each is a single category, taken apart.

• Your Backpack Got Worse On Purpose — VF Corporation and the slow death of the lifetime warranty.

• Your Power Tools Got Worse On Purpose — two conglomerates, the same playbook, opposite results.

• Your Glasses Got Worse On Purpose — how one company quietly became the entire eyewear aisle.

• Your Favorite Brands Got Worse On Purpose — the licensing machine and the art of reputation laundering.

• Your Dinner Got Worse On Purpose — Sysco and the convergence of every menu in America.

• Your Mattress Got Worse On Purpose — the name game and the merger the FTC could not stop.

• Your Pet Care Got Worse On Purpose — Mars, KKR, and the roll-up of the neighborhood vet.

• Your Appliances Got Worse On Purpose — badge engineering and machines built to die on schedule.

• Whole Foods Got Worse On Purpose — a mission, an IPO, a raider, and Amazon.

Late May 1981. Around ten inches of rain fell on Austin in a single day, Shoal Creek went over its banks, and the worst flood the city had seen in decades left the first Whole Foods Market store sitting in eight feet of water. The registers, the coolers, the entire inventory, gone. Losses ran to about $400,000, and the company carried no insurance.

The store had been open less than a year. By any sane accounting it was finished.

What happened next became the company's founding story, and unlike most founding stories it appears to be true. Customers came with mops and buckets. Neighbors came. Staff who had no promise of ever seeing another paycheck came anyway and worked for free. Vendors extended credit they had no reason to extend. Investors put in more.

Twenty-eight days later the doors reopened. Co-founder John Mackey built his entire operating philosophy based on his experiences that month. A business survives, he came to believe, because the people around it decide it deserves to. The workers and the suppliers and the neighbors and the shoppers.

Keep that in mind. What follows is really two questions asked across four decades: Who did Whole Foods answer to, and what happened when that answer changed?

It started stricter than you remember

Before the flood there was SaferWay, a tiny vegetarian store Mackey opened in 1978 at twenty-five, funded with $45,000 borrowed from family and friends. A college dropout living in a vegetarian co-op, he sold no meat, no sugar, no processed food, and when he and his girlfriend got evicted for storing product in their apartment, they moved into the shop and bathed using the dishwasher hose.

The first store to carry the Whole Foods name opened in Austin in 1980. It sold meat and beer and wine, the things SaferWay refused, because purity does not pay rent. Mackey's own line for it was that they were a "whole foods store," not a holy foods store. The instinct for scale was there from the first store.

The company defined itself by what it would not sell. It banned added hormones in meat in 1981, refused the artificial colors and sweeteners that filled every other grocery aisle, and built the whole identity around a "no" list when no national chain was daring enough to do the same at the risk of losing business. That was the offer, and people paid up for it.

Big, and good, at the same time

Whole Foods got very big without the principles becoming a pose. For a long stretch, Mackey's values were the engine of the company's growth, not a tax on it.

The proof is that the standards grew with the company instead of getting left behind. It dropped partially hydrogenated oils in 2003, more than a decade before the FDA banned them, and cut high-fructose corn syrup in 2011. It became the first national grocer to go certified organic. It kept a $25 million loan fund for small local producers and kept "foragers" on regional payrolls hunting products too small for any national chain to bother with. A company chasing margin alone does not write stricter rules for itself or bankroll the suppliers it could squeeze. The idealism and the growth pulled the same way, and that is why the brand earned the loyalty it later burned.

Whole Foods went public in January 1992 at $17 a share. The money raised funded a buying spree, and that is where the first cracks show. Take Mrs. Gooch's, the Southern California chain Whole Foods bought in 1993 for about $56 million. Gooch's had been even stricter than its new parent, no alcohol on its shelves at all. Within a few years Whole Foods had added beer and wine and folded every Gooch's store into the Whole Foods banner. A tighter brand absorbed and loosened into the bigger one. That move would repeat, store by store and region by region, for the next twenty-five years.

By 2006 the outside world had clocked the gap between the marketing and the supply chain. Michael Pollan built a long section of The Omnivore's Dilemma around Whole Foods, coining "Supermarket Pastoral" for the reassuring farm-fable on the label, then tracking the actual farms and finding feedlots and sheds. Organic at scale, he wrote, had come to look considerably less like a movement than a big business, and he warned that scaling it up would eventually hollow it out. Mackey fired back with an open letter rejecting the Big Organic label and noting the company had grown from one store to 184. He was proud of the 184. Pollan's point was about what the 184 cost.

The cracks ran through the founder too. Mackey had once compared unions to having herpes, a line he later said he regretted. For roughly eight years he posted on Yahoo Finance message boards under the handle "Rahodeb", an anagram of his wife Deborah's name, talking up Whole Foods stock and running down a rival chain, Wild Oats, which he was quietly preparing to buy. He boasted, in the third person, that the company was destroying Wild Oats market by market, city by city. When Whole Foods moved to acquire Wild Oats in 2007 for $565 million, the FTC sued to block it, warning that the merger would let the combined company raise prices and reduce quality and services.

In 2009 Mackey wrote a Wall Street Journal op-ed titled "The Whole Foods Alternative to ObamaCare", opening with a Margaret Thatcher line and arguing that Americans had no "intrinsic right to health care." The title put the store's name on the position, and a lot of his shoppers were exactly the people who backed the law. They organized a boycott.

Six years later the company gave them a blunter reason. In 2015 New York City inspectors found that every one of the 80 prepackaged product types they tested was mislabeled, with 89% of packages overcharging customers past the federal limit, in what the city called "the worst case of mislabeling they have seen in their careers." Whole Foods paid $500,000 to settle.

By 2015 the company plainly answered to two masters. The mission, and the stock. The next thing that happened decided which side won in the end.

The sale

By the middle of the 2010s Whole Foods’ edge was waning. Walmart and Kroger had stocked organic, Trader Joe's and Sprouts were undercutting on price, comparable-store sales were sliding, and the stock had given back roughly half its value from the 2013 high. "Whole Paycheck" had hardened from a joke into a diagnosis.

In April 2017 the activist hedge fund Jana Partners disclosed a stake of around 9% and immediately pushed for a sale. Mackey called them "greedy bastards" in Texas Monthly. Years later he described the meeting flatly. They told him they would take over his board, fire him, fire his executives, and sell the company to the highest bidder, and there was nothing he could do about it. He called Warren Buffett, who passed. He looked at Albertsons. He landed on Amazon.

On June 16, 2017, Amazon agreed to buy Whole Foods for $42 a share, about $13.7 billion, all cash. Bezos, in the announcement, called it the place that offered the best natural and organic foods. Jana, which had paid between $29 and $32 a share, sold its entire stake within two months of the announcement and walked away with roughly $300 million. The announcement alone wiped an estimated $22 billion off the stock value of competing grocers in a single day. That $300 million was a reward for financial maneuvering. The value got pulled out of a company other people had spent thirty-seven years building.

Mackey sold the deal publicly as a win-win-win, good for shoppers, workers, suppliers, shareholders. Privately, by his own later account, he did not want to sell at all. In his 2024 memoir he wrote that he regretted the circumstances that made Amazon the best option, though he would make the same call again.

What gets left out of his telling is that he built the trap himself. The 1992 IPO he celebrated is what created the quarter-to-quarter pressure he came to resent, and it is what left the company exposed to a raider who could force a sale. He spent decades preaching that a business should answer to all of its people, and then handed it to the most efficient shareholder-value machine ever built, because the structure he had chosen left him cornered.

What got worse

The decline under Amazon shows up in the documentation, and it runs straight through the founding principles one at a time.

Start with the suppliers, the small producers the whole thing was built to champion. Whole Foods made its name as the rare national grocer that actually sourced locally, giving each region its own buyers and foragers to dig up the local products bigger chains ignored. Under Amazon, the buying moved to headquarters in Austin, and the model inverted. The system now rewarded suppliers who could serve a national distribution network and pay for their own shelf space, the exact kind of supplier the old Whole Foods had defined itself against.

The new terms did the rest. Whole Foods began charging suppliers for prime placement, told any vendor selling more than $300,000 a year to hand over a 3 to 5% discount, routed in-store demos through a paid outside firm, and scrapped the minimum-order and shipping arrangements small producers had counted on. The bar became logistical. A producer who used to walk a few cases into one store now had to supply a centralized operation, meet its volume and delivery terms, and carry the warehousing to do it, or come off the shelf. A great product made in small batches, the kind the foragers used to find, increasingly could not clear that bar no matter how good it was. One pasta-sauce maker watched its sales fall 75% after the chain cut back its product, the kind of small producer that used to be the entire point.

Continuing the centralization trend, Whole Foods now leans on one primary distributor, United Natural Foods, locked in on a deal that runs to 2032. In June 2025, a cyberattack knocked UNFI offline and Whole Foods shelves went bare for days, refrigerators and bread aisles and produce sections emptied out, the company warning staff the fix could take "several days." The grocer that built its name on hundreds of local suppliers and regional foragers had wired itself to a single point of failure, and the point failed.

Then the shopper, who got the loudest promise and the quietest delivery. The day the deal closed, Amazon cut prices on staples and the Bezos-owned Washington Post announced that "Whole Paycheck" was finally dead; organic apples dropped 43% overnight. The independent checks were duller. One research firm priced 110 items a few weeks later and found the overall basket had actually risen about 1%, the cuts concentrated on produce while dry goods and snacks crept up. The deepest savings were walled off behind an Amazon Prime membership. A year in, most people who skipped Whole Foods still gave the reason they always had: too expensive.

The price was never really the point. People paid up for the curation, the prepared foods, the organics, and the small, strange, local finds the foragers used to bring in. That is the part that thinned. National brands and the 365 private label spread across the shelf space, the assortment drifted toward what every other grocer already carried.

Then the workers. Under Order-to-Shelf, the inventory system Mackey rolled out in early 2017 and Amazon then leaned on, managers walked the aisles with checklists, and if a single item was out of place, the manager responsible got written up. Three write-ups and you could lose your job. Management called it correcting errors. The staff called it punishment. Seeing someone cry at work became normal.

The kicker is that it failed at the one thing it was for. The system built to keep the shelves full ran them empty instead, because the back rooms had been gutted and there was nothing left to restock from. While that was happening, the company's own operations chief had told Wall Street analysts the staff were "really excited" about it.

In September 2019 Amazon-owned Whole Foods cut medical benefits for its part-time staff, raising the eligibility bar from 20 hours a week to 30 and stripping coverage from around 1,900 people. The following spring, reporters revealed an internal heat map scoring all 510 stores on the risk that their workers might unionize, built on more than two dozen inputs including employee "loyalty," proximity to a union office, local poverty rates, and the racial diversity of the staff. Critics called it Minority Report union-busting.

By the time the workers pushed back, the founder was gone. Mackey stepped down in 2022, handing the chief executive job to Jason Buechel, his longtime operations deputy. In January 2025 the Center City Philadelphia store voted 130 to 100 to unionize with UFCW Local 1776, the first union at any of Amazon's 500-plus Whole Foods stores. The company said it was disappointed, challenged the result, and as of mid-2026 the fight is still grinding through the NLRB.

Tally the founding principles against today. Ingredient standards: mostly intact, and credit where it is due, the "no" list survived. Decentralized local sourcing: inverted. Stakeholder treatment of workers: degraded into scorecards and heat maps. Independence: gone. The store still looks like Whole Foods on the way in. What changed is who it answers to on the way out.

Why the founder mattered, and what to actually watch

It would be easy to end on "a founder built something good and a corporation ruined it"… but that doesn’t tell the whole story. Mackey was the founder through Rahodeb, through the FTC fight, through the overcharging. His presence prevented none of it.

What actually tracked the quality decline was the company's accountability. At the start it answered to a mission and the people who saved it from a flood. The 1992 IPO let Wall Street into the room. The 2017 raid put Wall Street in the driver's seat. The Amazon sale handed the whole thing to a parent that measures a grocery store the way it measures a warehouse. The founder was present the entire time. What changed under him was the audience he was performing for, and the product followed the audience down.

So the useful signal is narrower than "buy founder-led." A founder who still controls the company is a solid indicator that the business still answers to the principles it was founded under, rather than to outside capital that was never interested in the product except as a number.

Founder control plus real independence is the strong version of that signal. Public and raidable is weaker. Owned outright by a giant optimizing for throughput is weaker still. Sadly, that is where a lot of the brands you grew up trusting now live.

Whole Foods was at its best when it answered to people who would show up with a mop. It got worse, measurably and on the record, as it came to answer to people who never would. That is the test. Before you hand a brand your money and your loyalty, find out who it actually answers to.

P.S. Big thanks to Steven Shonts, who helped a great deal in sharpening my understanding of Whole Foods’ supplier dynamics. Check out his squid jerky below (I was skeptical… it is awesome):

The Hermit makes jerky from wild-caught American squid (a single fishery, a single species, delivered direct by the brand to consumers and independent retailers). They are committed to always being an artisanal operation focused on producing quality food.

Try The Hermit →

Nowhere is the illusion of choice more potent than in the appliance showroom. Dozens of brands fan out across the price tiers, a wall of brushed steel under the lights, and behind all of it sit five or six companies. The labels compete. The machines mostly don't.

That arrangement pays off when your machine dies young, and these days it dies young on schedule. Americans spent 43% more on appliances in 2023 than in 2013 even as sticker prices fell, because the things wear out faster and get replaced more often.

Cheap parts and efficiency rules often take the blame, but consolidation is the real cause. The decline came on deliberately, built into the machines one brand at a time, as the companies that once competed on quality stopped having to.

A handful of holdouts still build appliances that last for decades, and that is telling. Durability is clearly still possible. Most of the industry just decided it was worth less than the next sale.

One company, four price tags

Start with Whirlpool. It owns brands on every rung of the price ladder.

Amana sits at the bottom as the budget badge. Maytag plays the dependable workhorse, KitchenAid the premium brand for people who care about their kitchen, and JennAir the luxury tier. Every one of them is Whirlpool, which the company lists plainly in its own corporate materials. They share wash systems, motors, compressors, and control boards across brands, built in the same plants. A Maytag and a Whirlpool of the same class are often the same machine in different sheet metal. The higher tiers add real materials and features, but you're still buying up a ladder one company controls end to end.

Kenmore takes the trick further. It never built a single appliance, not one, in its entire history. For decades it was a Sears nameplate bolted onto whoever won the contract that year, Whirlpool or LG or Frigidaire. Today it survives as a licensing shell owned by the hedge fund that picked Sears apart.

Consolidation like this carries a cost. When Whirlpool bought Maytag in 2006, it closed the Iowa facilities that had built Maytag washers since 1893. The plant shutdowns that followed cost around 4,500 jobs. At least those were American companies hollowing out American towns. Most of the names in your kitchen aren't even American anymore.

The brand on the box

When people think of American appliances, they think of General Electric. The name goes back to Edison and it still reads as homegrown on the showroom floor. GE Appliances has belonged to a Chinese conglomerate since 2016, when Haier Group bought the business for $5.6 billion and locked in the right to keep selling under the GE name through 2056. The headquarters and the factories stayed in Louisville, which is the part that lets it still feel American. The ownership and the money went to Qingdao.

GE's route to Chinese ownership ran through Washington first. The company lined up a sale to Electrolux, the Swedish appliance giant, back in 2014 but the US Justice Department sued to block it on antitrust grounds, worried that one company would control too much of the American range and oven business. The deal fell apart in 2015, and Haier stepped into the opening a year later. Washington had stopped an American brand from going to a Swedish buyer and instead cleared its path to China.

GE is only the most familiar case. Frigidaire, the name that became a synonym for the refrigerator itself, is Swedish now, owned by Electrolux. Bosch and its sister brands answer to a German parent, Samsung and LG report to Seoul, and the rest of the wall sorts the same way. The American kitchen is mostly a rack of nameplates owned between Qingdao, Stockholm, Stuttgart and Seoul. The flags on the boxes stopped meaning much a long time ago.

Foreign ownership is only half of it. The other half is that a lot of these competing brands roll off the same line. Badge engineering shows its real face in a recall, and a big one landed in June 2025, when the Consumer Product Safety Commission recalled about 1.7 million window air conditioners. They had been sold under at least nine names: Frigidaire, Insignia, Danby, Keystone, Mr. Cool, Comfort Aire, Perfect Aire, Sea Breeze, and LBG. A shopper could stand in the aisle, weigh them against each other, and feel like the choice meant something. Every one was built by the same company, a Chinese manufacturer called Midea that most buyers have never heard of, and every one carried the identical defect: a drain that let water pool inside until it grew mold.

Nine names on the box, and behind them one machine, broken the same way every time. The brand was decoration, the cheap part of the whole arrangement. The expensive part starts once the thing is in your house.

Why it dies young

The reason these machines fail early is built into them. Older appliances were mostly steel and mechanical controls, while newer ones run on plastic and circuit boards and Wi-Fi, and every one of those additions is one more part that can fail. The connected ones fail the most of all. J.D. Power's 2025 data put smart appliances at roughly 40% more problems per hundred units than the plain versions, which means the upgrade you paid extra for is the part most likely to quit.

When something does break, the math is rigged toward the landfill. A failed control board can cost a third of what the whole appliance did, and tariffs on imported parts, the compressors and boards and motors, pushed repair bills up another 5 to 20% in 2025. Throwing the thing out becomes the rational move, which is exactly where the design was pointing you.

LG's linear compressor is the clearest case of all this. The company sold it as the better option, quieter and more efficient, and marketed it on a twenty-year life. It became one of the most notorious refrigerator failures of the past decade. A 2020 settlement covered roughly 1.5 million people whose LG refrigerators stopped cooling, and LG denied any defect the entire way through. The part itself was never fixed. Fresh lawsuits in 2024 and 2025 allege the same compressor still dies in newer fridges.

So LG charged a premium for the feature that killed the fridge, then took the position that nothing had gone wrong. Somewhere around here, someone always blames the government.

It isn't the government

The theory goes that federal efficiency rules wrecked the machines by cutting the water and the heat, and there is a real tradeoff in there worth discussing. But a deeper look shows that the efficiency rules are only a small factor. The Department of Energy's own testing found the standards could be met with cleaning and cycle times close to the old machines, and low-water front-loaders often clean better and treat fabric more gently than the water-guzzlers people miss. The rules also swing with each administration, tightened under one and loosened under the next, which makes them a thin explanation for a decline that ran straight through all of it.

When the rules do bite, manufacturers just game them. Efficiency gets scored almost entirely on the Normal cycle, so Speed Queen built a Normal cycle weak enough to pass the test and kept a Deep Fill option that floods the full 24-gallon tub, which the test never touches. Dealers reportedly stuck labels over the Normal button telling owners not to use it. The tested cycle is theater, and blaming gallons-per-load is the comfortable story that keeps the real culprit, cost-cutting under consolidation, off the hook.

What still lasts

Good machines still exist. They cost more, and the showroom is built to sell you everything but them, so knowing which to choose is a real challenge.

The principles matter more than the brand here, and they hold at any budget. Every feature is one more thing that can break, so the simplest version of whatever you need is usually the most durable one: skip the Wi-Fi and the touchscreen, take a top-freezer or side-by-side over a French-door, and choose the model with the in-door icemaker left off, since icemakers sit among the most common failure points in any fridge. Favor knobs over circuit boards. And when you compare prices, compare them across fifteen years, not across one weekend.

For a normal budget, a plain Whirlpool or Maytag is the honest answer. Repair techs name them most often for parts and easy service, and Whirlpool sits near the top of Yale Appliance's reliability data, mostly because the company kept building straightforward mechanical machines while everyone else bolted screens onto every surface. It still makes them here, too, so a failed part is usually a quick trip to a local shop rather than a six-week wait on a container ship. These are ten-year machines and not heirlooms, and the reliability runs model by model, but under a grand it is the best bet going. The punchline ties back to the top: this is the same company as KitchenAid and JennAir, so the premium badge buys you a nicer finish and the same lifespan.

Spend more, and only a few brands are built to last for decades. Miele is the strongest of them. It is still family-owned and still built in Germany, and it tests its core machines to the equivalent of twenty years of use before they ship, which no other major maker claims. Set that against LG, where the same twenty-year promise is now propped up by a settlement and a second wave of fraud suits. One company means it. The other is in court over it.

The rest of the short list runs along the same line. Sub-Zero and Wolf are still independent and family-controlled, and their refrigerators routinely run past twenty years, though Wolf did recall some dual-fuel ranges, so nobody here is spotless.

For laundry, Speed Queen is the pick, steel where the others use plastic and light on the electronics that tend to fail, and a roughly $1,600 top-loader built to last two decades pencils out cheaper per year than a $550 machine you replace twice. And yes, this is the same Speed Queen that gamed the Normal cycle. The gaming cuts in your favor here: skip Normal, run Deep Fill, and the machine washes the way the old ones did. The asterisk is that the company is now private-equity-owned and recently public, so it earns a watchlist rating in The Brand Ledger.

Bosch makes the most reliable dishwashers by a wide margin, though it answers to a German conglomerate, which makes it a vote for the dishwasher more than the company.

One caution about the rankings before you shop on them. LG and Samsung post decent first-year service numbers, and that gets misread as durability, when first-year service says nothing about how a machine ages. LG's compressors tend to fail later, out of warranty. Yale dropped Samsung from its data altogether because the company has almost no parts-and-service network behind it. A machine you can't get fixed isn't reliable, whatever it scored in its first month.

Every brand named here, plus the ones that didn't make the cut, sits in the appliance section of The Brand Ledger with the reasoning for each call.

The fix is bigger than your next fridge

None of this is permanent, and the fix requires more than careful shopping. The European Union already makes manufacturers supply parts for at least ten years and build machines that ordinary tools can open. A handful of US states are starting down the same path. Right-to-repair laws won't un-consolidate the industry, but they take away the part that pays: the machine engineered to die just after the warranty does. They are worth your attention and your vote.

Until then, the defense is knowing the game. A plain fridge from one of the holdouts, knobs and a freezer on top and no screen, will outlast the $3,000 model that texts you and cost far less over its life.

You have been to one of these clinics. Maybe it was a Banfield tucked into the corner of a PetSmart, or a VCA with a bright clean logo. Maybe even a place with a warm local name like ‘Lakeside Animal Hospital’. At some point the visits started changing. The prices climbed faster than you remembered. The vet you trusted left without much explanation. A routine appointment came back with a recommendation for bloodwork and a dental cleaning and a bag of prescription food all at once. You paid it, because of course you paid it.

What is never clear when sitting in the lobby is who owns the building. The largest provider of veterinary care on the planet is Mars, the company that makes Snickers and M&Ms and Pedigree. It also runs more than 2,000 clinics in the United States under names like Banfield, VCA and BluePearl.

A decade ago, fewer than 10% of American veterinary practices were corporate-owned. The buying has concentrated almost entirely on the clinics that treat dogs and cats, while equine and farm-animal medicine have been largely left alone. In that companion-animal world corporate consolidators now own roughly a quarter of primary care clinics and about three-quarters of specialty and emergency hospitals, together accounting for around half of all veterinary revenue in the country. The neighborhood vet, as a category, is being absorbed by a small number of conglomerates and private equity firms that would prefer you not notice.

The two acquisition strategies

There are two kinds of buyer.

The first buys clinics and keeps them. Mars is the primary example here. It took a stake in Banfield in 1994 and bought the chain outright in 2007. It then acquired the specialty and emergency group BluePearl in 2015, and in 2017 paid roughly $9 billion to take VCA private. A single deal that added around 800 hospitals and handed Mars control of Antech, the lab that runs diagnostics for much of the country (Fortune; FTC). Mars does not flip what it buys. It accumulates.

The second operates on a clock: buy a platform, borrow against it, fold in smaller practices, squeeze costs, and sell in three to seven years at a multiple of earnings. KKR's PetVet Care Centers has grown from about 125 hospitals in 2017 to more than 450, financed in 2023 by a $2.3 billion loan the clinics themselves now have to service. JAB Holding, the family office behind Krispy Kreme and Panera, built a rival empire through National Veterinary Associates and was ordered by the Federal Trade Commission to divest clinics twice in 2022 alone.

Either way, the outcome does not favor the pet owner. Whether a buyer means to hold a clinic forever or sell it in five years, it is run as a financial instrument first and a medical practice second. The landlord still sets margin targets, and the flipper still has to make the debt payments. In both cases the care your animal gets and the price you pay for it were settled in a meeting you weren't in, by people who will never meet your dog.

Why dogs and cats

The reason this particular corner of medicine attracted billions of dollars in acquisition money is due to the structure of the business. Veterinary care is paid in cash at the point of service, with none of the insurer-reimbursement lag that complicates human medicine. Demand is close to inelastic, because the person holding the leash will find the money when the alternative is watching a family member suffer. And the liability is minimal, since pets are legally classified as property, which means a botched outcome that would trigger a malpractice catastrophe in human medicine produces, at most, a claim for the replacement value of a dog.

The practices themselves are profitable and, until recently, cheap. Veterinary valuation brokers (the firms that exist to sell clinics to consolidators) describe the economics with startling candor. A healthy independent clinic runs an operating margin somewhere in the low-to-mid teens. Brokers advise sellers that anything under 10% is a red flag for buyers, while a margin of 18% or higher is what attracts a premium price. One brokerage refers to practices outright as "flippable houses," properties a buyer can pick up, tidy, and resell at a profit. The gap between what an independent vet is willing to earn and what a corporate owner intends to extract is the entire engine of the story, and closing that gap is done in the exam room, on the invoice and in the staffing schedule.

The last ingredient was supply. American veterinary medicine was, until about fifteen years ago, a cottage industry of some thirty thousand independent practices, a large share of them owned by veterinarians old enough to be thinking about retirement. A consolidator can offer a retiring owner two, five, sometimes ten times what the practice would fetch from the younger associate who has worked there for years and assumed they would one day buy in. The associate, carrying student debt that now routinely runs past two hundred thousand dollars, cannot come close to matching that offer. So the practice sells, the name on the door stays the same and the new owner is a holding company in another state.

The workaround

Eighteen states prohibit anyone who is not a licensed veterinarian from owning a veterinary practice, on the old and sound principle that a person whose job is to generate returns should not be making medical decisions or directing the people who do. The consolidators worked around this with a structure borrowed from human healthcare, and they did it in the open, with law firms advertising the method as a service.

It works like this. The corporation creates a management services organization (MSO) which owns everything about the practice that is not, strictly, the practice of medicine: the building, the equipment, the inventory, the staff, the scheduling, the billing, and the bank account. A veterinarian, frequently one chosen by and friendly to the corporation, owns the medical entity on paper and is said to control clinical decisions. The two sign a management agreement under which the vet-owned entity pays a fee to the MSO. That fee is calibrated to route essentially all of the profit back to the corporation.

The American Economic Liberties Project, which has drafted model legislation to ban the structure, concluded that these arrangements render the state ownership laws "all but meaningless." This is not a loophole someone stumbled into. Veterinary mergers-and-acquisitions attorneys market MSO setups to buyers as the standard tool for entering states where direct ownership is illegal, and the trade press explains the mechanics in how-to articles aimed squarely at the people doing the buying.

What it costs you

The clearest measure of the change is the price. Veterinary service prices in the United States have risen more than 60% over the past decade, far outpacing inflation.

Some of that increase buys genuinely better medicine: the MRIs and CT scanners and oncology that can extend an animal's life in ways that were impossible a generation ago. Some of it pays technicians the raises they were long owed. But a meaningful share is just new ownership deciding the practice would charge more.

A veterinary-telehealth executive who studied regional pricing found that increases tended to land "immediately after the sale to a private-equity-owned group," and one emergency vet described prices rising 20% in a single year.

The mechanism by which a spreadsheet target becomes your invoice is the way these chains pay and pressure their veterinarians. Vets at corporate practices are frequently paid on production (a percentage of the revenue they personally generate) which quietly turns every recommendation into a transaction with a stake for the person making it. Veterinarians told the New York Times that corporate managers pushed them to see more animals and order more tests and sell more wellness plans and food. One vet recounted quitting after she was told her "cost per client" was too low. A peer-reviewed study in 2025 found that veterinarians in corporate practices reported significantly more pressure to generate revenue and to see more clients per shift than their peers in independent practice.

The pressure reaches past the clinic into the businesses it buys from. Mars owns Antech, one of the two labs that dominate veterinary diagnostics in North America. So when a Mars veterinarian orders bloodwork, the sample goes to a Mars lab. JAB runs the same play across more of the chain. It owns hundreds of clinics, a string of diagnostic labs, and a pet-insurance operation covering millions of animals.

The most revealing artifact of the whole model is Banfield's Optimum Wellness Plan, the monthly-payment package Mars sells inside PetSmart stores as a way to save on routine care. The plan is a twelve-month contract that renews automatically, and the advertised savings are real enough on paper, but the structure is built so that leaving is expensive. The valuable services, the dental cleaning and the vaccines, are loaded into the front of the year, so a customer who cancels partway through is told they have already consumed more value than they have paid in and must settle the difference. Cancellation without penalty is only possible during a narrow window before the renewal date.

The complaint records at the Better Business Bureau and ConsumerAffairs run for pages, and a striking number follow the same script: a pet dies, the grieving owner calls to cancel, and a representative explains that the contract is still owed. Owners report being billed $151, or $250, or in one case $719 that was turned over to a collections agency, for wellness plans attached to animals that were already dead. The practice is old enough to have produced a federal class action back in 2013, which alleged that Banfield's upselling wiped out the very discounts the plan advertised. Mars has been collecting on dead pets for more than a decade, and the plan is still for sale.

The vets

The squeeze reaches the veterinarians too, and it lands on a profession that was already struggling. New graduates leave veterinary school owing around $150,000 against starting salaries that do not come close to covering it. They walk into a workplace where, more and more, the protocols and the prices and the pace are set by corporate. They lose the clinical autonomy that drew them to the work, and they carry the moral weight of recommending care that families cannot afford while having no legal standing to treat an animal whose owner says no. The profession's suicide rate runs well above the general population's. While corporate ownership did not create that crisis, the loss of control and the constant revenue pressure aggravate the conditions that drive it.

What anyone is doing about it

Under Lina Khan, The Federal Trade Commission treated veterinary roll-ups as a genuine antitrust problem and forced divestitures in the deals large enough to notice. For instance, it ordered Mars to sell a dozen clinics as a condition of the VCA purchase and also forced JAB to divest clinics twice in 2022 as it swallowed SAGE and Ethos. The agency even imposed a standing requirement that JAB give advance notice before buying any specialty or emergency clinic within twenty-five miles of one it already owned.

The trouble is that the great majority of acquisitions are too small to trigger federal review at all, and a roll-up assembled from dozens of individually unremarkable purchases can reach regional dominance without a single deal ever crossing the threshold that would land it on the Commission's desk.

So the fight has moved to the states, where it is mostly losing. The American Economic Liberties Project has published a model bill, the Save Our Pets Act, that would ban the MSO workaround outright and require disclosure and review of veterinary transactions. A version of the transaction-review idea was introduced in New York in late 2025. But as of early 2026 more of these state bills have died than passed, and the federal posture under a new administration is an open question.

There is no cavalry coming over the hill. What exists is a slow and underfunded effort to make ownership visible, which is worth supporting precisely because visibility is the one thing the model depends on you not having.

What is within reach of any pet owner is smaller, though it is not nothing. You can learn who owns your clinic before the day you urgently need it. Ownership is rarely posted on the door, so ask at the front desk whether the practice is independent or part of a group, and notice how readily they answer. You can look the practice up on your state veterinary board or business registry, where an out-of-state parent or a separate management company is often the giveaway. You can run the clinic's name against the major portfolios: Mars (Banfield, VCA, BluePearl, AniCura), JAB's National Veterinary Associates, KKR's PetVet, and groups like Thrive, VetCor, and the combined Mission and Southern operation, which hold hundreds of clinics between them.

The signals that matter most are harder to check from a website. Ask how the veterinarians are paid. A practice that puts its vets on salary rather than a cut of what they bill has removed the reason to sell you things your animal does not need. Ask for a written estimate, and ask which items are necessary and which are optional, and watch whether the vet treats that as a fair question or an irritation. The good ones are still out there. They have usually kept the same owner for years, they do not rush you, and they will talk plainly about money. That kind of clinic is worth driving past three corporate ones to reach.

Coming for the food bowl next

Step back far enough and the veterinary story is one limb of something larger. The same Mars that owns your clinic and the lab it sends bloodwork to also makes Pedigree, Whiskas, Royal Canin, and Iams. Petcare is the single largest division inside Mars, bigger than the candy, and it accounted for 59% of the company's sales in 2023.

The food aisle has been consolidated by the same handful of conglomerates that took the clinics: Mars, Nestlé, General Mills, and a few others. The brands that look like spirited independents on the shelf mostly answer to one of them. You eat a tiny fraction of any single food company's output across a week. Your dog may eat one company's product, exclusively, for its entire life. That is a separate essay, and it is coming soon.

I've been shopping for a mattress lately. The plan was simple: figure out what was actually in the bed I'd be sleeping on for the next decade, pick the best one for the price, and get back to my life.

Two weeks in, I had a thirty-row spreadsheet, a dozen browser tabs open and was still no closer to my answer. I was looking at Tempur-Pedic, Sealy and Stearns & Foster, the three big premium names you see in every showroom. I noticed that a certain coil count would appear on one retailer's product page and differ from another's selling what was supposedly the same mattress. Foam densities listed by retailers didn't match the manufacturer's spec sheet either. Independent review sites described materials that the official product pages didn't list. Each brand had between three and eight "models" with subtly different names I couldn't reliably tell apart from each other, much less from competing brands. The information was was contradictory in ways that seemed engineered to prevent comparison.

The practice has a name. Mattress retailers call it the "name game," and it exists to deliberately confuse buyers. The manufacturer makes one mattress, then ships it to ten different retailers with a different cover and a different model name on each. Macy's carries the Sealy Posturepedic High Point Hybrid. Mattress Firm carries the Sealy Posturepedic Elite Brenham II. Sleep Outfitters carries the Sealy Posturepedic Plus Albany. Same coils, same foam, three different names, three different prices. The retailer advertises a low-price guarantee, knowing it'll never have to honor it because no competitor sells that exact model.

The industry doesn't really hide this. Industry-funded comparison site GoodBed explains the mechanic in detail. US-Mattress (one of the largest online mattress retailers in the country) markets itself as the place where you can crack the code and regional independents publish blog posts on the practice.

So I started digging. And it gets worse.

The trick is older than the showroom

Mattresses are the perfect product for this kind of obfuscation. You buy one every eight or ten years. You can't really test one in a fifteen-minute showroom visit. The materials inside are invisible once the cover is sewn shut. The product is large, awkward to return, and emotionally entangled with sleep, which is something everyone is bad at and worried about.

Walk into any mattress store in America and three brand names dominate the premium wall. Tempur-Pedic, originally founded to commercialize a NASA-developed pressure-relieving foam, now synonymous with the entire memory foam category. Sealy, the innerspring giant that's been making mattresses since 1881 and selling them as the default American bed for most of the twentieth century. Stearns & Foster, founded 1846, marketing copy heavy on words like "master craftsman" and "heirloom," positioned at the top of the wall where the salesperson takes you when you say price isn't really the issue.

Standing in the store, you would think you were looking at three competitive brands.

They're one company.

Tempur-Pedic acquired Sealy in 2012 for $1.3 billion, and Stearns & Foster has been a Sealy brand for decades. The combined entity, then known as Tempur Sealy International, did about $5 billion in annual revenue and accounted for roughly a third of the entire U.S. mattress market.

When you stand in a Mattress Firm and a salesperson walks you from the Tempur-Pedic LuxeAdapt to the Sealy Posturepedic Plus to the Stearns & Foster Estate, comparing feel and price, the salesperson is walking you between three offerings from the same manufacturer.

Then they bought the store

Tempur Sealy makes the mattresses. Mattress Firm is the largest mattress retailer in the United States. More than 2,300 stores, about $2.5 billion in annual revenue. In May 2023, Tempur Sealy announced it was buying Mattress Firm for $4 billion. The deal eventually closed at $5.1 billion in February 2025.

The Federal Trade Commission, in a rare bipartisan 5-0 vote, tried to block it. The FTC's argument was that the world's largest mattress manufacturer buying the country's largest mattress retailer would allow Tempur Sealy to crowd rival manufacturers off of Mattress Firm's shelves. Vertical integration as exclusion. The FTC took it to federal court in Texas and lost.

Federal Judge Charles Eskridge of the Southern District of Texas, in his 115-page opinion denying the FTC's preliminary injunction, did not find that the FTC was wrong about the merger's likely effects. He found that the FTC was probably right and approved the merger anyway.

A legal analysis of the ruling by Paul Weiss summarized the court's acceptance of the FTC's core finding: “the combined firm will have a profit-aligned incentive to increase the sales of Tempur Sealy mattresses after acquisition by excluding rivals from the Mattress Firm floor.”

In other words, the court agreed that Tempur Sealy now has both the means and the motive to use Mattress Firm's 2,300 stores to push rivals out of the market. The court approved the deal because the foreclosure share, the percentage of total mattress sales that Mattress Firm represents, was estimated at "approximately one and nine percent, which isn't concerning given the competitive nature of the mattress industry." The deal closed on February 5, 2025.

The 'remedy' Tempur Sealy offered to get the deal through, and which the court accepted as sufficient, was a commitment to reserve 43% of premium ($1,500+) floor slots at Mattress Firm for third-party brands for five years. The other 57% (and 100% of the non-premium floor space) is now contractually permitted to be Tempur Sealy products. After 2030, even that 43% commitment evaporates.

The combined entity now operates about 3,000 retail stores, 30 e-commerce platforms, 71 manufacturing facilities, and 4 R&D facilities globally. The world's biggest mattress manufacturer owns the country's biggest mattress retailer, and you cannot walk into one without primarily seeing the other.

Thirteen days after the deal closed, Tempur Sealy International renamed itself Somnigroup International. The ticker changed from TPX to SGI. The consumer-facing brand names did not change.

The play wasn't new for Tempur Sealy. They had already acquired Dreams, the UK's largest specialty bed retailer, in 2021 for £340 million. Dreams now operates 217 stores and generates roughly £412 million in annual revenue.

The Mattress Firm bankruptcy is its own story

Some of you will remember when Reddit briefly convinced itself Mattress Firm was a money-laundering front. The theory was that there were too many Mattress Firms on too many corners doing too little visible business, and the answer had to be illicit cash flowing through the registers. The conspiracy was wrong about the cause but accidentally onto the symptom.

Mattress Firm sued its own former employees and outside real-estate brokers in 2017 for what its court filings called a "massive, multi-year fraud, bribery, and kickback scheme" involving hundreds of store leases dating back to 2010. The defendants (including senior real-estate managers at Mattress Firm and brokers at Colliers International) were running kickbacks from developers and landlords to sign leases regardless of whether the locations made business sense. That's why there were three Mattress Firms on one block.

Layered on top of that, the South African parent company, Steinhoff International, which had bought Mattress Firm in 2016 for $3.8 billion, turned out to be running the largest corporate accounting fraud in South African history. An independent forensic investigation found Steinhoff had recorded $7.4 billion in fictitious transactions between 2009 and 2017, enough that without the fraud the company would have been loss-making across most of that period.

Mattress Firm filed Chapter 11 in October 2018, closed about 700 stores, and emerged from bankruptcy 47 days later. The reason I'm flagging it here is that this is the chain Tempur Sealy just bought. The retailer at the center of the largest vertical mattress merger in American history was, ten years ago, a wholly-owned subsidiary of a multinational accounting fraud whose parent company also presided over a domestic real-estate kickback racket. The chain has cleaned up since. But the fact that the largest mattress retailer in the country has been an asset traded between fraud-adjacent owners for most of the last decade is worth considering before you decide to spend your money with them.

What it looks like when a company actually answers the phone

Here's where my own shopping comes back in. I ended up buying a mattress from My Green Mattress, a family-owned operation in Illinois that makes latex hybrid mattresses with GOLS organic latex, GOTS organic cotton and wool, and a 365-night home trial. The company was founded by a guy whose daughter had chemical sensitivities and who couldn't find a mattress that wouldn't off-gas in her bedroom, so he started making them.

The reason I'm telling you about MGM specifically is what happened before I ordered. Their 20-year warranty has a slat-spacing requirement: the bed frame supporting the mattress can't have gaps wider than 3.5 inches between slats, or the warranty is void. I'd already bought a Room & Board Hudson bed frame. It has 14 slats spread across 80 inches. About 4-inch gaps. Over spec.

I called MGM to ask whether I had a problem. The person who answered the phone actually knew the answer. They ran the math on my specific frame, confirmed the gap was over their threshold, and recommended a bunkie board to bridge it. The whole call took maybe fifteen minutes. I bought the bunkie board, ordered the mattress, kept the warranty intact.

Now imagine making that call to whichever third-party warranty service handles Sealy claims. The big-brand mattress warranties typically void for any stain, require a specific approved foundation, and only honor sagging claims past 1.5 inches. The 20-year warranty on a Sealy Posturepedic is a sales tactic, not a warranty.

What I keep finding

The mattress industry is not a unique offender. The pattern is everywhere now. Three brand names that look like competitors are owned by one company. The largest retailer in the category is owned by the largest manufacturer. Information is deliberately fragmented, prices are deliberately impossible to compare, warranties are deliberately written to never pay out, and the regulatory bodies that exist to prevent this kind of vertical capture are impotent.

What's specific to mattresses is that the category is so opaque, the purchase is so infrequent, and the product is so universal that the trick scales. Everyone needs one. Almost no one buys one often enough to notice the game changing under them.

If you're in the market right now and want the short version: the brands worth your money are small, independent, and owner-operated. A starting list:

• My Green Mattress. Family-owned in Illinois. GOLS organic latex, 365-night trial. The one I bought.

• Spindle. DIY latex layers, around $2,000 for a king, fully swappable/rebuildable. Structural opposite of the rest of the industry.

• SleepEZ. Third-generation family-owned in Arizona. Customizable latex layers under one zippered cover.

• Sleep On Latex. Hand-sewn at their own factory in Niles, Illinois. Single product, no upselling, 100-night trial.

• Charles P. Rogers. Family-owned in New York since 1855. Probably the oldest independent American mattress maker still operating.

• Naturepedic. Family-owned in Chagrin Falls, Ohio. GOTS-certified organic factory, modular construction means individual components are replaceable.

• OMI. Made to order in Colorado. America's first certified organic sleep brand. Expensive

• PlushBeds. Founder-led in California. Vertically integrated, manufactures in their own facility.

The full mattress section of the Brand Ledger goes deeper, including the brands that look independent but aren't. None of these brands spend money on national TV campaigns or showrooms on every corner. Most of them will tell you over the phone exactly what's in the mattress, who made each component, and what the warranty actually covers. The whole field of "mattresses you've heard of" is, with few exceptions, a list of brands you should be skeptical of.

I've started doing this with every category I touch. Backpacks. Power tools. Eyewear. Now mattresses. Pick a category, pull on the thread of who actually owns the brand, and the same architecture keeps showing up. Private equity rolls up the small American companies that built reputations over decades. Conglomerates buy the manufacturers and then buy the retailers. The brand names persist because the brand names are the most valuable asset. More valuable than the factory, more valuable than the people, more valuable than the product. The product is the thing that gets quietly worse while the brand stays the same.

The whole system depends on a buyer who recognizes the logo and assumes the company behind it hasn't changed. Brand equity is rented credibility from dead founders. Knowing who actually owns what changes what you buy, and what you buy is the only signal these conglomerates respond to. A dollar spent at a small family-owned company is a dollar the conglomerate didn't get. Enough of those and the math of rolling up small companies starts to break.

This is enshittification with a longer time horizon. It's the same thing that happened to your favorite app, except it's been happening to your mattress for forty years and you only notice when the new one sags faster than the old one did.

I'll keep pulling on threads. There are a lot of them.

A truck pulls into the alley behind two restaurants. Same truck, same hand cart, same flats of frozen jalapeño poppers walking through two different kitchen doors that share a back wall. Two different menus, two different price-points… the exact same food supplies.

The truck is Sysco. They deliver to more than 400,000 of the ~749,000 restaurants in America. Roughly one in every two. The steak and eggs at a diner in the Texas Panhandle and the steak and eggs at a breakfast joint in northern Maine taste functionally identical because they came off the same pallet at the same distribution center, processed against the same private-label spec, on the same line, by people who never knew which restaurant the boxes were headed to.

This is what the system was built to produce. The same dinner, served to 400,000 different rooms, by people who think they are running their own restaurants.

The truck stops everywhere

Sysco does not just feed independent restaurants. They feed hospitals, federal prisons, military bases, public schools, and the food service companies that supply the cafeterias of the United States Capitol. Fiscal year 2025 closed at $81.4 billion in net sales. The customer count sits at roughly 730,000 across 10 countries, with 337 distribution centers and around 1,719 employed drivers.

The thing people should understand is what those numbers do at the supplier layer. When Sysco moves a spec on a chicken breast, the spec moves on the plate of a restaurant-goer, a public school kid and a federal prisoner in the same week. When Sysco strikes a single supplier deal for frozen seafood, the cafeteria at the United States Congress and the chow line at the Bureau of Prisons end up with the same case from the same boat.

That institutional reach is the leverage point. A corner cut at the supplier layer reverberates through hospital trays, school cafeterias, military mess halls, and the booth at the local diner, all from a single procurement decision in Houston.

What Sysco told the FTC in 2015

In 2015, Sysco tried to buy US Foods (America’s second largest distributor) for $3.5 billion. The Federal Trade Commission sued to block it. Sysco's CEO Bill DeLaney published a public statement defending the deal, arguing the FTC had drawn the relevant market too narrowly. The FTC was, DeLaney said, "ignoring the existence of myriad local suppliers, including broadline companies, specialty companies, cash-and-carry, and club stores" with whom Sysco competed daily. The argument did not persuade. On June 23, 2015, a DC federal judge granted a preliminary injunction, and Sysco abandoned the merger the following week.

Eleven years later, on March 30, 2026, Sysco announced the acquisition of Jetro Restaurant Depot at $29.1 billion in enterprise value. Restaurant Depot is the largest cash-and-carry operator in the United States. Cash-and-carry means a small restaurant owner can drive to a warehouse, walk in, pay wholesale, and load the truck themselves, with no contract, no sales rep, and no delivery minimum. It is the version of food distribution that does not require you to be big to get the chain-restaurant price. The exact category Sysco invoked in court to prove the broadline market was already competitive is now the category Sysco is consolidating.

One of those two Syscos was lying. Either 2015 Sysco was lying to a federal judge about who their competitors were, or 2026 Sysco is lying to investors about what this acquisition does to the structure of restaurant supply.

For the operators who have spent two decades using Restaurant Depot as the wholesale price benchmark against which Sysco contracts got negotiated, the question is academic. Erika Polmar, executive director of the Independent Restaurant Coalition, put it on the record: Restaurant Depot has been "the great equalizer" for independents, a place where a small operator could walk in and pay the same wholesale price as a chain, with no contract and no leverage required.

For two decades, an independent restaurant owner getting squeezed on a Sysco contract could drive to Restaurant Depot, see the real wholesale price of a case of chicken, and walk back into the negotiation with that number. After the deal closes, the wholesale price is whatever the company on the other side of the contract says it is.

Each Worse on Purpose essay names the corporation behind a different category. Subscribe to get the next one in your inbox.

The big three

Sysco was founded in 1969 as a nine-company merger from day one. They were a roll-up from the moment of incorporation, and they have acquired over 100 companies since. Austin Frerick, the antitrust scholar and author of Barons, has the cleanest version of the argument: Sysco grew by buying its competitors, decade after decade.

The implication for the food supply chain is massive. Sysco arrived at $81.4 billion in revenue by buying the regional players, then absorbing the routes, then closing the warehouses, then standardizing the spec, then writing new contracts on terms that left independent restaurants with fewer alternatives every year.

The company that delivered to a small-town restaurant in 1995 was probably a regional distributor with relationships going back decades, a sales rep who knew the chef by name, and a price book that was negotiable. That distributor was acquired somewhere between 2001 and today. The rep who knew the chef got laid off in the integration.

This pattern is shared across the industry. The FTC's 2015 court filing calculated that Sysco and US Foods alone would have controlled 75% of the national broadline market if their merger had closed. They are now back to controlling that market separately, joined by PFG, and a 2025 industry analysis puts the three companies' combined share of the independent restaurant supply chain at over 50%.

Three companies decide what gets delivered to most of America's restaurants. If a restaurant owner does not like the contract Sysco offers, the alternatives are US Foods or PFG. In every case, they are getting squeezed.

How scale produces sameness

Five mechanisms, drawn from Frerick's reporting in The Nation and his Barons chapter on broadline distribution, describe how scale produces convergence at the plate (read Barons - it is a fantastic book on this subject).

First, center-of-plate as loss leader. Sysco wins accounts by undercutting on protein at razor-thin margins, then makes the actual margin elsewhere on the order. Restaurateurs told Frerick the same thing repeatedly: nobody beats Sysco on the steak and the fish. The protein gets you in, and everything else is where Sysco turns the screw.

Second, spec compression. Sysco buys to a spec, and the spec drifts down between contract renewals. The Milkweed first reported on a Pennsylvania diner whose burger patties stopped frying the same on the grill. When the owner finally read the label, he discovered that the patties contained soy protein as filler. What was once 100% beef when he signed with Sysco had degraded without his knowledge. The contract had been quietly rewriting itself for years.

Third, private-label push. Sysco Classic, Sysco Imperial, Block & Barrel, and Arrezzio. The house brands carry better margins for Sysco and lower nominal prices for the operator, so they crowd out branded suppliers at the case level. By 2024, Sysco brand products made up 36% of US broadline cases, and an even larger share of Sysco's independent restaurant sales. The "Chef's Special Wings" at one restaurant and the "House Buffalo Wings" at another are the same wings, out of the same Sysco case.

Fourth, frozen, engineered for durability. The chicken breast on a Sysco truck has been injected with a brine solution of saltwater, phosphates, and binding agents that keeps it juicy after a long freeze. The shrimp has been soaked in sodium tripolyphosphate so it holds water through thawing and arrives looking firm. Every product on the truck is a chemistry problem solved for shelf stability, and the chemistry costs flavor.

Fifth, regional identity collapses. The clam chowder in a New England diner and the clam chowder in a Florida diner come out of the same Sysco can. The biscuits at a Tennessee breakfast joint and the biscuits at a Wisconsin one come from the same frozen case. Regional cuisine, the kind that used to be the reason people drove to a particular restaurant in a particular town, requires regional ingredients and regional suppliers and a chef with the leverage to source both. As Frerick put it, “every independent diner becomes an off-brand Denny's."

Among line cooks, the saying is simpler. “When a Sysco truck pulls up to the loading dock, the kitchen has stopped trying.”

The receipts

Sysco’s enshittification pattern shows up in court records, settlement filings, and government investigations spanning more than a decade. Four of the most egregious are below:

The first receipt is in California, 2014. The state's Department of Public Health investigators found that Sysco had stored 405,859 food items, including raw meat, in 25 unregistered sheds across the state for a total of 23,287 days. The sheds were unrefrigerated. Some held raw meat at 80°F next to insects and rat traps. Sysco's sales representatives loaded the food into their personal cars and drove it to schools, hospitals, and private restaurants. The state collected a $19.4 million settlement, and Sysco employees in subsequent reporting confirmed the same shed practice in Nevada, Washington, Utah, Tennessee, Illinois, New York, the District of Columbia, and Ontario.

The second receipt is older but more revealing of how the spec game works at the plate. In 2006 and 2007, the Florida Attorney General's office investigated grouper fraud across the state. Investigators DNA-tested fish from 24 Tampa Bay-area restaurants and found that 17 were serving substituted species labeled as grouper. Sysco's West Coast Florida subsidiary was the supplier for 14 of them. Sysco settled by paying $300,000 and agreeing to test future grouper shipments, without admitting wrongdoing. State regulators reported that mislabeled grouper had also turned up at the cafeteria inside the Florida State Capitol. Sysco settled, paid the fine, and went on selling grouper.

The third receipt is bigger and more recent. The Outlaw Ocean Project traced Sysco's seafood supply chain to Chinese processing plants using Uyghur and North Korean forced labor, including Chishan Group, Rongcheng Haibo, and Shandong Haidu. Congressman Jared Huffman's office found that more than $200 million in federal seafood purchases over five years are tied to that supply chain. Sysco supplies the food service contractors (Sodexo, KSC, Restaurant Alliance, etc…) that serve the cafeterias and dining rooms of the United States Capitol and Executive Office Buildings. As Huffman wrote in the letter, it is "quite possible" that members of Congress, their staff, and Capitol visitors have unknowingly been served seafood from this tainted chain, in the same building where they could pass the law that would stop it.

The fourth receipt was adjudicated just a couple months ago. A $52 million whistleblower verdict in California Superior Court in February 2026. Five former Sysco Riverside employees, mostly drivers and yard workers, were ordered to load perishable food into trailers operating at temperatures of up to 70°F when the legal requirement was 40°F. The supervisor who led the retaliation against the whistleblowers was promoted to Warehouse Director and remains in that role today. Twelve years after the sheds settlement, the same behavior, the same workaround on temperature controls, and the same retaliation against the people who flagged it.

Sysco won’t change. This is what the system produces continuously, by design, because the cost structure rewards exactly this kind of corner-cutting. Fines, settlements, and now whistleblower verdicts are all priced into the operation. The corner-cutting is worth more than the cost of getting caught.

The algorithm that ate the margin

In fiscal year 2022, Sysco's net earnings rose 159%, climbing from $524.2 million to $1.358 billion on a GAAP basis. That number is in Sysco's own Q4 fiscal year 2022 SEC filing. The same year, Sysco returned $1.5 billion to shareholders.

What was happening at the same time was drastic inflation in food costs as a result of the pandemic. Restaurants were absorbing that inflation, raising menu prices, and watching margins compress. On the Q4 fiscal 2022 earnings call, Sysco CEO Kevin Hourican told investors that the company was using a pricing algorithm that scrapes the market in every region to set customer-specific prices. The algorithm was optimizing for how much extra inflation Sysco could pass through to restaurants without losing the account.

For the operator at the end of the chain, that meant paying the higher prices. For the people working in those restaurants, wages did not rise to match. And Sysco's own drivers, the ones physically moving the food, have seen real wages fall 30 to 40% since trucking deregulation in the 1980s, per BLS data.

Interestingly, Sysco is also a plaintiff in the meatpacker price-fixing litigation, suing JBS, Tyson, Cargill, and National Beef for fixing beef and pork prices upstream. The company's court filings allege that the upstream packers used coordinated pricing to extract margin during the same window Sysco's own pricing algorithm was extracting margin downstream from restaurants. Sysco is a victim of the exact behavior they perfected one rung up the chain. Diners and restaurant workers, as always, are the ones paying the price.

What the deal actually does

The Sysco/Jetro Restaurant Depot deal is set to close in early 2027.

The Independent Restaurant Coalition has filed a petition with the FTC asking the agency to block the deal. Their core argument: Jetro Restaurant Depot was the wholesale price independent restaurants could check Sysco's number against. Once Sysco owns it, there is no number to check against. The price Sysco names is the price.

Sysco is projecting $250 million in annualized "synergies" within three years of closing the deal, primarily through procurement and supply-chain consolidation. In the financial press, "synergies" is the polite word for vendor squeeze and headcount cut. Anyone who has watched a roll-up closely understands what synergy delivery looks like in practice. Regional warehouses get consolidated, salespeople get laid off, supplier contracts get rewritten on Sysco's terms, and pricing flexibility for the buying side gets reduced to whatever the central pricing algorithm decides this quarter.

The 2015 case that blocked the Sysco/US Foods merger turned on a 3-2 commission vote and a Bureau of Competition that was actively litigating broadline concentration. In the current commission it seems that the political appetite to block the deal is meaningfully thinner than it was a decade ago. The pre-merger filings will be made and the hearings will happen, but the most likely outcome is that the deal closes. This will cause the price benchmark for independents to disappear, and the broadline oligopoly will tighten by another notch.

For the operator at the end of the chain, the consequence is dire. The 2015 broadline market had three dominant players plus one functioning cash-and-carry alternative that any small restaurant could walk into and use as a price check. The 2027 broadline market will have three dominant players, the largest of whom now owns the cash-and-carry layer too. The price floor is gone, and the next time an independent restaurant negotiates a Sysco contract renewal, they will have fewer cards to play.

This is what corporate consolidation in the food supply chain looks like when it has metastasized for fifty years. Bland, frozen, shed-meat, forced-labor-seafood food served continuously to restaurant-goers, schoolchildren, federal prisoners, and the members of Congress who could have regulated it.

The cost shows up in something other than money, in the slow decade-by-decade collapse of regional food identity, in the disappearance of the diner that tasted like a place, in the convergence of every menu in every town toward the same private-label pallet.

The Sysco/Jetro deal is the next beat in the food quality collapse. The same truck pulls up to the alley behind two restaurants. Same boxes. Same dinner. The next time you sit down to eat, you'll know what's on your plate.

If you want to eat at restaurants that aren't on the truck, the honest answer is that no good national directory of locally-sourced restaurants exists. The closest thing was the Eat Well Guide and it shut down. What does exist is the supply side. The USDA Local Food Portal lists farmers markets, CSAs, and food hubs in just about every state. EatWild maintains a state-by-state directory of pasture-based producers selling direct.

Once you know which producers operate in your region, the question to ask a restaurant is whether they source from any of them. The menu usually answers it for you. Restaurants sourcing locally name their farms and fishermen on the menu. Restaurants on the broadline truck don't, because their supply chain doesn't have any names worth mentioning.

In much of the country (rural counties, small towns, large stretches of the South and Midwest, anywhere outside a metro with a working farmers market scene) the options thin out fast. Sometimes the broadline truck is the only truck. But where alternatives exist, the shorthand is simple. If the menu names the farm, that's a meal worth paying for.

This is the kind of consolidation Worse on Purpose investigates. The corporate decisions, hidden ownership structures, and quiet quality cuts that make the products and services you rely on get worse over time.

Most of those investigations end up on The Brand Ledger, a searchable database of every brand I've covered across tools, bags, apparel, eyewear, and footwear. Who owns them, what they used to be, and whether they're still worth buying.

→ The Brand Ledger — Every brand, with ownership details and verdicts
→ Approved brands — The ones still making their own stuff

Know a line cook, restaurant owner, or diner who'd want to read this? Forward this to them.

If someone forwarded this to you and you want the next one, subscribe.