Alrighty, today we’re in for a treat from my pal Isaac.

You might remember him from our piece on Buffer’s radical build-in-public playbook.

Well since then we’ve both moved from Sydney 🏝️ to London 💂, still co-work all the time, and still furiously debate whether brands should have a personified logo or not.

Over to Isaac.

— Tom

Hey folks, Isaac here 👋 I write the Brand Chemistry newsletter, unpacking how brands connect with their customers, and run Pistachio, where we help B2B brands build content engines that earn trust instead of just attention.

Today we’re unpacking Notion’s AI agent.

Every software company has now rolled out some kind of AI or agent features, but there’s been public push back on how they’ve been rolled out - even Duolingo’s brand was torched last year.

Yet, Notion have walked away with nothing worse than a few bug complaints, all because of some very considered and clever design choices.

Deals don’t die in meetings. They die between them.

The follow-up nobody sent, the signup nobody researched, the buyer signal nobody saw. Attio’s agentic CRM closes those gaps: revenue agents keep every deal moving between the moments your team is in the room.

Every customer signal becomes your next move - that's how pipelines stop leaking.

Your best ideas shouldn’t live everywhere

You know the drill: ideas in Slack, context buried in docs, and Jira tickets you swear you saw somewhere. That was Maple too.

See how they rolled out Jira Product Discovery to get ideas and prioritisation into one place, plus the month-by-month playbook they used to go from one team’s pilot to an org-wide standard.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

How Notion made AI cute

Chess Move

The what: A TLDR explanation of the strategy

When Notion’s AI is thinking, its eyebrows do a wave. Hit an error, and the whole face comes apart, then puts itself back together.

The team nicknamed it Nosy.

Underneath that face is an agent that works across hundreds of your documents on its own, about as close as software gets to the “AI employee” everyone’s either excited about or terrified of.

In 2025, Duolingo’s CEO said the company would phase out contractors for work AI could handle, on top of cutting 10% of staff. Users who’d held 100-day streaks ended them in protest. Duolingo lost 400,000 TikTok followers in a few weeks and wiped its own accounts.

Notion released something with a similar risk profile: an agent working autonomously across your work. It got criticised for being slow, for hallucinating, and for chasing AI while the basics were still broken.

Software complaints. Not a revolt.

The difference wasn’t the technology. It was a year Notion spent designing a face before anyone saw the agent behind it.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1. Reject the default symbol

Open any AI feature in any product and you’ll see a sparkle icon. It’s shorthand for magic. That’s exactly the problem. Magic means you don’t need to understand it. And when every competitor reaches for the same icon, the icon stops telling you anything about who made it.

Notion’s team named the problem directly. Design lead Ryo Lu said they needed “to do something that differentiate ourselves from the sparkles in the world.” Colleague Linus Lee went further:

Early concepts leant into the magic anyway: wizards in different hats for different tasks and a wand that turned into a pencil or a brush depending on the job. All of it got thrown out, because introducing magic didn’t make the product any easier to understand.

Everyone else designs for how advanced the technology is. Notion designed for how much of it you can understand.

2. Build the character before the capability

In September 2024, Notion’s Head of Creative Rob Giampietro said the quiet part out loud: an agentic future was coming, but “right now, I think we have to make people comfortable with being helped [by AI]… that’s the first step.”

The face designs shipped that month. The agents themselves didn’t arrive for another year.

The character is deliberately minimal: two eyes, a pair of eyebrows, and a nose. “The nose is almost like a baton that conducts your eye,” Giampietro said. Processing shows up as a wave of the eyebrows. An error makes the whole face fall apart, then reassemble.

None of this is decoration. Notion asked Steven Sinofsky, who ran the Office team that shipped Clippy, whether he’d give an assistant a face again. He pointed to a Stanford study, The Media Equation, which found people generally treat technology as “right” by default. Give them something softer, and “they stop blaming themselves for mistakes and are more open to questioning back and forth.”

Notion pushed the idea further than a static drawing: you can rename your agent and dress it in accessories. A character you customised yourself is hard to be afraid of.

3. Let the language do the rest

Notion never says “AI employee”. You won’t find the word “headcount” anywhere near the feature. They won’t even call it an “assistant”, the same instinct that has Apple refusing to call the Vision Pro a “VR headset”. Name the category, and you import every fear that comes with it.

Cofounder Ivan Zhao said he wanted the product to feel less like productivity software and more like playing Tears of the Kingdom: easy to build with, endless complexity hidden underneath. Hide the difficulty. Keep the user inside the work.

Compare that to Duolingo, which said the quiet part out loud: AI would replace contractors. It paid for that in cancelled streaks and 400,000 lost followers.

Same underlying technology. Opposite framing. Opposite outcome.

Every AI agent asks the same question of the person using it: what is this doing, and what does it mean for me? A sparkle answers with mystery. A pair of eyebrows answers with a colleague, thinking it through alongside you.

That’s all from me!

If you want more of this, brand strategy explained through the actual decisions companies made, subscribe to Brand Chemistry here and follow me on Linkedin and X.

And if you’re a B2B brand trying to earn the kind of trust Notion just did, that’s the work we do at Pistachio. You can book time to chat with me here.

Rabbit Hole

The where: 3x high-signal resources to learn more

(Re)designing Strategy Breakdowns ↗

[6 minute read]

The raw, behind-the-scenes scrapbook behind the Strategy Breakdowns rebrand.

→ The initial moodboard and vibe sliders

→ The concepts that got thrown away

→ The iteration loop with Jade (product designer @ Atlassian) who brought the brand to life

Clippy walked so Notion AI could run ↗

[8 minute read]

Notion’s own AI face interviews Steven Sinofsky - the exec who shipped Clippy - on the Stanford research behind friendly interfaces, and why Clippy was “early and wrong”.

The most interesting part: Why softer interfaces make users question the interaction instead of blaming themselves

The life and death of Microsoft Clippy ↗

[9 minute read]

The full cultural history:

• Picked from 250 candidate characters and focus-grouped as “most trustful and endearing”

• Hated in practice

• Retired in 2001

• Reborn as a meme and, per the 2025 update, a protest symbol.

Also contains illustrator Kevan Atteberry’s original designs - super cool!

💬 Do you like reading guest posts? Why or why not? Reply to this email and let us know!

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

When I started Strategy Breakdowns, I worked at Atlassian.

Strategy and Business Operations (”SBO”) in the Jira integrations team.

Last day on the job 🥹 Always loved those plants hanging from the air con ducts

It was a dope place to work for lots of reasons. One of the big ones was the hardcore commitment to “dogfooding” - using Atlassian’s tools to build Atlassian.

Every single employee lived inside of Atlassian tools. Every project, line of code, strategy doc happened in the same stack we were actually shipping to customers. Jira, Confluence, Bitbucket, Trello, Loom, etc.

It’s not like Atlassian makes accounting software that only finance could use - the company literally makes dozens of tools for enterprise collaboration, so everyone was a target user of some kind.

I’ve wanted to write this piece for a while.

With all the wild AI coding / agentic collaboration / automated development stuff they’ve been shipping recently (I left a few years ago, right as the first AI features were shipping externally), I had to reach out to the team to get a behind-the-scenes deep dive on the books.

Here’s the latest. Enjoy.

— Tom

While you slept, your website ran 3 experiments.

A normal night on Ploy. Its agents watch your site and your competitors around the clock, then act: fixing SEO, testing landing pages, spotting the enterprise visitors worth chasing and drafting the outreach - all before standup.

We know because we've been testing it on ours - our new media kit shipped in an afternoon. Built by Webflow co-founder Bryant Chou, it already powers growth at Hex and Clay.

Strategy Breakdowns readers: use code Tom-Alder at checkout.

Your best reps shouldn't be doing data entry.

Every hour spent updating fields is an hour not spent selling. Attio flips the ratio: its revenue agents handle the busywork - account research, follow-up drafts, buyer signals - so your team stays on the deals.

It's the agentic CRM built to run alongside your team, not add to their admin.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

How Atlassian builds Atlassian

Chess Move

The what: A TLDR explanation of the strategy

AI was supposed to make work faster.

And it definitely has, at an individual level.

But at a team level? It’s complicated…

In Atlassian's State of Teams 2026 report, they found something surprising.

89% of executives said AI had sped up work.

6% could point to real ROI.

Everyone is sprinting → nobody is syncing.

Atlassian calls the gap the “fragmentation tax”:

It’s duplicated work, misaligned priorities and general coordination chaos from rapid individual AI adoption.

It’s worse now that AI floods teams with more output than they can absorb.

Across the Fortune 500, it’s costing $161 billion a year (that’s ~$400M a day)

Plenty of companies are racing to sell the fix. But Atlassian has a unique advantage that allows them to solve it faster than anyone else: 14,000 employees, inside Jira, Confluence etc every day, pushing AI to its limits.

So instead of building for a customer they have to imagine, they build for themselves:

→ Rebuild Jira, Teamwork Collection and the Teamwork Graph around AI agents to give them the same context and coordination Atlassian has spent 24 years giving humans
→ Run it on the hardest enterprise customer on their books, themselves, until the numbers move
→ Ship the proven tools externally, then push the same pattern past engineering into IT, design, ops and marketing

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Dogfooding 101

“Dogfooding” is a term that comes from "eating your own dog food".

Basically using your own product.

Do it well and you get the fastest feedback loop there is - your own staff hit the bugs, friction, and missing features before a paying customer ever does.

Plenty of companies dogfood.

But for Atlassian, it’s a way of life.

Atlassian sells enterprise collaboration software, and it’s an enterprise itself.

~14,000 employees added over 24 years since its 2002 founding.

That's 24 years of the exact procurement cycles, security reviews, cross-team dependencies, and legacy workflows its own customers wrestle with every day.

A 40-person startup selling enterprise AI has to imagine what enterprise scale feels like.

Atlassian folks can just walk down the virtual hall.

I witnessed this firsthand when I worked there, and it is like clockwork.

• They use Jira at scale to manage the development of every single Jira feature.

• They have an internal Confluence instance called ‘Hello’ that’s been running since Confluence first existed, with early feature decisions, founder strategy docs, and defining product debates still living in the comments.

• They ship and automate code using Bitbucket.

• They send Looms as the default for too-long-for-Slack-but-lets-save-ourselves-a-meeting communications.

And employees aren’t just the power-users. They're the beta testers too.

All from my ‘Atlassian hall of fame’ screenshots folder. Wish I took more while I was there tbh!

Every day, new experiments and updates are tested by thousands of employees.

Research teams run internal interviews.

Data scientists track real retention curves.

Engineers go bug-bashing at scale.

The result: by the time external customers touch new features, they’re already proven.

2. Jira as the system of record for AI agents

So how does dogfooding solve the ~$400M a day fragmentation tax?

Well, Atlassian's own engineers live in Jira, Confluence and Bitbucket all day.

But even they had their own version of the fragmentation tax, too.

In a DX study Atlassian cites, engineers' AI usage is up 65%, while developer velocity has risen only around 10%.

The gains leak out at the seams:

• agents missing project context

• unclear source of truth for planning

• AI slop (PRDs, code, docs, etc) cleanup

• bottleneck shifts from writing code to QA

Their fix was to turn Jira into the system of record that orchestrates AI agents across the entire SDLC, and not just code.

With humans and agents coordinated in one place, agents assigned work straight from a ticket, and every action visible on the board.

Jira Planner turns a rough idea into a technical spec

Jira Coding Agent turns a work item into an agent request

And the cool thing is - they actually document their research in public, publish the real data, and explain how the learnings get folded into the customer-facing products:

• 80% fewer engineering hours lost to maintenance chores, roughly a full engineering week back per team, per month

• 52% of security vulnerabilities resolved automatically, with issue cycle time cut in half

• 19% more pull requests, and 2 to 3 hours saved per developer, per week.

3. The same pattern, everywhere else

Here's what makes AI-native Jira an org-wide step-change strategy rather than an isolated dev-centric shift.

Atlassian knew this fragmentation tax - the gap between individual speed and team output - showed up far outside just engineering:

• ops

• design

• marketing

• everywhere teamwork is required

Engineering was simply where the instrumentation exists to prove it first.

But since all types of Atlassians are power-users of Atlassian products, that same dogfooding loop can run across all their non-engineering teams too:

→ Event Marketing ran the entire planning for Team '25, Atlassian's flagship conference, on Jira and Confluence

→ Creative Studio ran its largest-ever internal Jira campaign, with Rovo pulling work items straight from the project plan

→ Community and Learning Ops unified the work of 6 teams around a single source of truth

Customers are living the same story too.

→ Booksy uses Jira Product Discovery to put product, engineering and business teams on one shared view.

→ Procore, on the same setup plus Rovo, saves 23 hours per month per employee.

→ Tempo builds its own agents in Rovo Studio to automate reporting and finish tasks up to 30% faster, without writing a line of code.

The unit of work changes (a campaign or a creative brief instead of a bug ticket or a pull request).

But the graph underneath remains.

Which means Atlassian doesn't have 1 dogfooding loop. It has one per department, all feeding the same products.

Rabbit Hole

The where: 3x high-signal resources to learn more

The State of Teams 2026 ↗

[25 minute read]

Atlassian's Teamwork Lab surveyed 12,035 knowledge workers and 172 executives to put a number on why AI isn't paying off at the team level.

→ Where the $161B fragmentation tax actually comes from

→ The 3 pillars that separate the top teams

→ Why only 14% of teams are getting AI right

Evolving Jira for AI-native software development ↗

[8 minute read]

Atlassian's own blueprint for what comes after AI code generation.

Plus, plenty of videos demo’ing how these things actually work at scale, live. More show, less tell.

How Atlassian's own teams run on Jira, Confluence and Loom ↗

[15 minute read]

This one’s super cool.

A running collection of Atlassian's own teams showing their real setups, and not just sales materials.

→ Event Marketing, Creative Studio, Community and Learning Ops and more

→ Actual workflows, screenshots and lessons learned

It’s build in public meets enterprise, the closest you'll get to sitting inside Atlassian HQ without the swipe-card.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

We’re about 2-3 months away from looking like a COMPLETELY different business.

2 massive launches planned.

1 will lift our revenue ceiling by 3-4x.

1 will fundamentally change the business model.

Both have been dreams of mine for years. Wild to see them come to life.

Can’t wait to share more very soon.

Reply with your guesses 👀

— Tom

You read how Attio works. Now put it to work.

Our recent Attio breakdown covered the strategy; this is the product. Attio is the CRM where agents run alongside your team: researching accounts, drafting follow-ups, surfacing what buyers are saying, and moving work forward on their own.

Every customer signal turns into your next move, which is what go-to-market velocity actually looks like in practice. It's free to try.

Is your roadmap a list of features or a strategy for outcomes?"

Most product teams have plenty of feedback – it’s just scattered across Slack, call notes, and support tickets. Jira Product Discovery (JPD) brings that evidence into one place so you can stop debating opinions, and start prioritizing based on customer data.

Connect your roadmap directly to Jira and show exactly why you’re building something.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

Why Airpods are actually a subscription

Chess Move

The what: A TLDR explanation of the strategy

Bloomberg Intelligence put AirPods at roughly $22 billion in revenue for 2024, up 22% year on year, and around $18 billion a year every year since 2021.

Spin that out as a standalone company and that would make rank AirPods #199 on the Fortune 500.

Apple's Wearables, Home and Accessories segment boasts $35.7 billion in FY2025 and would be #129.

And the product only launched in December 2016.

It’s easy to think AirPods look like a one-time hardware sale, but data shows they’re more like a subscription.

You pay $250 once → then the atrophy begins…

→ The battery degrades hard after about 2 years of daily charging.

→ It's glued in with no user-serviceable swap

Every generation is "essentially disposable by design."

AirPods are a one-off purchase, but are engineered to recur.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Engineered to expire, not to break

The lithium coin-cell in every pair loses meaningful capacity after about 2 years of daily cycles, and Apple has glued it in since generation 1.

scene of the crime.

You can't cheaply fix the one part that wears out because Apple's battery-service price sits awkwardly close enough to a new pair that most people just re-buy.

And people do re-buy, reliably.

Compute the cost for yourself and ask - “would I pay for a new battery after 2 years or just upgrade?”

CIRP found AirPods sit on a shorter replacement cycle than the iPhone itself, driven by exactly this mechanism:

1. batteries wearing out

2. buds getting lost (guilty).

AirPods are now one of the most frequently reordered items on Amazon - far from an occasional top-up.

A one-off sale essentially becomes a recurring one, without it ever being called (or considered) a “subscription”.

2. The replacement always stays in-house

So a dead battery forces what gets replaced.

But lock-in decides what it's replaced with.

Path of least resistance is another pair of AirPods - it’s the ecosystem argument we’ve all heard before:

• one-tap pairing via the H-chip and a protocol that’s closed to 3rd party devices

• Handoff across every device (Macbook, iPad)

• Spatial audio

• Find My

Ditching all of it for Sony or Samsung means giving up the whole ecosystem experience, not just the earbuds, and most people never even weigh the trade-off.

Above Avalon estimated 61 million pairs sold in AirPods' first 3 years, with the installed base clearing 100 million people by 2021.

Each one now sits inside a repeating cycle where the only replacement is another pair of AirPods.

That's the flywheel.

Lock-in retains users and also re-sells them the same category every time the battery gives out. And by the time their battery dies out there's 1-2 generations' worth of upgrades that users can jump to.

3. How to sell a product that is “solved”

Honestly. If we zoom out, AirPods stopped changing years ago.

AirPods Pro 1 landed ANC in 2019. (huge)

AirPods 3 added spatial audio in 2021. (also quite huge)

But every generation since has iterated at the margins

→ AirPods Pro 3 (2025) advertises 2x more noise cancellation than Pro 2

→ New generations are listed as "the most secure and best-fitting AirPods ever"

And the "what's new" keeps getting more niche.

→ An FDA-authorised hearing aid mode (for users with hearing loss).

→ A heart rate sensor in Pro 3 (fitness users).

→ 5-language Live Translation (travellers).

→ Sleep-triggered auto-pause (sleep-with-AirPodders).

“Best Fitting AirPods Ever”?

Each one is real and well-built - each one is just engineered for a narrower slice of the buyer base than the last.

None of it is why someone buys their first pair of AirPods.

But all of it exists to give someone who already owns a pair a few more plausible reasons to buy the next one right around the time that battery starts to show its age.

Rabbit Hole

The where: 3x high-signal resources to learn more

AirPods Pro 3 Teardown: Still a Tragedy ↗

[8 minute read]

This one’s obviously an opinionated teardown written by the repairability community, but from a strategy/moats PoV it’s quite interesting

The new heart rate sensor is packed in so tight iFixit says there's no realistic way to replace it on its own - and this generation even glues the charging case shut, so the battery in there is just as unreachable as the ones in your ears.

"Essentially disposable by design."

Cook: wearables is almost a Fortune 300 company ↗

[3 minute read]

"The wearables business is the size of a Fortune 300 company."

Tim Cook said it at Apple's February 2018 shareholder meeting - back then, #300 on the Fortune 500 was Alcoa, doing about $9.3 billion a year.

By the end of that same year, Cook had already upgraded the line to Fortune 200.

AirPods sales topped $18bn, more than all of Nintendo ↗

[4 minute read]

Bloomberg Intelligence's math: AirPods alone did $18 billion in a year - more than Nintendo, more than X, more than eBay or Airbnb's entire annual revenue.

Bloomberg's own follow-up call: at this pace, AirPods overtakes the iPad before the decade's end, becoming Apple's third-biggest product behind iPhone and Mac.

Where the viral "bigger than Nintendo" number everyone quotes originally comes from.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

Dear Strategy Breakdowns readers,

What’s the worst part of the product?

Bugs

Broken links

Pages that don’t load everything

Automated emails that feel like they come in the wrong order

→ Anything you’ve noticed, across any touchpoint (website, emails, social, etc), that we should fix.

I’m going to spend the next 60 days in the weeds making SB 10x better.

Where should I start?

— Tom

The best-run companies all made the same hire.

One message in Slack or Teams, and marketing gets a board-ready Google Ads + Stripe revenue report. Engineering gets a PR review cross-referenced with its Linear tickets. Ops gets three vendor contracts summarised before standup. Same colleague, every department.

Viktor works right inside Slack and Teams, connects 3,200+ tools, and 60,000+ teams already run this way.

(Strategy Breakdown readers get $100 in credits)

Give every caller a human-sounding agent

ElevenAgents brings emotional intelligence to voice AI - it reads tone and responds at the right moment.

• Adapts in real time with Eleven v3

• Smarter turn-taking = smoother conversations

• Nuance that scales across 70+ languages

The result → agents that respond in the right tone.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

Attio's agentic bet

Chess Move

The what: A TLDR explanation of the strategy

In April 2025, we called Attio an "AI-native CRM".

From auto-generated database fields, to sales calls transcribed straight into deal context - all wrapped in the cleanest interface in the game.

Our punchline was:

"Sales Data gives AI knowledge. But tooling gives AI limbs."

A year (and a $52m Series B) later, Attio is betting that the next generation of knowledge workers will expect every tool they touch to be agentic.

Tools that meet you where you already work, instead of adding steps to your workflow.

Tools that predict your next move and act automatically, instead of waiting for your prompt.

The thing is, CRMs have never worked like that.

There's always been a translator between you and your data:

A sales ops hire, a data engineer, an implementation partner.

As AI rolls into the sales org (and beyond), a shift is needed.

Where users build and query their CRM by just…asking it.

And then setting up a Claude Routine to “run this report every day at 9am”.

And then an automation trigger for “every time we receive an email asking about our services, research the enquiring company’s funding, revenue, team size, and social following, then populate this tracker with what you find”.

The implications of agentic software are huge.

My team, my agents, and myself are obsessed daily-active-users of Attio. Let’s jump into it what the cutting edge of agentic GTM actually looks like.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  The CRM that answers back

When we last wrote about Attio, its AI worked inside the sales records.

Twitter tweet

→ Estimating ICP fit or ARR for individual companies

Twitter tweet

→ Pulling in context from calls with prospects

Attio’s new Ask Attio flips it.

As well as using AI to enrich data, users can now ask AI to use their data.

Ask about any deal, get briefed before any meeting, pull account history without knowing where anything lives.

And it doesn't just answer your questions. It acts - updating records, drafting emails, building entire workflows from a plain-English description.

Not only does this save time navigating a UI to understand and update records.

But it also expands the TAM that Attio can sell to within an organisation.

Picture this:

A 100-person startup with 35 people in sales.

The other 65 (engineering, marketing, product, finance, etc) have always been effectively locked out of a CRM customised beyond recognition, too nervous to click around in case they break something live.

So when they have questions, those 65 people:

1. Ping a rep

2. Who is taken out of flow

3. Who then has to navigate the UI to find the data

4. Who then has to interpret the data to answer the question

5. Who then reports back

Ask Attio cuts out all 5 steps.

The finance lead is empowered to ask → “what’s our Q4 revenue forecast looking like?”

The delivery lead can determine → “what did we promise the client on that call?”

The leadership team can build → “a custom sales dashboard covering just our UK-based enterprise prospects”

And the RevOps lead can skip the workflow builder entirely.

Describe the automation in a sentence - "when a deal hits negotiation, alert the account team and create a renewal task" - and Ask Attio assembles it on the canvas in real time. Trigger, steps, logic. Review it, hit publish.

A build that used to mean an engineer, a ticket, and a week of waiting.

Founder Nicolas Sharp laid out this thesis 2 years ago in our interview:

2. The CRM leaves the building

CRMs used to force you to build a second product just to use the first: dashboards, pipelines, custom views.

There was a whole application layer that existed because clicking was the only interface available, and it took legitimate specialists to build anything useful.

Now, with the Attio MCP, the interface is your Claude Code / ChatGPT / OpenClaw.

Attio's MCP server makes the same workspace readable and writable from wherever you already work.

The layer underneath is what Attio calls Universal Context:

Every record, conversation and signal kept in a form agents can navigate, always in sync with the data itself.

Ask Claude to "Log this deal" → it writes the record directly into Attio.

"Show pipeline by stage" answers the question.

Want a throwaway dashboard to screenshot and put in a single email ? Just tell your agent to build it

No more sales ops tickets required.

Keep sending in feedback on the infographics! Iterating daily, trying to make them dope. We reply to every email :)

When we wrote about Attio's AI Strategy in April 2025, the ceiling was directing its agents to research 1,000s of companies while you moved on to your next task.

Impressive, but sequential:

1 instruction

1 agent

Inside the app.

The MCP removes the ceiling and unlocks agent swarms.

Any number of agents can now work the same workspace at once, from outside it, running on whatever schedule or trigger your harness desires:

1 monitoring deal activity.

1 prepping your meetings.

1 flagging renewals.

None of them waiting on a human, or on each other.

Ask Attio and the MCP are 2 doors into the same thing.

Ask Attio lives in Attio’s UI.

Attio MCP lives in Claude / ChatGPT etc

They share the same power, and both doors open onto the same Universal Context.

Attio meets the user wherever they are.

Twitter tweet

3. Strategy Breakdowns runs on it

We're not just writing this from the outside.

Every $ of Strategy Breakdowns' sponsorship revenue is touched in some way by various custom AI skills built on the Attio MCP.

Let’s take our /brief skill. Every morning it:

→ Pulls every active deal straight from Attio

→ Reads the real Slack, email, LinkedIn and WhatsApp conversations attached to each one

→ Works out what the ‘Stage’ field hasn't caught up to yet: who's waiting on a reply, who's gone quiet, who's already agreed to something

Every deal, read and reconciled in parallel, before I’ve opened my laptop.

Our /brief skill file would literally fall apart without Attio.

/reply goes even further.

→ Claude drafts the strategic response to any sponsor message (applies frameworks, logic, principles from my favourite business books)

→ Updates the Attio record itself with the latest context

→ Deal status gets updated if we sent a proposal or closed the deal

Our CRM is now always up to date. And we don't even have to open it.

Once again, Nicolas Sharp already predicted exactly this in our 2024 interview:

Turns out you might not even need to log in at all.

I just fire up a skill from wherever I am, or look at a dashboard, and everything stays live up-to-date.

One of our many internal dashbaords - live sales data, pulled directly from Attio. What a time to be building.

Rabbit Hole

The where: 3x high-signal resources to learn more

GTM Atlas ↗

[30 minute read]

Honestly one of the coolest pieces of content marketing I’ve seen this year. Essays “from the trenches” from real operators at Lovable, Anthropic, Linear, Stripe and more.

Almost like a GTM-focussed Strategy Breakdowns but written from the perspective of those creating + executing that specific strategy.

Genuinely useful. Steal the ones that apply to you.

Introducing Universal Context ↗

[8 minute read]

Attio's engineers on the layer that makes everything above possible.

Legacy CRMs bolt a vector database onto the side and pray it stays in sync. This is the story of building context directly into the system of record, so an agent never reads stale data.

The engineering receipts behind takeaways 1 and 2.

Reinventing CRM with AI: A Conversation with Attio CEO Nicolas Sharp ↗

[11 min watch]

Straight from the investor who wrote the $52m check.

This is the version of the pitch reserved for people deciding whether to fund the next 3 years of growth.

Phenomenal piece of content.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

Most companies I break down are machines. Today's is closer to art.

No funding rounds. No growth team. No IPO dreams.

Lofi Girl started as 1 guy in France with a quirky aesthetic, a playlist, and an animated loop of a girl studying.

10 years later it's a record label, a games studio, a café in Paris, and one of the most fiercely loved brands on the internet.

I love indie businesses like this. Unique idea, taken very seriously, built with craft and soul, executed well for years.

1 of many inspirations for Strategy Breakdowns tbh. Lmk if you want to read about more of these internet gems.

— Tom

P.S. My personal Lo-fi playlist, for on-brand vibes while you read. Includes many of my all-time favourite instrumental hip-hop artists like J Dilla, Nujabes, Madlib, FKJ, Clams Casino.

Take your meeting context to new places

Run complex work through Claude or ChatGPT? You know the drill: feed it docs, then copy-paste your meeting notes for context.

Granola - the AI notepad - now skips that with its new MCP integration, connecting your notes straight to your AI app. Ask Claude to review last week's calls and update your CRM, or have ChatGPT pull tasks into Linear - no copy-paste.

Built for engineers, PMs, and operators who want their AI to actually understand their work.

(Strategy Breakdowns readers get 1 month trial - use code BREAKDOWNS)

Your AI feature is getting used. But is it paying off?

Most AI features don’t fail because of pricing. They fail because the AI is not valuable enough to justify what it costs to deliver. Usage grows, costs scale, and margins start disappearing.

See the patterns behind AI features that actually pay off: the pros and cons of popular pricing models, the traps that thin your margins, and practical tips from Kyle Poyar, founder of Growth Unhinged, and Aakash Gupta, founder of Product Growth, based on their work with teams shipping AI at Atlassian, Bolt, Salesforce, Cursor, and dozens of other companies.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

How Lofi Girl owns lo-fi

Chess Move

The what: A TLDR explanation of the strategy

Picture a stadium packed with 40,000 people.

None of them know each other.

None of them are even speaking.

They're all there for the same thing:

A looped animation of an anime girl studying at a desk, with chill lofi beats playing.

That's an average Tuesday for the Lofi Girl youtube livestream.

IYKYK

The obvious explanation is that lo-fi hip hop is good background music.

But ~15.85M subscribers, 2.5 billion total views means there has to be more to it.

There is. Much more. Let’s dig in.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Co-presence makes content 3D

Open Lofi Girl and you're sitting in a virtual room with 1000s of strangers, all working in silence.

The music is pretence for the real product → The co-presence.

Dr Steven Gamble, a music researcher at Oxford's Digital Flows, spent the pandemic reading Lofi Girl's comment section and turned it into an academic paper.

Quite the heartwarming read

What he found was a genuine community of users:

• Greeting new joiners, "welcome friend, come for the music, stay for the vibe"

• Reminding each other to "drink enough water and have regular meals today"

• Assuring exam-stressed strangers, "we'll get through this, keep going!"

In 2023, ~40,000 people were connected to the livestream at any given moment.

This sense of community was clearest during the 2 false copyright claim takedowns.

In February 2020 and July 2022, the stream went down, and the global community responded en masse with hashtag campaigns, fan art, public pressure on YouTube, and emotional posts about feeling like they'd lost a friend. (#FreeLoFiGirl)

The music didn't go anywhere. You could still find their playlists on Spotify and other Youtube channels.

They were grieving the place - not just the music.

The stream had been running continuously for 2 years, 4 months and 18 days when it was cut. The archived version is a single video 20,843 hours long, with 668 million views.

Didn’t know Youtube put in comma’s for hour counts that long…

Users had lost somewhere they went every day for years.

Brands typically compete for foreground attention.

Lofi Girl wins by being deliberately ignorable - in the background.

Background-first content has a different physics:

Low effort to consume → infinite session length → co-presence at scale

2. Brand the category

Lo-fi hip-hop existed for over a decade before ‘ChilledCow’ (original name) opened a YouTube channel. Even multiple lo-fi livestreams already existed on YouTube by 2013.

Dimitri (the creator of the livestream) claimed the lo-fi x anime aesthetic even if plenty of others got to the sound first.

The original visual was a copyright-unsafe loop of Shizuku from Studio Ghibli's Whisper of the Heart.

When that became untenable in 2018, Dimitri, commissioned Colombian-French artist Juan Pablo Machado to design a bespoke character: a studious anime girl at a wooden desk, an orange lamp, a rainy window looking out over Lyon's La Croix-Rousse neighbourhood.

He called her Jade.

New Version Left, Original Version Right

Within 3 years, that single image had locked in the visual default of an entire genre.

Every other lo-fi creator now operates downstream of Lofi Girl's aesthetic.

→ It’s the Calm move (didn't invent meditation, became its default app).

→ Or the Red Bull move (didn't invent energy drinks, branded the category).

→ Or the Liquid Death move (water in a can, but with an identity nobody had bothered to give it).

The category was sitting ownerless. ChilledCow showed up with a character, a colour palette, and a name.

Expanded into Synthwave, another 40 year old genre. Here’s one of my faves.

3. Own the catalogue

The stream isn't the business → It's the shop window.

The streams stay live, permanently, which deliberately gives up the mid-roll ads that would pay out to most YouTubers

Ads interrupting the music would break the whole thing, so Lofi Girl instead created a model that put’s YouTube as the entry point to the community, rather than a real revenue line.

The real business is a record label.

Which means the distribution channel and the rights-holder are the same company. Very powerful, and quite unique.

Lofi Records released 156 EPs and albums plus 3 compilations in 2021 alone.

1,453 tracks.

Roughly 3 releases a week.

Artists get paid monthly, not quarterly - rare enough in music to be a recruiting pitch on its own.

In Feb 2024, Warner Chappell Music France picked up publishing admin for most of the roster.

Here’s how the loop works:

40,000 people are sat in the room at any moment → the room plays music the company owns all the masters to → those tracks land on the Spotify playlists → 1.3 million monthly listeners → streaming revenue and massive artist distribution → which attracts more artists to play in the room.

It's also why the 2022 takedown got reversed within 3 days. Every track in the stream was released through their own label. They were being accused of pirating themselves.

The team has ~30 staff now, across an in-house animation studio (hand-drawn only, zero AI policy), a games studio, radio stations, vinyl and plushies, a physical Lofi Café opening in Paris, and licensing/brand collabs with Fortnite, LEGO, Ubisoft (Assassin's Creed Shadows), Renault, Chess.com, Youtooz figures, and an esports sponsorship.

What a cool business.

Rabbit Hole

The where: 3x high-signal resources to learn more

Beats to Quarantine to ↗

[25 minute read]

A music academic at Oxford spent the pandemic reading Lofi Girl's comment section and turned it into a peer-reviewed paper.

It covers what an "imagined community" actually looks like at 3am plus the line between background music and shared infrastructure.

A must read for the more scholarly breakdown readers among us. I know you’re out there.

The 13,000-hour video YouTube accidentally created ↗

[5 minute read]

The 2020 takedown that turned a 548-day livestream into the largest single static asset in Youtube history.

Turns out it was a “fat-thumb” moment on Youtube’s end, but what it revealed about the community is even more interesting.

Interesting case for an always-on asset could beating a thousand weekly uploads.

Lofi Records lands a Warner Chappell deal ↗

[3 minute read]

159 releases. 2.23 billion Spotify streams. ~€5.5M in revenue. One year.

That's what Lofi Records did in 2021 alone, before Warner Chappell France showed up to take over the publishing admin.

An interesting endgame for creator brands, showing how a major signed the indie.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

What you think of when you hear “EV” says a lot about what type of nerd you are:

• Expected Value = gambling/trading nerd

• Enterprise Value = finance nerd

• env var = programming nerd

• Exposure Variable = photography nerd

• Electric Vehicle = hardware nerd

Shoutout to all the legends dropping feedback on our recent run of hardware breakdowns 🥹 you know who you are

• 🎯 The reason every DJ booth looks the same

• 🎯 How Sony sold 400 million Walkmans

• 🎯 How Logitech turned a $20 commodity into a $100 status symbol

• 🎯 DJI used GoPro to kill GoPro

I think there’s a little bit of hardware nerd in every software nerd. Our twisted twin.

Today’s is a banger - you’ll leave understanding why, seemingly out of nowhere, you’re probably seeing more of these than Teslas in the last few years.

— Tom

Security backlogs are the new tech debt

CVE disclosures could top 50,000 this year - and AI coding tools are adding to the pile. Atlassian's new playbook shows how their own team resolves 52% of vulnerabilities automatically and cut issue cycle time in half with Jira's Coding Agent.

Inside: what to automate, what still needs human review, and how to quantify the hours you claw back.

You don't need more AI advice. You need AI that does the work.

Most tools tell you what to fix. Viktor fixes it - audits your ad spend and pauses the waste, groups duplicate bugs into scoped Linear tickets and opens the fix as a PR, builds the revenue dashboard and ships the link.

One AI employee in Slack or Teams, across 3,200+ tools. 50,000+ teams have made the hire. SOC 2 compliant.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

The battery shop that beat Tesla

Chess Move

The what: A TLDR explanation of the strategy

BYD beat Tesla in quarterly EV deliveries for the first time in Q4 2023 (526,409 vs 484,507).

Then they took the full year clean in 2025:

2.26 million EVs sold, up 28%, against Tesla's 1.64 million, down 8%.

But most explanations of BYD start in the wrong decade.

To really understand we need to go back.

Before any of the following existing:

• A US/China Trade War

• Tesla

• Electric Vehicles

In 1995, BYD was a 20-person battery workshop in Shenzhen, hand-building lithium-ion cells.

Founder Wang Chuanfu couldn't afford the automated production lines Sony and Sanyo were running in Japan.

So he built manual jigs instead.

Basically cheap tools that let low-cost labour produce the same quality a robot would.

It meant that BYD reportedly could build battery cells for $1.30 each.

That same cell would cost Sony or Sanyo $4.90.

3 decades later, that same instinct, own the production rather than outsourcing it, is why BYD builds:

• It’s own batteries

• Its own semiconductors

• Its own motors

• Even the cargo ships that get its cars out of China

Crack open a BYD Seal (their rival to the Tesla Model 3) and you’ll learn something crazy.

Roughly 75% of BYD’s parts were made in-house.

A Tesla Model 3 (built in china) was 46% in-house.

BYD built a cost structure so low it could start a price war with Tesla (often 15% cheaper for comparable models)… while maintaining an operating margin (4.8%) more than 3x Tesla’s (1.4%)… while still growing profitably.

Here’s how they pulled it off. Buckle up.

Did someone say: SB-branded social-native infographics?

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Own the entire stack.

BYD doesn't buy its batteries.

It designs them + builds them

It even mines the lithium that goes into them.

The Blade Battery, BYD's flagship cell, uses lithium iron phosphate chemistry instead of the nickel-cobalt-manganese mix most rivals rely on.

So what.

Well It’s far cheaper, more thermally stable, and most importantly entirely BYD's own patented design.

Charging a car in the time it takes to read this newsletter wasn’t on my bingo card.

Power semiconductors are the same story.

BYD Semiconductor, founded in 2004, now controls 28.9% of China's power-module market. 2x the market share of it’s next competitor, Infineon at 14.5%.

Motors, too.

Since January 2024, BYD builds and operates its own car-carrier ships, with the 9,200-vehicle BYD Shenzhen, currently the largest car carrier in the world.

This is a battery, car, semiconductor logistics company

Since inception, BYD has obsessed over owning as much of the stack as possible to preserve control and quality.

With assets as large as carrier ships to as small as windshield wiper fluid.

2. Beat Japan on cost, not on patents

Wang Chuanfu wasn't a car guy. But he was a metallurgist (new word unlocked: scientist who studies metals), with an education from the Beijing General Research Institute of Nonferrous Metals.

Japan's battery makers ran automated production lines that cost 10s of millions of dollars.

BYD couldn't afford that, so it didn't try.

BYD stands for “Build Your Dreams”. Rumour is that Wang named it that to show up at the top of exhibitor lists at conferences.

Engineers broke each automated step into a manual one. No matter how small, or complex, BYD would reverse engineer it.

Step-by-step they built cheap custom jigs that let low-cost labour hit the same tolerances a robot would.

HBS even wrote a whole case study around this in 2006, before BYD had sold a single car, arguing that they were primed to enter the automotive industry.

3. Turn the cost edge into a price war Tesla couldn't win

Owning your supply chain is only leverage if you use it at scale.

In February 2023, BYD launched the Qin Plus DM-i "Champion Edition" at RMB 99,800 (~$15K USD).

The first plug-in hybrid ever priced under the RMB 100,000 floor. It took 32,000 orders in its 1st week and effectively opened China's EV price war.

BYD kept cutting.

Over 100 model trims repriced over the next 12 months.

Cuts of up to 34% across 22 models by 2025.

BYD knew they were the only ones who could survive, since their cost base starts where everyone else's margin ends.

Ok but seriously - thoughts on the graphics? Would you share this? Why or why not - reply to this email and let us know!

Rabbit Hole

The where: 3x high-signal resources to learn more

Plugged In: BYD's Wang Chuanfu Explains How China's No.1 EV Maker Caught Up With Tesla ↗

[6 minute read]

A rare thing, an actual interview with Wang Chuanfu himself, not a profile written about him. He's notoriously press-shy, so direct quotes from him are genuinely scarce.

Probably the closest you'll ever get to hearing the strategy in his own words.

Why Are Chinese EVs So Cheap? ↗

[15-20 minute read]

The analyst deep-dive that actually quantifies the vertical integration moat instead of just asserting it.

This ones the finance/trading nerds who self-selected earlier.

Blade Battery Nail-Penetration Test ↗

[65 second watch]

BYD's own live demo, and one of the most-shared pieces of BYD content that exists. A nail through a charged cell, no smoke, no fire.

An extension of drinking battery fluid, and an iconic example of Wang’s "safety theatre" marketing playbook.

And it’s only 65 seconds.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

My 2009 ‘Music’ folder was 1.2GB of naming-conventionless MP3s, 13 songs that were secretly other songs, and 1 virus.

I'd queue a LimeWire download before bed and find out in the morning whether "W0nderwall.mp3" was Wonderwall.

Last week we broke down how iTunes got the world to pay 99 cents a song.

Apple's offer was basically: give us a dollar and you can skip the parts of the album you didn’t want anyway, along with the viruses + guesswork + ‘am I gonna get busted’ factor that came with torrenting.

Plenty of us still torrented away. Most of us ended up on Spotify.

Today's piece is on how Spotify ended up beating iTunes and piracy at the same time.

Enjoy.

— Tom

How healthy is your email setup, really?

Omnisend now works with ChatGPT and Claude. Ask it to score your email setup based on campaign cadence, automation coverage, and catalog completeness, then get your top three improvements in seconds.

Not on Omnisend yet? Brands average $79 back for every $1 spent - migration is free in under 5 days, and pricing runs up to 35% lower than Klaviyo's.

Be the future of AI

Hunting for AI roles takes hours you don't have. Careers pages, group chats, internal referrals, a dozen job boards, all updated at random. The Athyna AI Job Board does the watching for you.

• Track openings at frontier AI companies like Anthropic, OpenAI, Mistral, and Perplexity

• Match them to your profile

• Ping you when a role hits 75% matching index

Set up a profile once. The matches come to you.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

Spotify's obsession with speed

Chess Move

The what: A TLDR explanation of the strategy

Blink once.

Knock a domino over.

Say the word “Spotify”.

Each of those things take more time than it takes for Spotify to start playback of a song.

In 2009, a Wired writer described the Spotify beta as "a magical version of iTunes in which you've already bought every song in the world."

Not bought. Not downloaded. Just there, the instant you hit play.

→ That feeling was the product.

Spotify launched in 2008 into a market that had already lost. The fastest way to get a song was to steal it, and no legal company could beat the convenience of a file downloaded straight to your hard drive, for $0.

Spotify founder Daniel Ek stopped competing with iTunes and started competing with The Pirate Bay, on the one important axis that piracy hadn’t quite nailed: speed.

The obsession was utterly total.

They got a song to start in a median of 265 milliseconds, faster than most brains register a wait.

Behind-the-scenes, they assembled an engineering masterpiece, with only 8.8% of the audio coming from Spotify's own servers.

The rest came from your computer’s cache and, for years, from strangers' laptops.

You read that right. Let’s learn how.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Spotify vs iTunes vs piracy

Every music startup in 2008 was trying to out-iTunes iTunes. That would mean a bigger catalogue, a slicker store, better price.

Spotify ignored all of it.

They noticed that the real winner in the digital music battleground was trending towards another player entirely: piracy.

And while piracy had a limitless catalogue and unbeatable price, users did have to spend a lot of time hunting for safe seeders, then wait for files to download and import into their media player of choice.

Copping an ungodly amount of traffic from my IP in 2013

So Spotify built against that. A streaming engine that felt like the file was already yours.

Spotify's own product and engineering lead put it bluntly:

Napster and LimeWire had already trained a generation to expect all music available at their fingertips. Spotify couldn’t argue with that expectation.

So it met it → full catalogue + artist pages + playlists on top → a near-perfect playback experience.

Piracy had already written the spec for user behaviour. Spotify just had to legally build and ship it.

My GOAT for Deep Work

2. The µTorrent hire that built the engine

You can't obsess over milliseconds without someone who can shave them. Spotify's answer was to buy the best low-latency engineer in the world.

In 2006, they acquired µTorrent, the wildly popular BitTorrent client.

Shoutout to all the seeders out there holding it down.

They ended up selling it only months later but that actually didn’t matter.

The code was never the point.

Strigeus, µTorrent's creator, went on to build Spotify's original streaming engine, the peer-to-peer core that made instant playbacks possible, using the same family of technology that powered piracy itself.

So what exactly did he build?

3. How to fake instant at scale

Instant playback is a trick of attention.

(Much like Instagram’s ‘instant upload’ UX magic we wrote about here)

Latency is dominated by the first few 100 milliseconds, so Spotify made the opening of a song as fast as possible, and let the rest arrive however was cheapest.

So what happens when you hit play? 4 things:

→ Cache first. 55.4% of the audio you heard came straight from your device's local cache. Zero network, totally instant.
That's music you already played, still sitting on your hard drive. Spotify never deleted it after playback, so hitting a song twice isn't streaming at all, it's just opening a file.

→ A server head start. If the track wasn't cached, Spotify streamed roughly the first 15 seconds from its own servers over an already-open connection.
The client holds a live connection to Spotify the entire time you're logged in, so there's no dial-up ‘time to get connected’ moment when you hit play. The request goes down a pipe that's already active, and it only asks for the opening seconds, not the whole file.

→ Peers for the bulk. While that intro played, your client pulled the rest of the song from other users who already had it, 35.8% of all data (those strangers' laptops).
You only need the first 15 seconds urgently. The remaining 3 minutes have minutes to show up, so Spotify handed that job to strangers to save themselves the bandwidth bill.

→ Pre-fetch the next track. Before the current one ends, Spotify loads the next song, so "hit next" is gapless too.
In the closing seconds of the current song, Spotify quietly pulls the opening of whatever's up next into your cache, so by the time you press skip it's already ready to play.

Throw these together and you have a ~265ms median start, with <1% of plays stuttering, and only 8.8% of bytes served by Spotify itself.

And when 265 milliseconds still wasn't instant enough (the average human brain can detect down to ~200ms), Ek admitted the last fix was psychological:

Rabbit Hole

The where: 3x high-signal resources to learn more

Spotify: Large Scale, Low Latency, P2P Music-on-Demand Streaming ↗

[30 minute read]

One of the denser reads, but well worth it.

2 Spotify engineers show their working: how half your music came from other people's laptops and you never noticed.

Shows exactly where the 265ms median actually comes from, how the cache → server → peer split measured, and why only 8.8% of bytes ever touched Spotify's servers

Daniel Ek on Masters of Scale ↗

[35 minute listen]

Reid Hoffman gets Ek to confess the 200ms miss, and how a moving loading bar covered for it.

→ Why they obsessed over latency when no one asked

→ The throbber trick, in Ek's own words

→ Building trust when the product has to feel instant

35 minutes on why perception is a product surface.

Acquired: Spotify CEO Daniel Ek, The Complete History and Strategy ↗

[1hr 38min listen]

Recorded at Spotify HQ, Ek walks the whole arc from racing pirates to running the audio internet. The definitive origin deep-dive, the business behind the milliseconds.

3 hours wasn’t enough.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

Today’s breakdown might (rightfully) ruffle some feathers.

iTunes was the catalyst for the music industry moving from albums → individual tracks, and the gateway drug for the streaming + playlists paradigm we live in today.

Majority of consumers prefer it this way.

But as a tragically obsessed music fan (fun fact: I did a music history major, wrote essays on J Dilla + The Doors + Talking Heads for 3 years, good times), I don’t think I’m alone in the “I feel like we lost something important by doing that” camp.

That said, we can’t bury our heads in the sand and pretend the shift didn’t happen. And there’s many lessons we can learn from how and why it happened.

That’s the purpose of today’s breakdown. Enjoy.

— Tom

Voice agents that actually sound human

Most AI voice tools sound flat, and flat can make frustrated customers angrier. With ExpressiveMode, ElevenAgents adjusts tone mid-call: de-escalating heated callers, reassuring customers mid-claim, guiding conversations to resolution.

No awkward pauses, no robot voice. Employees at 75% of Fortune 500 companies use ElevenLabs.

What would your email setup score out of 100?

Want to know how your email marketing setup is really performing? Omnisend now works with Claude and ChatGPT, making it possible with a single prompt.

Ask Omnisend AI to score your account out of 100 based on campaign cadence, automation coverage, and catalog completeness, then list the top three areas to improve. In seconds, you'll get a clear picture of your account's overall health, along with the first three things worth improving.

Not using Omnisend yet? Join more than 150,000 brands generating an average of $79 for every $1 spent. Migration is free and takes up to five days, while pricing is up to 35% lower than Klaviyo's

_{Thank you for supporting our sponsors, who keep this newsletter free.}

The most valuable thing Apple built in the 2000s wasn't the iPod

Chess Move

The what: A TLDR explanation of the strategy

On 28 April 2003, Apple launched the iTunes Music Store with ~200,000 songs available for purchase at 99¢ each.

They forecasted 1 million songs sold in the first 6 months.

They hit it in 6 days.

→ In 12 months: 70 million songs.

→ in 36 months: 70%+ of every legal digitally downloaded songs ran through iTunes

Low-key fire selection in this one Steve.

But just saying "Apple figured out selling songs online" doesn't capture what actually happened.

Apple managed to stack multiple strategic levers on top of each other to enable one of the most significant music industry shifts of all time.

From how they packaged the product (songs > albums)

To how they got the 5 major labels to cooperate

And how Apple collected more credit cards on-file (400 million) than any other company in history (Amazon had 152 million, PayPal had 110 million)

The effects of Apple’s iTunes strategy still echo in 2026.

Let’s get into exactly how it worked.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Change the unit of value, change the industry

Before iTunes, the album was the atomic unit of music.

A CD retailed for $14-$18. Contained 2-3 tracks anyone actually wanted, and 10+ that were alright.

Customers paid for those 10+ tracks whether they wanted to or not.

I lived for those leaderboards on the right.

That bundle of tracks wasn't just the package the industry shipped in.

It was the margin. Labels priced, produced, marketed and signed artists around the entire album, because that's where the economics worked.

iTunes broke it.

→ 99¢ per song

→ Any song

→ Albums available at $9.99, but optional

The unbundling shook up the industry.

Within a few years, CDs went from the dominant unit to a rounding error of total sales.

Apple understood that every business has an explicit product and an implicit product.

Labels thought they sold music. They actually sold bundles.

iTunes knew people wanted songs.

It’s not “iPod. A 5GB portable MP3 player with mechanical scroll-wheel.”

2. Look small enough to get the deal. Then become the gatekeeper.

By 2002, music labels were in free fall.

Napster had been sued out of existence the year before, but Kazaa, LimeWire and Morpheus were bleeding the industry from every angle.

Genuine nostalgia finding this screenshot

The labels had already tried to fix it themselves by controlling their own distribution:

→ PressPlay (Sony + Universal) → launched 2001 → flopped

→ MusicNet (Warner + EMI + BMG) → launched 2001 → flopped

Trust between the 5 majors was at a generational low.

Jobs saw an opening and spent ~18 months personally lobbying each label CEO to get their libraries on iTunes.

Doug Morris at Universal and Roger Ames at Warner were the first dominoes. Morris later said:

Once Universal and Warner were in, EMI, Sony and BMG had to follow or be locked out.

The thing that made it work was the opposite of what most founders do.

Jobs didn't try to look bigger than Apple actually was. He deliberately looked smaller.

→ Mac was ~3% of the PC market at the time

→ He sold iTunes Store as a Mac-only experiment, not an industry play

→ Contracts were 1-year exclusives and could be renewed flexibly

→ Labels could walk if it went badly

Then there's the buried detail almost nobody mentions.

The 99¢ price point wasn't Jobs' idea → Most labels actually wanted $3.49 a track.

Warner executives proposed 99¢ in the negotiation. According to Paul Vidich (then at Warner):

As digital music sales exploded, those 1 year term contracts renewed, and renewed, and renewed, and ever since the big 5 labels have been long-standing tenants on Apple's platforms.

Within 3 years, iTunes accounted for 70%+ of all legal digital music sales.

The App Store gets all the love, but this is a top-5 underrated moment in software history

3. Exchanging margin for (Credit Card) data.

So you buy a song for $0.99

After ~65¢ goes to the the label and ~25¢ in card processing, Apple was running roughly break-even per song.

That was the point.

Because The iPod was the actual product.

iPod gross margins were ~25-30%. A $399 iPod cleared ~$100+ in profit. One iPod sold was worth more to Apple than 1,000+ song downloads.

The numbers followed:

→ 939K iPod units sold in 2003

→ 4.4M in 2004

→ 22.5M in 2005 (that’s a LOT of new-category hardware sales)

→ US market share: 31% → 74% in 18 months

iTunes was the marketing expense that made the iPod indispensable.

But the iPod wasn't even the final answer.

Timeless.

Underneath the iPod was something more valuable: a credit card on file in tens of millions of homes, with users trained to one-tap buy.

Every 99¢ purchase did two things at once:

→ Sold a song of course

→ Trained customers to "give Apple your credit card" for very small amounts of money.

Music was the perfect habit-builder for repeat spending with card-on-file. Low cost. High frequency. Immediate reward.

By the late 2000s iTunes had the largest active credit-card-on-file databases in consumer tech.

7 years later when the App Store launched, every iPhone user already had a payment method on file from iTunes.

One tap = purchase. Zero signup friction.

Every previous mobile app marketplace (Verizon's Get It Now, Nokia, Palm, BREW) had died, primarily because of payment friction.

Apple had quietly solved that problem years earlier, under the cover of music.

Be honest - when a site/app has an option to pay with Apple Pay, using saved card details, without extra fees, you take it, right?

The same payment rails then unlocked:

→ In-app purchases (2009)

→ iBooks (2010)

→ iCloud (2011)

→ Apple Pay (2014)

→ Apple Music (2015)

→ Apple TV+ (2019)

→ Apple Arcade (2019)

→ The entire modern Apple Services line, now a $100B+ business and Apple's highest-margin segment

Worth more than Coke and Netflix combined.

Rabbit Hole

The where: 3x high-signal resources to learn more

Steve Jobs introduces the iTunes Music Store ↗

[65 minute watch]

The primary source. Jobs in 2003 walking through the launch on stage at Apple Special Event.

Notice how he frames the labels (hint: they’re not enemies)

→ The famous "we think this will go down as a turning point for the music industry" line

→ Tracks at 99¢, no subscription, no DRM hassle (his framing, not mine)

Hearing Jobs sell the move is its own masterclass.

Steve Jobs on the iTunes Music Store: the unpublished interview ↗

[20 minute read]

Harry McCracken's previously-unpublished Jobs interview, dug out of his archives. Jobs himself walking through how the deal with the labels actually came together.

An inside baseball look at how Jobs pulled together the big 5 labels onto iTunes.

A timeless lesson in the school of negotiation from Steve Jobs.

The deal that forever changed music and technology ↗

[10 minute read]

The label-side retelling, sourced from the executives who were actually in the room.

Includes the buried detail that Warner suggested the 99¢ price, not Jobs.

If you want the story from the labels' point of view, this is the one.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 150,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

If you ran Strategy Breakdowns for 24h, whats the #1 thing you’d add/remove/alter/improve?

Reply to this email and let me know!

— Tom

p.s. Olá from Douro Valley, Portugal

Most voice agents break on real customers. Yours won't

Anyone can spin up a voice agent. The hard part is one that survives real customers. ElevenAgents maps every decision point and handoff in a visual editor, and lets you test it against your hardest conversations in 70+ languages before a single customer hears it.

Same platform behind agents at Cisco, Klarna and Deutsche Telekom and trusted by employees at over 75% of Fortune 500 companies.

What would your email setup score out of 100?

Omnisend now works with ChatGPT and Claude. Ask it to score your email setup based on campaign cadence, automation coverage, and catalog completeness, then get your top three improvements in seconds.

New to Omnisend? Join 150,000+ brands.

Free migration in up to 5 days and pricing up to 35% lower than Klaviyo's.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

DJI used GoPro to kill GoPro

Chess Move

The what: A TLDR explanation of the strategy

In 2013, a little company called DJI shipped its first consumer drone with a GoPro mount glued to the bottom.

DJI Phantom 1 with GoPro Mount

This small piece of plastic went on to be the most important decision DJI ever made.

Remember this era? GoPro was worth billions, owned the action camera category, and was in every surf shop, ski resort, and Red Bull stunt video on the planet (take me back).

Consumer drones were barely a category.

So when the mount for DJI's drone came along, GoPro owners didn't need much selling. They already owned the camera. They just needed something to fly it with.

In 2014 DJI’s Frank Wang met with GoPro’s Nick Woodman to discuss a new co-branded drone/camera product. GoPro’s proposition:

• DJI would build the drone

• GoPro supplies the cameras and brand

• GoPro takes two-thirds of the profit.

Wang refused. He argued that the flight controller and autonomous stabilisation system is the genuinely hard part, the camera is … just a camera.

3 months later, DJI launched a standalone drone: Phantom 2 Vision+ with its own integrated camera. No mount and no GoPro.

By the time GoPro's own drone, the Karma, shipped (30 months after that meeting), DJI had already shipped 4 more generations.

GoPro’s Karma drone was recalled 16 days after launch (battery decoupling issues mid-flight = 2 pound drones falling from the sky)

By January 2018, GoPro had discontinued it entirely, and the rest is history.

💀

The drone war was over before most people realised it had started.

Here’s exactly how it went down.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  The mount is the wedge

DJI’s original Phantom wasn't trying to compete with GoPro. It honestly completed it.

By shipping with a GoPro mount, DJI borrowed a $3b+ brand to validate "the flying camera" with zero camera R&D and zero category education spend.

GoPro owners already had the camera, so the drone was an easier purchase decision.

DJI got customers, a brand halo, and 18 months of product-market fit data in return.

We’ve seen this pattern before:

→ Microsoft licensed MS-DOS to IBM for the original PC in 1981 and used the deal as a beachhead onto every PC that shipped. <10 years later, IBM lost the top position in the PC value chain.

→ Amazon spent years letting third-party sellers list products on its marketplace, quietly tracking which categories sold at what margins. Then they launched Amazon Basics versions of the bestsellers in 2009 and undercut those same sellers.

→ Apple shipped Intel chip Macs for 15 years while quietly building its own silicon design team then in 2020 ejected Intel from the Mac entirely with Apple Silicon, costing Intel roughly 10% of its volume overnight.

Compatibility with the incumbent is a cheap and fast way to validate new categories because you skip the hardest part: convincing people your brand is worth buying.

2. Bail from the bad partnership

In early 2014, Frank Wang met Nick Woodman to formalise a partnership.

Wang refused.

3 months later, the Phantom 2 Vision+ launched with DJI's own integrated camera.

Hmmm looks a bit different than the last model…

GoPro's offer showed Wang how GoPro saw the relationship: a camera company graciously letting a drone supplier into its orbit, at a 2:1 split.

But the initial mount showed DJI that the drone category was hot.

And while Wang knew GoPro's brand recognition mattered, he believed their flight technology was a wider moat to replicate.

Consumers wanted drones, and DJI was the only company that could make drones that fly well.

Much easier for DJI to ship a small camera (commoditised component) than for GoPro to ship a drone (niche, patented, intertwined hardware + software components).

GoPro needed DJI to compete in the drone market.

DJI didn’t need GoPro.

By leaving the GoPro partnership, they locked in their market status as the only drone worth buying, then bet aggressively on that thesis over the next 18 months to compound their advantage.

3. Out-ship as a strategy

After the failed meeting in early 2014, GoPro began work on their own integrated Drone/Camera product.

Meanwhile, DJI kept its head down and shipped 4 new generations:

→ April 2014: Phantom 2 Vision+ (integrated camera)

→ April 2015: Phantom 3 series (4K camera)

→ March 2016: Phantom 4 (automatic obstacle avoidance)

→ September 2016: Mavic Pro (foldable)

DJI even sneaked the launch of their Mavic Pro just 8 days before GoPro’s big product launch of the Karma.

The spec sheets told the whole story

It was a superior product in every way, and outsold the Karma by a landslide.

That is, it would have outsold the Karma if the Karma could survive on the market for more than a month.

The Karma was recalled 16 days after launch. Mechanical issues and lawsuit risks led GoPro to discontinue it in January 2018. DJI’s “they can’t build good drones without us” thesis was right.

It’s a perfect case study of why iteration speed wins.

A competitor's launch window is their most exposed moment.

• Hardware pecs are locked in 12-18 months prior

• Marketing and positioning are fully committed

• The audience is paying attention

• They cannot pivot

DJI used those 30 months to ship 4 products, each one cheaper to manufacture and harder to compete against. By the time GoPro finally arrived at the starting line, DJI was literally years ahead.

The saga may have ended, and GoPro is now a shell of its former self, but that hasn’t stopped DJI from innovating.

→ Their Osmo cameras have become the de facto camera for vloggers.

→ Their DJI Mics have become the de facto microphone for vloggers.

And DJI has poetically become the category-owning 'GoPro of drones' in 2026.

^{The Strategy Breakdowns behind-the-scenes tapes was filmed entirely on DJI’s legendary 2-product combo. Click here to be the first to watch when it drops.}

Rabbit Hole

The where: 3x high-signal resources to learn more

The GoPro Karma launch announcement ↗

[3 minute watch]

GoPro’s launch video outlining the features of the Karma. Immediately ran to the comments - top one?

"It's a complete system" \ Camera not included*

A monument to the moment the drone war was lost.

Funny to watch knowing the product gets recalled just 16 days after launch.

Drone Overlord Frank Wang On DJI's Milestones, Miscarried GoPro Partnership & Corporate Espionage ↗

[12 minute read]

Frank Wang's one and only deep interview (2015). The reclusive DJI founder broke a decade of silence twice. This is the first time.

The GoPro saga is in here in Wang's own words, but the wilder thread is the corporate espionage subplot.

If you want the primary source on the deal that changed drone history, this is it.

The story of drone pioneer DJI ↗

[13 minute watch]

The rise of Frank Wang, told in full. From building flight controllers in his HKUST dorm room to running a $15B company from a Shenzhen tower nobody is allowed to photograph.

How dorm-room prototypes that became the Phantom and why Wang refused outside investors for years (and what that did to DJI's culture).

If the Karma launch video shows you the loser's stage, this shows the winner's workshop.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 125,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

I’ve always adored this little workhorse.

The horizontal scroll wheel. The thumb action buttons. The way it hugs your hand.

When I made the switch from spending my free time farming gold / XP / loot to farming followers / subscribers / $USD, and wanted a mouse that matched the identity shift, every single YouTuber, Redditor, and 1m traffic/mo SEO affiliate blog recommended the same bit of gear:

The Logitch MX Master series.

Go to any co-working space and you’ll see everyone rocking the same piece.

You might even be scrolling this article using the very same (reply to this email and send in a pic if you are!)

Today we’re figuring out how Logitech became the undisputed winner of the non-gaming mouse category.

— Tom

P.S. In Porto for the week - any reccs?

Deploy a voice agent in minutes, not months

Voice AI used to be a months long engineering project. ElevenAgents by ElevenLabs cuts it to an afternoon: upload your knowledge base, set your tone, and go live across voice and chat. No code at any step.

Your agent speaks 70+ languages and plugs into Salesforce, HubSpot, Zendesk and Zapier out of the box. 4M+ agents already deployed and used by employees at 75% of the Fortune 500.

Turn back-to-back meetings into real follow-through.

Leave a call and forget half of what you agreed to? That's not a memory problem - it's a meetings problem.

Granola is the AI notepad built for it: jot quick notes while it transcribes in the background, then it turns your raw notes into clear summaries with real next steps. After the call, chat with your notes to pull exactly what to do next - no re-reading.

(Strategy Breakdowns readers get 1 month free - use code BREAKDOWNS)

_{Thank you for supporting our sponsors, who keep this newsletter free.}

How Logitech turned a $20 commodity into a $100 status symbol

Chess Move

The what: A TLDR explanation of the strategy

It's literally just a mouse. A small plastic dome you push around a desk to move a tiny arrow on a screen.

Logitech calls it the MX Master, charges $119.99 for it, and has developers, designers, editors, and knowledge-workers all calling it the de facto standard.

The mouse of the decade.

It’s the best-selling productivity mouse (i.e. non-gaming mouse) of all time.

And for a category that revolves around the $25 mark, the product had no business existing at this price point.

Mice were textbook commodities. They were the freebie thrown in with the laptop or the thing your employer reissues for free.

So how did Logitech get customers to pay 5x the going rate?

In 2013, Logitech made 3 pivots:

1. Pulled R&D dollars out of the commodity mouse business and reinvested them into design for premium product ranges.

2. Hired ex-Nokia design chief Alastair Curtis as their first Chief Design Officer to launch The MX Master in March 2015 at $99.99

3. They over-engineered it. Patents, hand-sculpted prototypes, electromagnetic scroll wheels, ergonomic research.

Then COVID hit. All of a sudden - 1.5 billion WFHers needed a serious home setup, all at once, and Logitech's revenue went from $2.98b → $5.25b (+ 76%) in just 12 months.

One of the most underrated winners from the remote work boom. Worth studying how they created the right conditions for establishing and winning a new consumer market, just in time to ride the WFH wave.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Rebuild the company before you rebuild the product

Logitech had shipped 100+ mice before the first MX Master.

What changed?

When Bracken Darrell joined Logitech in January 2013 from Whirlpool, P&G, and GE, he described the company as:

“One of the quintessential examples of a PC-related company, and our profits were dropping."

So when he diverted R&D dollars from the “commodity” mouse business into design and growth categories, then hired Alastair Curtis, the designer behind the Nokia 3210 (150m+ units shipped) as Logitech's first Chief Design Officer, whilst profits were falling, analysts thought the company was doomed.

Moving away from short-term revenue into riskier, longer-term, design-first bets created a dramatic culture shift. Darrell famously kept Dieter Rams (legendary industrial designed of Braun products) 10 principles of good design on his office wall.

The MX Master launched March 2015 at $99.99 and was the first flagship out of this new system.

The product aligned with the company’s philosophical (and literal) restructure, but it required unlearning and rethinking in order to market successfully against the commoditised market…

2. A thousand lines of scroll per second

Once you've decided to charge $100 for a $20 product, you have to make the premium obvious.

Here’s how Logitech did it:

→ Process theatre. 

Every MX Master generation is hand-sculpted in clay before CAD. Real engineers with real hands, testing each each prototype for craftsmanship over economic efficiency.

This is how Herman Miller chairs and Ferrari’s gets made. It is also, now, how a mouse gets made.

→ Patents. 

The ‘MagSpeed electromagnetic scroll wheel’ replaced traditional gear ratchets with electro-permanent magnets.

Logitech holds the patents, and it scrolls 1,000 lines per second.

Even if most users aren’t scrolling 1,000 lines per second, the marketing invokes a feeling of immense productivity - gliding your way through 1,000 lines of code, 1,000 rows of Google Sheets, 1,000 frames of animation etc.

→ Productivity stats with implausibly tight numbers. 

87% of MX Master users reported a productivity uplift in Logitech's own survey.

The MX Master 4 marketing claims 33% less work time.

How they calculated that? Who knows. But it’s a hell of a stat.

The buyer doesn't have to obsess over the design details or buy into the numbers word-for-word.

The buyer just has to differentiate.

Logitch did everything they could to associate the MX Master line with productivity, ergonomics, and quality.

3 things no other mouse at the time took seriously.

3. The WFH wave

The MX Master had been on the market for 5 years and was already the default "serious productivity mouse" on every dev-Twitter recommendation list.

Then COVID happened, and the entire knowledge-worker population was forced into a home-office buying decision at the same time.

Many employers offered Work From Home budgets for employees to beef up their setup.

A product that had been a cult dev tool for 5 years became the go-to upgrade for anyone setting up a real workstation in a spare bedroom.

Naturally the Creativity & Productivity category (think keyboards, webcams, and of course the MX Master) grew double-digits in every quarter of the pandemic.

An interesting leading indicator for “areour products absolutely flying off the shelves right now?” - advertising spend. Can you spot when COVID happened? 🤓

The renewal cadence is the rest of the moat. Every year or 2 they ship a new version with a few extra gizmos.

MX Master 4 launched October 2025 at $119.99 with the same shape, the same buyer, and yet another "record-breaking" launch.

Plenty of MX users now own 2 or 3 generations (including yours truly).

Rabbit Hole

The where: 3x high-signal resources to learn more

 Logitech was dying a slow death. A design-focused CEO helped it thrive ↗

[10 minute read]

"Every category can be reinvented using design and design thinking." 

Bracken Darrell's full thesis on how he turned a commodity peripherals company into a margin business.

Logitech MX Master 3 vs 2S Teardown ↗

[12 minute read]

For my hardware nerds out there. The MagSpeed electromagnetic scroll wheel, photographed and explained.

→ Where the money actually goes inside the housing

→ How magnets replace traditional gear ratchets

→ Why MX Master 3 was a step-change over the 2S

If you ever wondered how a mouse can credibly cost that much, this is how.

The Razer DeathAdder makes history: 10 million units sold ↗

[3 minute read]

An explainer on the gaming mouse that made today’s $100+ mice possible.

3-5x market prices.

10 million units sold.

Arguably the heavyweight champ of the premium-peripheral market.

I played so much Dota 2 on one of these guys

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 125,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

In 1979, Sony shipped a cassette player with no record button.

Every piece of market research said it wouldn't sell.

Every retailer agreed.

Sony’s own engineers begged their CEO to keep the button, because it cost almost nothing to leave in.

He removed it anyway.

That product went on to sell 400 million units.

“Technostalgia”: Fond reminiscence of, or longing for, outdated technology.

I can’t stop thinking about this story.

Trusting the data vs. gut feel.

Sometimes less is more.

Let's get into it.

— Tom

AI agents in Jira cut KTLO hours by 80% inside Atlassian's Eng org - all under Engineer control

The Jira team got tired of the same drill. A test fails for the third time this week. Someone digs in, reproduces the issue, determines the problem, and patches it. Two hours gone.

So they tried something Jira now supports natively. They codified the flaky-test playbook as an agent skill, assigned the tickets to a Jira agent, and let it handle the repetitive tasks like investigation, diagnosis, and open PR. The Engineer signs off.

Result: What took hours now takes minutes. In other words, we reclaimed enough time savings to equal a full engineering week back every month.

This is Jira.

The work your team keeps pushing to next week.

Every team has it. The report nobody wrote. The vendor email stuck in drafts. The code review that waited a week. You can't hire fast enough to clear it, so the pile just grows.

There's a new way to handle it: an AI employee that works inside Slack and Teams and picks up the work no one got to. Meet Viktor.

How Sony sold 400 million Walkmans

Chess Move

The what: A TLDR explanation of the strategy

The first working prototype for the Sony Walkman was a portable cassette player, built on the chassis of their existing Pressman dictation recorder.

The Walkman Prototype could play tapes, but it could also record them, because the recording component was already in the Pressman.

Ripping it out would actually have been more work than leaving it in…

But Sony founder Akio Morita told the engineers to rip it out anyway.

They pushed back:

Recording was the feature customers said they wanted. Sony had run the research.

Retailers claimed nobody would buy a tape device that didn't record. The engineering team agreed. Morita couldn’t care less.

He launched the Walkman at ¥33,000 (about $150 USD) on 1 July 1979, with a 30,000-unit run and an ice-cold promise to the board:

"If we haven't sold 100,000 Walkmans from now to the end of the year, I will resign from my position as President of the company."

The first month was rough. Sony moved only 3,000 units. Tracking towards 36,000 for the year.

Then word of mouth hit.

→ The next two months sold the entire 30,000-unit run.

→ By Christmas, Japan alone had bought 50,000.

→ US sales grew 41.3% in the first twelve months after the American launch.

→ The cassette Walkman sold 200 million units before Sony retired it in 2010.

→ Across all formats - cassette, CD, MiniDisc, MP3 - cumulative Walkman sales hit roughly 400 million.

The feature everyone asked for was the feature that would have killed it. Let’s understand why.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  The deletion was the product

We all know the expression “less is more”, but it takes real courage to spend time and money on removing features.

The Pressman chassis already recorded. Leaving the record button in would have cost Sony nothing in engineering hours and saved them a redesign.

The "lazy" (even “rational”) path was to ship the prototype as-is and let the market decide.

Morita chose the hard path: rip out a working feature, redesign the case around its absence, and ship something narrower.

The deletion did four things at once:

1. Halved the bill of materials (no record head, no record electronics)

2. Shrunk the form factor small enough to fit in a coat pocket (iPod inspiration anybody?)

3. Hit the $150 price point (vs the $200+ price point of other cassette players)

4. most importantly… Killed user confusion about what the device was for.

More of an Apple-style lineup - different devices for different jobs

Steve Jobs ran the same play in 2007.

The iPhone shipped without a physical keyboard, against every piece of evidence that BlackBerry users would never switch. Deletion was the product. A keyboard would have made the iPhone a worse BlackBerry.

Every builder knows that addition is intuitive. Every PM, every roadmap, every "what's next" conversation pulls toward more … but subtraction?

That takes conviction, because the case for keeping the feature is always concrete and the case for deleting it is always speculative.

You can always find the customer who asked for record, then bought.

But you can't find the customer who wouldn't have bought a heavier, more expensive, more confusing Walkman, but would buy the alternative, because that customer doesn’t exist yet.

2. Your most current data is for an outdated category

Sony's market research wasn't wrong, but it was answering a different question than Morita was solving for.

When you ask a 1979 consumer about a portable cassette player, they answer using the only reference point they have: existing cassette players, which all recorded.

So their answer is "yes but make sure it records". They're not lying. They're describing the only product they can imagine, which is a variation of the product that already exists.

“Remove the big red button” ~ Akio Morita (probably)

Said another way.

Said another way

This pattern appears everywhere.

Loom, AirBnb, and Spotify would never survive a 2001 focus group.

→"Do you want to record a 90-second video instead of writing a few sentences?"

→"Do you want to stay on a stranger’s sofa bed instead of a hotel?”

→ "Do you want to pay a monthly fee to rent, not own, digital music files, with no physical CD/record to add to your shelf”

The data would have said ‘no’ because the category didn't exist for the data to describe.

The rule isn't "ignore data". The rule is "is this data relevant to the new category I’m trying to build?".

3. Design for the behaviour, not the feature

The TPS-L2 had 2 headphone jacks so 2 people could listen together.

Makes sense.

It also had a button labelled "Hotline" that activated a built-in microphone.

Press Hotline, the music paused, and you could talk to your listening partner without taking your headphones off.

Innovative right?

Nope. It was pointless, since they could just take off their headphones, so Sony quietly killed them in the next model.

They also learned from the first batch that people wanted to listen alone, not together. The behaviour the Walkman actually created was solitary - the personal soundtrack, the bubble, the world tuned out. So they removed the 2nd headphone jack too.

Morita's team was last focussed on what people would do with the product, not what features the product should have.

Their guess about the behaviour was wrong. The discipline of guessing about the behaviour, then adapting based on reality, was right.

Most product teams reason at the feature level by writing spec docs, benchmark competitors, list capabilities.

But when you're inventing a category, you're not really shipping features. You're actually shipping a verb.

• We google things (typing into a search box)

• We uber somewhere (having strangers drive you around)

• We snap friends (send disappearing pictures to friends)

The Walkman shipped walking around with music.

If you're staring at a roadmap right now → ask yourself: what verb am I shipping?

Rabbit Hole

The where: 3x high-signal resources to learn more

Made in Japan ↗

[10 hour read]

Morita's own account of the Walkman decision, the founding of Sony, and his philosophy on why market research kills new categories.

The chapter on ‘inventing demand’ should be required reading for every founder.

The Story Behind Sony's Iconic Meditating Monkey Walkman Commercial ↗

[8 minute read]

Yup. That monkey meme was a 1988 Walkman ad.

A 10-year-old performing macaque refused to cooperate for two days. On day 3, during an unscripted break, he sat down, closed his eyes, and the camera was already running.

That accidental shot became the ad.

Iconic.

SONY on Acquired ↗

[4hr 22min listen]

Ben and David spend 4 hours on the company that invented portable music, the CD, the PlayStation, and the format wars.

• The Morita/Ibuka founding dynamic

• Why Sony almost didn't ship the Walkman

• How the company lost the iPod war it should have won

💡Any ideas for future breakdowns? PLG mechanics, viral loops, GTM tactics. Reply to this email and send ‘em in!

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 125,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

Me + friends gatekeeping the aux button-mashing our way through uni.

Back when b2b meant Back-to-Back, not Business-to-Business 😮‍💨

I've been chronically obsessed with electronic music since I found this deadmau5 remix in high school. (Still certified sauce in 2026.)

Somewhere along the way, noticed every single DJ set on YouTube has the same branded decks.

21-year-old lifemaxxing me shrugs, downloads Rekordbox, and starts cooking tonight’s playlist.

30-year-old tech-pilled me wonders if the monopoly happened because a hardware-company built a software-moat, or vice versa.

Here’s the strategy behind the only dog in town when it comes to DJ-tech: Pioneer.

— Tom

P.S. Chuck on Strategy Breakdowns FM: Housey Flow while you read this one. Or Deep Work if you’re procrastinating rn and going straight back into it after reading. Or Techy Breaks if today might be the day you pivot your entire business model. 297 saves across the 3 of them 🎧

Your AI employee lives where your company already talks.

Until this week that meant Slack. Now it means Microsoft Teams too. 

Viktor is an AI employee you message like a colleague: he reviews code, writes the board update, builds and ships a landing page, and pushes back when your plan has a hole.

3,200+ tools connected, SOC 2 certified, and your data never trains models.

Want 125,000+ tech founders and operators to meet your product?

You’re reading this. You know why our audience converts. You’ve seen the world’s fastest-growing startups like Clay, Attio, Granola, Viktor etc run ads with us.

71% of our readers influence budget / procurement decisions.

Grab a sponsorship slot and reach decision-makers where they're paying attention.

Quarterly price increase kicks in on 1st July.

The reason every DJ booth looks the same

Chess Move

The what: A TLDR explanation of the strategy

Walk into any club, there’s a good chance you’ll see 2-4 of these:

Pioneer DJ owns roughly 70% of the global DJ hardware market and close to 100% of professional club booths.

Their software Rekordbox ships free with every piece of hardware - it’s used by DJ’s to mark cue points, set loops, apply effects and mix tracks together. It’s the Microsoft Word for DJ’s.

In their 88 year history (damn), they’ve managed to keep on top of every major tech shift impacting the industry.

From Vinyl → CD → USB → SD card → Cloud, Pioneer’s hardware and software has become the de facto industry standard.

Let’s break down why.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Vinyl, CD, USB, SD, Cloud. Win the extinction events

First, what is a CDJ?

Short for Compact Disc Jockey: a CD player engineered for DJing, jog wheel and all, launched by Pioneer in 1994.

In 2026, they no longer ship with a CD slot, but the name stuck because the form factor did.

It's a music player engineered around the things DJs need to do during a live set.

That matters here because the CDJ is now on its sixth media format in 30 years.

And Pioneer has won every transition.

Vinyl → CD → USB → SD Card → Cloud.

These platform shifts are usually when monopolies die.

→ Kodak couldn't make the jump from film to digital.

→ Blockbuster couldn't make the jump from disc to stream.

→ BlackBerry couldn't make the jump from keyboard to touchscreen.

Format changes are extinction events for incumbents because the new buyers don't carry the old switching costs.

Pioneer's playbook: they always ship the next format inside the current deck.

• CDJ-1000 (2001) read CDs, then the MK2 read early MP3 CD-Rs.

• CDJ-2000 (2009) kept the CD slot but added USB and SD Card. DJs could carry both during the changeover.

• CDJ-3000 (2020) killed the CD slot. By the time it shipped, every working DJ had already migrated to USB / SD Card.

• CDJ-3000X (2025) added CloudDirectPlay. The deck streams your prepared library straight from your Dropbox / Google Drive. USB and SD still work. Cloud is layered on top.

Pioneer controlled the migration path by bridging to the next format with each hardware release.

2. A USB is a contract

At midnight on a Saturday, a $200k DJ walks into a club, plugs in a USB.

The music HAS to play.

Every cue point, every loop, every hot cue intact. The booth manager doesn't get to say "sorry, can you reformat that for our setup."

A DJ's USB isn’t just their playlist - it’s their entire show.

• Loops

• Hot cues

• Beatgrids

• 8hr set notes

It’s all metadata baked into a USB prepped in Rekordbox.

Plug it into a Pioneer CDJ and it plays correctly.

Plug it into anything else and that metadata gets buckled.

Every track a DJ plays has cues, notes, and data in Rekordbox that make it easier to mix with.

The flywheel:

→ Clubs buy CDJs → Reads Rekordbox USBs perfectly → Every DJ preps in Rekordbox → So their USB plays anywhere → Clubs buy CDJs, because every DJ prepped in Rekordbox

Pioneer forced the industry to standardise. Clubs choose Pioneer because it's the only deck every A-list DJ shows up expecting to use.

Some argue that Denon DJ has technically been making better gear for years.

Everyone argues they’ve also been technically losing for years.

The safe choice beats the better choice when incompatibility isn’t an option.

Pioneer even has a deal with San Disk for branded USB sticks. Genius.

3. The DJ Deck is the most valuable billboard in music

Watch any DJ set on YouTube.

The deck sits between the camera and the DJ.

The Pioneer logo is the most visible piece of branding in the entire shot.

^{A personal favorite. Effortless swagger.}

Multiply that by the entirety of recorded DJ culture for the past 15 years and you have what is probably the largest free brand marketing campaign in music.

This is the inverse of how sponsorships usually work. Most brands pay to be in frame for a few seconds. Pioneer is being paid to be in frame for every second.

It also creates a self-fulfilling status moat.

A 19-year-old learning to DJ has only ever seen one deck on YouTube, in every set, in every venue.

→ They buy a Pioneer controller.

→ They prep in Rekordbox.

→ They show up at their first paid gig with a Rekordbox USB.

→ The club has CDJs.

→ Everything works.

They never even consider an alternative.

The product is the billboard and the curriculum. The next generation of DJs are being trained on Pioneer hardware right now, every weekend, on every clip on the internet.

Make your product the camera-facing surface.

P.S. Since you got this far, you've earned some more music → crack open StrategyBreakdowns.com and scroll to the very bottom 🎵

Rabbit Hole

The where: 3x high-signal resources to learn more

30 Years of the CDJ ↗

[25 min watch]

DJ Mag spent 25 minutes inside AlphaTheta (Pioneer’s infamous 2020 rebrand 🤷‍♂️) with the engineers who built the deck that ate the booth.

→ How a Japanese hi-fi company ended up controlling 100% of professional booths

→ Why USB beat the laptop

→ The CEO's plan to "surpass" his own product

If you only watch one thing on Pioneer, watch this.

Why Pioneer DJ is the industry standard (whether you like it or not) ↗

[8 minute read]

Phil Morse has been writing about DJ gear for 20 years. He's tried every alternative to Pioneer. He still ends up back on CDJs.

The clearest case of why "better gear" loses to "the gear in every booth" and why no one is dethroning Pioneer in our lifetime.

 The untold history of the CDJ ↗

[10 min read]

The full timeline of every CDJ from the 1994 CDJ-500 to the CDJ-3000X, with the strategic context for each launch.

• 1994: First pro CD player for DJs

• 2001: CDJ-1000 breaks vinyl's grip

• 2009: USBs replace CDs, Rekordbox lock-in begins

• 2020: CDJ-3000 ships with no CD slot at all

The migration path played out in slow motion.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 125,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

Quick callout to the legends reading this before we dive into today’s piece:

→ We're hiring our first Partnerships Manager for Strategy Breakdowns

Looking for a unicorn who:

→ Can sell AND build
You can close a $100k annual partnership call in the morning and ship an enrichment agent in the afternoon.

→ Is high-agency
20% uncapped commission on new business + renewals, and the freedom to build + run the whole sponsorship engine.

→ Builds trust
Our sponsors ( Atlassian, Attio, Granola 💚 etc) don't see us as a media buy - they see us as a partner. Success in this role is fundamentally about building relationships and helping our sponsors win.

Remote. Full/part-time. Working directly with me. Team of 4 atm.

Not you? Send this to someone who should apply!

— Tom

You wear every hat here.

Which means you already know the business better than any prompt does: the invoices nobody chased, the campaign quietly burning money.

Viktor is the AI employee in your Slack. Tell him what you do in plain words, and he spots the work worth doing and starts on it.

No magic prompt required. Describe your world. He writes the to-do list.

Most mid-market companies have not started with AI.

RSM’s 2025 mid-market research found that the majority of mid-market companies have not adopted AI, and a still-larger majority describe themselves as only “somewhat prepared” to implement it.

BetterBrain runs an AI Readiness Assessment built for that gap. It surfaces where AI is actually already in use across your organisation (almost always more than the leadership team knows), names the highest-leverage starting point, and produces a defensible plan you can take to a board.

You will get a clearer answer in twenty minutes than most companies get in a six-month strategy engagement.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

The AI notetaker winning by not joining your meeting

Chess Move

The what: A TLDR explanation of the strategy

Every day it seems AI notetakers are competing to be more present in your meetings.

Bots in the room, transcript popups, email notifications to everyone before and after the call.

Many such cases.

Granola is winning by doing the opposite. They made the AI invisible. And it’s working. A recent $1.5B valuation with 70%+ weekly retention built mostly on word of mouth.

But that invisibility is just the beginning.

Once the notepad earns trust, the meetings themselves becomes Granola's richest context source - a queryable, action-ready knowledge base you can chat with after the call ends.

And the growth strategy is just as essentialist as the product.

→ No over-engineered marketing funnel

→ No "we do everything" AI promise.

Granola serves one job, for one user, in one moment, exceptionally well.

Today we're breaking down why less is more. How Granola built the fastest-growing AI notetaker by getting out of the way.

Big shout out to the Granola team for making this breakdown possible - plus, a little perk for you if you sign up using the code BREAKDOWNS 👇

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Elegant interface hijacking

Every other AI notetaker sends a bot to join your call.

We’ve all seen this by now.

It sits in the corner, records the room, and kills the vibe instantly.

Yes, it creates a PLG surface, but it’s really there because its how they record the call’s audio.

Granola is filled with thoughtful design details that make it feel seamless, starting with the fact that they refused to add a bot to your calls.

Instead, it transcribes at the macOS layer - grabbing the audio going in and out of your computer directly.

No call participant.

No awkward “hey can you let my notetaker join?”

Nothing to take attention away from what’s being discussed.

The best innovations often seem obvious in hindsight.

It sounds like a minor technical choice. But it’s a deeply embedded characteristic that enables downstream product innovation.

Because Granola listens at the operating system, it doesn't care which platform you're on. Zoom, Meet, Teams, a phone call on speaker, a demo playing in your browser - if the audio reaches your computer, Granola hears it.

One product. Every platform. No integrations to maintain, no APIs to beg permission for. 

Universal by default - which quietly expands the market in a way bot-based rivals can't copy without rebuilding from scratch.

Another simple-but-magical interface hack is the pre-call notification.

Since Granola authenticates with your calendar, it always knows what meetings are coming up.

Three minutes before each meeting, a notification slides onto your screen. One click does 3 things:

• Fires up the call in your default browser

• Opens Granola

• Starts transcribing

The clever part: it isn't a native macOS notification - Granola built their own. It sits above every window, even if your Granola app is closed, and your notifications are hidden.

The result: You never miss a call. And you never forget to ‘Granola’ the call either. Every meeting, the habit gets built for you, because the friction is so minimal.

2. Notes that meet you where you are

How many 4,000 word transcripts do you have sitting in your inbox collecting dust?

AI can transcribe meetings perfectly, but different parts of the conversation matter to different types of people.

Founders Sam and Chris Pedregal started with a different question: what if AI just made your notes better?

Granola's interface is a clean notepad.

You take notes the way you always have, but with less pressure. One word here. Website someone mentioned there. Just a few things that stood out.

Granola transcribes in the background, then uses your notes as the scaffold for the summary - adding more detail to what you jotted down, but filling in the gaps too.

Any line of the summary can be expanded into a deeper recap, including relevant quotes and context

Notes you helped shape are notes you actually trust, revisit, and act on.

Users feel a genuine sense of “*ownership*” and those notes become something you’re more willing to revisit.

It’s the IKEA effect for meeting minutes.

And for the first adopters, that ownership was contagious.

Granola's earliest fans were VCs, YC founders, and operators living in back-to-back meetings - the most platform-agnostic users in tech.

One hour they're in Zoom → the next in Teams → the next in Meet.

A native notetaker built into one of those platforms locked to one stack couldn’t follow them, but since Granola was a local app that transcribed audio at the MacOS layer, it could.

So word of mouth spread, because the product experience felt so seamless and personalised for every user.

If “can everyone see my screen” was the quote of early 2020’s then “mind if I record” is doing the same in the mid 2020’s.

3. Meetings are your richest context source

Once notes are trusted, the meeting becomes the most valuable knowledge base you own.

What users do with that is where things get interesting.

Granola Chat lets you query any past meeting in plain English:

"Based on my convo with Tom, draft a follow-up that covers the action items"

Teams Spaces let you share meeting notes with your team. User interviews, sales calls and team meetings all in 1 place, so everyone stays in the loop.

Recipes are pre-built prompts users save and share (which turns the feature into a viral loop).

E.g. Granola's own marketing team built a “Write Briefs” recipe, which they use when briefing collaborators like Strategy Breakdowns.

Each one is a tiny product a user can create and use, expanding the products reach far beyond “just another AI note taker”, laying the groundwork for their transition to a full-blown AI knowledge base platform.

Matt Mochary’s “Coach Me Matt” recipe went viral on X when every Granola user got a good hard look at themself in the mirror using some infamous, self-reflection coaching techniques

Twitter tweet

Rabbit Hole

The where: 3x high-signal resources to learn more

The 5 Hidden Rules Behind Successful AI Products - Chris Pedregal on Behind the Craft ↗

[45 min listen]

Most AI products win the first week and lose by month two. Chris explains how Granola hits 70%+ weekly retention and remains the AI tool of the day, week, month, *year*.

→ The 5 product principles he won't ship without

→ Why "magic" is the wrong goal

→ How to design AI that users actually keep opening

If you're shipping anything with AI in it, this is the one.

Introducing Briefs, prepare you for your next meeting as you join ↗

[2 minute read]

As you join an external meeting, Granola now helps you prepare for it.

2-3 lines of prep – who they are, what you discussed last time you met them, recent company news, relevant links – gathered for you overnight from your meeting notes, the web and your email.

Granola’s first big feature drop beyond meeting notes. Built for people in back-to-backs. Works auto*magically* - no setup required. Lick the rest of the product - v slick.

Chris Pedregal on the PLG-as-GTM playbook ↗

[45 min listen]

“Our strategy was to mimic Slack or Dropbox — someone inside a company discovers it, falls in love, tells colleagues, grows organically, then the company pays for enterprise.”

It’s one of those ideas that makes sense when you hear it, but in practice it requires a relentless obsession over quality that most founders won’t prioritise.

Short-term pain, while you allocate resources away from marketing and sales and towards product excellence.

→ Pushing the boulder up the hill.

Long-term gain, where growth feels effortless because the product spreads itself.

→ Rolling the boulder down the hill.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 125,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

For the last 18 months, I've been chipping away at a digital products arm of Strategy Breakdowns.

It's called StrategyHub, and it teaches the exact research-and-hypothesis system I built over 6 years of working in strategy - the same tools, tactics, and templates that power this newsletter.

Today’s breakdown isn’t about StrategyHub. But it is about the product that powers it: Kajabi

Kajabi has always been a leader in the creator economy. They found their market wedge as an all-in-one course platform back in 2010, and quickly became the de facto default option for online educators looking to productise and monetise their knowledge.

In 2026, both sides of the creator economy look very different to 2010:

→ Creators rely on AI tools to plan, build, and launch their products

→ Consumers expect AI tools and personalisation to implement what they learn

In the last 2 weeks, Kajabi have launched 5 (!) new products that embody this shift:

• Program personalisation

• Affiliate network

• The Kajabi MCP

• Student-facing agents trained on your content

• AI business partner that builds products

Today’s riff is a collab with the Kajabi team on why the original creator economy premise is dying, what's replacing it, and how Kajabi is quietly rebuilding itself to own the next chapter.

One reason I’m particularly excited about this breakdown: it’s less playbook, more stream-of-consciousness. Something a bit different. The legends over at Kajabi wanted the sauce. Front-line observations on an industry I’m genuinely obsessed with, and living in every day.

So today’s Kajabi breakdown is ultimately a collection of scribbles and thought bubbles from the last few years of building this thing. V fun to pull together.

I hope you enjoy it too.

(Plus, Strategy Breakdowns readers get 50% off all Kajabi 6 month plans, this week only)

Enjoy.

— Tom

Lean teams now operate like ones twice their size.

This week one Slack thread closed a DTC founder's books, audited Klaviyo and rewrote 3 flows, summarised 2 vendor contracts before standup, and opened a pull request engineering had pushed off for a month.

Same employee. Four functions. 20,000+ teams in Slack and Microsoft Teams already operate this way.

"Like adding another employee. Or two."

- Krysta, landscaping owner who built 63 workflows in two weeks.

Most mid-market companies have not started with AI.

RSM’s 2025 mid-market research found that the majority of mid-market companies have not adopted AI, and a still-larger majority describe themselves as only “somewhat prepared” to implement it.

BetterBrain runs an AI Readiness Assessment built for that gap. It surfaces where AI is actually already in use across your organisation (almost always more than the leadership team knows), names the highest-leverage starting point, and produces a defensible plan you can take to a board.

You will get a clearer answer in twenty minutes than most companies get in a six-month strategy engagement.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

The creator economy is dead. The expert economy replaced it.

Chess Move

The what: A TLDR explanation of the strategy

Back in 2010, the creator economy was a maximise-distribution game played on borrowed attention.

You probably know how the story goes:

• Platforms change

• Organic reach plummets

• Growing your audience means constantly chasing algorithms

But as the industry matured, it’s created space for a middle-class of creators.

A new model for indie media / education, built on 4 pillars:

1. Distribution of tribal knowledge developed from…

2. Earned authority and delivered to…

3. Niche, but highly engaged audiences via…

4. Productised offerings filled with value

The “expert economy”.

The most interesting 'creators' I follow generally aren’t the biggest ones.

But they do share those same traits.

1. They have a unique POV stemming from…

2. Expertise I respect, and deliver it to…

3. A small community who love their stuff, that powers…

4. A productised business that's itself proof of what they teach.

They walk the walk.

For a decade, the metagame was "build an audience" and reach was the proxy for value.

More viewers/subs/impressions was the north star.

But that proxy is breaking because AI is commoditising top-of-funnel content, and audiences worth building have stopped rewarding volume.

They reward judgement, taste, experience, and the kind of credibility that only shows up when your own business reflects the worldview you're pitching.

That's the bet Kajabi is making.

Once a course platform → today the operating system for expert-economy businesses.

5 new products to usher in the next era of online education.

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Your favourite creators are rarely the biggest creators

Think about it.

The creators you keep coming back to aren’t the ones with the biggest audiences. They’re the ones who you trust.

The marketer who teaches positioning → runs a business that is tightly positioned.

The early-stage investor teaching founder selection → has actually invested in generational founders you've heard of.

The health coach who you resonate with most → less juicy meathead, more nutrition + lifestyle + workouts that align with your aspirational self.

I think about this paradigm a lot.

Mass market + rides thumbnail/format/hook trends + 1,000,000 followers = 0 followers who consider that creator their #1 favourite creator

Niche market + always brings something unique to the table + 10,000 followers = 100s or even 1000s of followers who consider that creator their #1 favourite creator

Same goes with musicians, authors, etc

Audiences worth building may be smaller and more self-selecting, but are often far more valuable and sustainable in the long run, because those people actually share your work, buy your products, and enjoy your content for life.

In 2026, the strategy for 99% of people playing the great online game:

→ Stop optimising for the follower count

→ Start optimising for delivering maximum unique value

2. Why I built StrategyHub, and what most people get wrong about strategy

Many people see strategy as archaic, academic, and theoretical. It’s five-year plans gathering dust in a GDrive. It’s Harvard Business Review case studies that put you to sleep.

But after writing 150+ issues of this newsletter - I can tell you that’s simply not true.

Strategy, done properly, is research + data + hypothesis generation.

It’s taking a complex question, rapidly generating thoughtful hypotheses, and backing the best one using evidence from real data.

StrategyHub is the curriculum of that method.

A course that teaches the underlying skill so you can apply it to your own company, your own decisions, your own career.

Kajabi was the only platform I tested where the productisation work didn't fight me.

I had years of tactics, frameworks, tools, and methodology written down already.

What I needed was something that could ingest all of it, multimedia formats, package and price it in a way that delivers the underlying promise, and then keep helping me extend it as the catalogue grew.

A peak behind the curtain.

Now, with the addition of Kajabi Cofounder, I can keep extending the catalogue.

Cofounder reads my existing content, products, audience, and pricing history, then drafts new assets that align with the curriculum. Layer in a baseline understanding of what's converting for the thousands of other experts on the platform, and it's less "AI tool" and more "strategist who already knows the business."

Some use cases I’m thinking through atm:

Launch the next product → I describe the idea, Cofounder drafts the offer, the landing page, the email sequence, and a launch plan calibrated to my list size and past conversion rates.

Add to the library → it surfaces the natural next module, workshop, or follow-on bonus based on what's already there and what readers keep asking about.

Spin up lead magnets → it takes assets I've already written (lessons, templates, etc) and repackages them as standalone freebies, with opt-in copy that pulls them into the funnel.

Sharpen what's already live → it reads existing pages, flags the under-performers, and rewrites the copy in my voice.

Build-out stops being the bottleneck.

3. Trust is the new moat

Let’s zoom out of StrategyHub for a second.

AI is making content abundant, meaning trust is the thing that’s still scarce.

Kajabi’s bet: "infrastructure for productised trust" becomes one of the most strategically valuable products they can own in their category.

Here's the thing most creators miss though: trust is a curve, not a number.

Every interaction someone has with you is a trust transfer. The whole game is traversing the curve:

→ Social impression: "I’m aware you exist"

→ Signup: "I'll let you into my inbox"

→ Welcome sequence completion: "I'll keep reading"

→ Landing page click: "I'm interested enough to evaluate"

→ Purchase: "I trust you enough to pay"

→ Product engagement: "My trust was justified"

→ Ascension to a high-ticket offer (if you've got one): "I trust you enough to go deeper"

→ Testimonial or referral: "I trust you enough to vouch for you to strangers"

For 18 months, I've been guessing where trust was leaking on mine. Google Analytics over here, Stripe over there, Zapier somewhere else, a Notion doc trying to stitch it all together.

The new Kajabi products are aimed squarely at improving this trust journey. 2 interesting examples:

Backstage. The most trust-rich products on the internet are still 1:1 coaching relationships, and they typically live in the chaos of Zoom + Notion + Calendly links. Backstage gives them a single home, assigned curriculum, direct chat, live or recorded sessions, that feels polished, brand-native, and high-value.

(No plans to launch a high-ticket offer myself, just an interesting move for Kajabi. These programs are getting much more popular in the post-AI market.)

And then there's the Kajabi MCP, which is the one I'm most excited about.

Full disclosure, I'm a certified MCP addict so plugging Claude or ChatGPT directly into my Kajabi account to query data or build workflows is kind of massive.

First thing I’m building: A custom dashboard that finally visualises my end-to-end trust curve in one view.

TOFU social impressions → email signups → welcome sequence completion → clicks to the StrategyHub landing page → purchases → product engagement → testimonials and referrals.

Gears are already turning…

My particular dashboard isn't really the unlock. The unlock is that you get to architect it however you want.

Kajabi has its own dashboards and its own UIs, and they're great defaults.

But every business is different. Different funnels, different offers, different tech stacks, different tiers. With the MCP, Kajabi becomes the underlying infrastructure for your version of productised trust. You operationalise and visualise it however your business actually works with the MCP.

That’s about it from me on this. Shout out to the Kajabi team for jamming with me. Got Strategy Breakdowns readers 50% off all Kajabi 6 month plans if you want to start building today:

Rabbit Hole

The where: 3x high-signal resources to learn more

The Kajabi MCP launch + product tour ↗

[5 minute read / 10 minute watch]

If "manage your business through a conversation" still sounds vague, the launch demo makes it concrete.

→ A Claude or ChatGPT prompt that drafts a landing page, an email sequence, and an offer inside Kajabi in under a minute

→ The 6 surfaces the Kajabi MCP covers at launch: Pages, Emails, Offers, Contacts, Orders, Blog posts

→ The Kajabi MCP builds it, you launch it

1,000 True Fans ↗

[8 minute read]

The 2008 essay that called the expert economy 18 years early. Kevin Kelly's thesis: you don't need millions of fans, you need 1,000 true ones willing to pay you directly.

→ Why the maths works even at small scale

→ What "true fan" actually means as a definition

→ The infrastructure gap that took 18 years to close

If you only read one piece on why audience size is the wrong number to chase, this is the one.

The Passion Economy and the Future of Work ↗

[12 minute read]

Li Jin's 2019 macro framing from Andreessen Horowitz. Reads as oddly prophetic in 2026.

• How individuals capture value from platforms

• What kind of infrastructure the trend demands

• The shift from one-size-fits-all to monetising niche expertise

Good read if you want the macro tailwind underneath everything in this breakdown.

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 125,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

Hey hey - Tom here 👋 

Quick callout today - I'm hiring, and the best person for this job probably reads this newsletter (or knows someone who does)! 30 seconds, then back to your day.

I've sold $500k+ in newsletter sponsorships in 24 months.

Without a sales team. 95% inbound + renewals.

Now I'm hiring my first Partnerships Manager 🙌

(Not for you? Forward this email to someone who should apply)

What makes this role interesting:

→ Legendary audience: 125k (+10k/mo) readers at Google, Meta, Netflix, Stripe, etc
→ Crazy sponsor lineup: Including Atlassian, Notion, Hubspot, Clay, Vanta, Attio, PostHog, ElevenLabs
→ Constant inbound: My audience brings in leads organically - you don't cold-start from zero
→ Existing systems: Pre-built skills + agents for outbound, proposals, CRM, follow-ups etc
→ Long-game: Same commission % on new business AND renewals (you keep getting paid on what you build)

What I'm looking for:

A closer who builds. Someone who can run a $100k annual partnership call in the morning and ship an enrichment agent in the afternoon.

Someone with high-agency who can read a multi-stakeholder situation, and then drive the next 10 actions: scoping proposals, follow-ups, stakeholder mapping, partner strategy, internal coordination, expanding existing accounts, etc.

You'll own and grow our sponsorships from $40k/mo to $100k+ - new business, renewals, inbound, outbound, proposals, calls, and the CRM + automation stack underneath it.

🔮 AI-native is non-negotiable.

Ideally full-time, part-time possible.

Bonus points if you've worked in media sales (newsletters, podcasts, events etc), but open to all sales backgrounds.

Important note: Our sponsors don't see us as a media buy - they see us as a partner. They DM me for advice beyond their campaigns - hiring, social tactics, broader strategy, etc. This is fundamentally a relationships role, not a transactions one. The person I hire must vibe this way too.

What you get:

→ 20% commission - uncapped. No base
→ Work directly with me, fully remote from anywhere in the world
→ A pre-built sales engine to plug into
→ Front-row seat to a fast-growing internet media biz
→ Freedom to follow your curiosity and shape the role over time
→ Access to pro versions of best-in-class tools (Claude, ChatGPT, Notion, Attio, Wispr Flow etc)

If you can close AND build, full role + application here

If someone you know would be perfect - forward them this email 🙏 

Either way, thanks for reading, and looking forward to jamming if this role sounds like you!

— Tom

I’d bet 33.3 million Sonnet input tokens that the reason you’ve heard the name Intel is because you had a chunky tower / laptop brick with one of these stickers on it.

The little swirl that lived on essentially every single PC for two decades.

I assumed it was the output of some carefully constructed 5-year brand plan.

It wasn't.

It was the outcome of Intel losing a court case nobody remembers.

And it accidentally became the most effective sticker-marketing campaign of all time.

Here’s how it went down.

— Tom

Your newsletter is sitting on data you can't see.

beehiiv's MCP connects Claude directly to your account - subscriber behaviour, churn signals, upgrade intent, all of it.

Ask: "Find free readers who opened 5+ times but never upgraded." Get the segment instantly.

No exports. No dashboards. Just answers. Setup takes 2 minutes.

Stop guessing what to build next

Shipping fast doesn't matter if you're shipping the wrong thing. Customer feedback is buried across tools, prioritisation comes down to the loudest voice in the room, and engineering builds without the context behind the decision.

Atlassian's Product Collection brings customer insight, prioritisation, and delivery into one connected, AI-powered system - so product teams can make better decisions, not just faster ones.

_{Thank you for supporting our sponsors, who keep this newsletter free.}

The best ‘sticker marketing’ strategy in history

Chess Move

The what: A TLDR explanation of the strategy

In March 1991, a federal judge in San Jose told Intel they couldn't trademark the number "386."

It was generic, so AMD (2nd largest US chip manufacturer) could use it, and create their own chips named “386”.

The entire naming convention Intel had built its brand on (286, 386, 486) was suddenly un-protectable.

CEO Andy Grove gave his VP of Corporate Marketing, Dennis Carter, the weekend "to figure what [we] would do about it."

Carter realised the court had handed Intel a deeper problem than a lost trademark. As long as Original Equipment Manufacturers (“OEMs”, i.e. PC-makers) were the customer, AMD's cheaper "386" would always win.

The fix was to create demand at the end consumer - the person at Best Buy who'd never thought about b2b chip manufacturers in their life. 

If they learned to ask for Intel by name, OEMs couldn't switch to AMD without losing the sale.

They came up with a strategy to win consumer mindshare by paying every PC maker on earth to advertise Intel for them.

Here's the playbook.

→ Intel set aside 3% of its microprocessor revenue and poured it into a co-op marketing fund.

→ Any PC maker who wanted to dip into it could get up to 50% of their print and TV ad spend reimbursed by Intel.

→ The only conditions:

1. The ad had to feature the Intel Inside logo.

2. Every machine had to ship with the Intel Inside sticker.

12 months later 300 Original Equipment Manufacturers (’OEMs’) signed up.

Another 12 months later, 500.

That year, 70% of all eligible PC ads were featuring intel’s logo, and their sales jumped 63%.

European awareness of the Intel brand climbed from 24% in 1991 to 94% (!) by 1995.

Advertising Age called it:

"the most effective co-op advertising program in history... a stellar run in which the chip giant built a brand, influenced a generation of PC users."

Breakdown

The how: The strategic playbook boiled down to 3x key takeaways

1.  Stickers live a long life

Nobody had ever ingredient-branded a microprocessor before 1991.

Intel borrowed the idea from Dolby (on your stereo), Nutrasweet (in your Diet Coke), and Lycra (in your gym fit).

The pattern across each of these: a B2B supplier that pays the end-product brand to display it.

Intel’s version came with one extra ingredient. The sticker.

The TV ads Intel funded were only paid for once → The sticker drove continuous brand marketing for the lifetime of the devices they shipped on.

Every Dell, every Gateway, every Hewlett-Packard became a physical billboard the buyer carried home, sat on their desk, and looked at daily for the next 5 years.

Ongoing media placements, paid for once.

The 3% of revenue Intel spent on the co-op fund bought millions of square inches of permanent ad space inside the most valuable real estate on earth: the customer's home.

Not a temporary Google Search Engine Results Page ad that only displays against specific search terms, before users click away.

Not an Instagram ad that gets scrolled past in 0.15 seconds.

A physical sticker, within eye-shot, for 100% of the time the end-user is online.

2. Audio identity is the forgotten brand vector

To compound the success of the sticker campaign, in 1994, Carter commissioned an Austrian composer named Walter Werzowa to write a sonic logo.

Five notes.

D♭, D♭, G♭, D♭, A♭. The rhythm was lifted directly from the syllables: "In-tel In-side."

Still pure heat 🔥

The result was one of the most iconic jingles of all time.

In their focus group studies, 80% recognised the melody when played on a violin but 100% recognised it the moment it was played with Intel's actual sound.

Including when researchers deliberately added a wrong note.

By the late 90s the Bong was playing somewhere on earth roughly every 5 minutes.

For the marketers reading this: Your visual brand competes with 1000s of logos a person sees per day.

But an audio brand competes with the small handful of jingles a person hears that week.

The acoustic environment is far more emotionally direct, and far less crowded than the visual.

(Quietly humming for a Strategy Breakdowns jingle that pops 🎵)

3. Choose your layer

Compaq and IBM both hated Intel Inside.

The campaign told consumers that every PC running a given Intel chip was fundamentally the same. The premium machine, the budget machine, the gaming rig. All of them. Intel inside.

By centralising brand value at the chip layer, Intel deliberately commoditised the layer above it.

The PC OEM became a swappable assembler of someone else's components.

And once buyers started shopping by what was inside the box rather than whose name was on it, the OEMs lost their pricing power.

Intel was even running Super Bowl ads, just to advertise their chips.

Over the next few years, Compaq and IBM (the 2 largest PC manufacturers) exited the PC business.

The chip had become more famous than the computer it lived in.

Every value chain has one layer that owns the customer's mindshare.

Most components stay anonymous because the assembled product gets all the attention.

Intel spent 3% of revenue making sure they became the layer the buyer remembers.

Rabbit Hole

The where: 3x high-signal resources to learn more

Dennis Carter: Behind the Intel Inside campaign ↗

[15 minute read]

The landmark interview with the guy behind the playbook.

Dennis Carter was Andy Grove's first technical assistant.

Then he weekend-built the most successful co-op ad program in history. This is his version of the story.

Ingredient Branding: The Intel Inside Story ↗

[3 minute read]

The official version. Intel's internal historians walk through the launch and campaign in their own words.

Includes a very cool 1991 original intel print ad.

1 of 16 pieces in their ‘Virtual Vault’. Click around. Super cool stuff in there.

The Most Expensive 3 Seconds in Tech ↗

[8 minute read]

How a former Austrian electronica artist scored a melody that became more recognisable than most national anthems.

Stay for the clip at the end.

If you made it this far - (a) mad props - appreciate you, and (b) you’re part of a secret club of readers that gets to submit ideas for future breakdowns. Just reply to this email (we always write back!)

Enjoy the rest of your week 👓

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 125,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

If you read my 3-part series with Tanguy Crusson last year, you already know Jira Product Discovery ("JPD") is one of Atlassian's fastest-growing products. Ever.

A few months back, Atlassian acquired the tech and team of Cycle - a 6yo startup that auto-captures customer feedback from across your stack and turns it into something a PM can actually act on. The capabilities were rebuilt into the new 'Feedback App', bundled with JPD and Rovo (Atlassian's agentic platform) inside the new 'Product Collection'.

I knew I had to wrangle my ex-Atlassian status into a sit-down with Cycle's founder, Mehdi Boudoukhane, now Principal PM for Feedback to get the low-down on:

The real acquisition story.

Atlassian's swing at AI usage-based pricing.

The unfair distribution math of selling into a 300,000+ customer base.

Enjoy.

– Tom

Don't SOC-block your best engineer

Your software needs to be compliant to win deals. But you also need your engineers focused on building your product — NOT pulling SOC 2 evidence.

The Vanta Agent is the sharpest GRC engineer you've never had to hire, working tirelessly across the platform to draft policies, complete questionnaires, and flag issues before they escalate.

Fast-moving companies like Ramp and Cursor use Vanta to get and stay compliant, simplify their audit process, and unblock deals — so teams can get back to building.

What if your AI showed up before you asked?

Viktor doesn't wait to be tagged. He flags contract blockers, surfaces ad performance, reviews PRs - before your team thinks to pull them. 

Half of paying teams spread him across 5+ Slack surfaces in their first 30 days. Strategy Breakdowns readers get $100 free.

_{Thank you for supporting our sponsors, who keep this newsletter free}

Why Atlassian acquired Cycle

Breakdown

The how: The strategic playbook boiled down to a few DMs

TA	Wassup Mehdi! Thanks for jumping on. I’ve heard a lot from the team about the acquisition. V excited for the inside scoop today. Take me back to the start. Why did Cycle exist in the first place, and how did the Atlassian conversation actually begin? 8:00 AM ✓

MB

Thanks for having me! Loved the last JPD interview series, so it’s cool to be here 🙂  The story is essentially this: I was a PM and I was frustrated with the tooling that was available to me, because everyone was talking about being 'customer obsessed' but the customer was nowhere to be found in the stack PMs used day-to-day. The most important thing in building trust with customers (and internal customer-facing teams) is making sure they feel heard. That when they share feedback, it doesn't enter a black box. That closing the loop was so important, and so hard to do. So I thought, okay, there's an opportunity to obsess over this problem. We did that for 6 years. The first 4 were pre-AI - very manual, PMs spending hours tagging insights. Then AI came and we started shipping the feedback autopilot. Capturing from Slack, JSM, Zendesk, Zoom calls, all into one place. The autopilot extracts insights, links them to ideas, closes the loop automatically when an idea ships. That's when we hit product market fit. And that's about when Tanguy from JPD reached out. 8:08 AM ✓

MB

In that first discussion, Tanguy was just curious. “Oh, that feedback use case, that's interesting”. No urgency. Then a year later, he noticed it was becoming the single biggest pain point in JPD customers. They were hacking feedback together themselves, polluting backlogs with ideas that should actually be feedback. So from the Atlassian perspective: they need this. Do they build it or acquire? They acquired us because it was effectively a 6-year head start on a problem we'd been obsessing about. 8:11 AM ✓

TA	I heard through the grapevine that Cycle had a few hundred customers vs JPD's 20k+. Pretty much a rounding error. So the acquisition probably wasn't really about the customer base. How do you see the acquisition from their pov? 8:13 AM ✓

MB

I think about it like this: in an acquisition you can buy 3 things: the business (customers, ARR), the product, and the team. In this case Atlassian didn't want the business. It was too small. It was really the other 2. When we joined, there wasn't a single shipped AI feature in JPD. We'd been obsessing about AI for 2 years, processing millions of pieces of feedback every month. That know-how, all the thinking that went into building a feedback app the right way - that's what they bought. And with it, the team that's going to take that know-how and rebuild it within Atlassian. 8:17 AM ✓

TA	When you joined - did you end up sunsetting Cycle? I imagine that would’ve been a hard call. How did that conversation play out? 8:18 AM ✓

MB

Yea we did - probably the hardest part of the journey! We had customers who'd built entire workflows around Cycle. The reaction to the shutdown told us how much they'd come to rely on it. But ultimately it was the right decision. Tanguy had been through other acquisitions before, and one of his learnings is that when you keep the legacy product while trying to integrate and rebuild, you end up maintaining 2 products at the same time. Very costly. Ultimately it didn’t make sense to spend a lot of effort maintaining a legacy product when we should be entirely focused on the future. The goal is how fast can we get from 20k to 100k customers together, and keeping Cycle running wasn’t a viable component of that strategy.

TA	And from your side as a founder, what tipped you toward joining vs staying solo? 8:23 AM ✓

MB

We had product market fit - the product finally worked with AI - but distribution was a huge challenge. I was spending 80% of my time on sales and content. I'm a product guy. But I really didn't have much time anymore to spend on the product. Atlassian has 300,000+ customers. Distribution wouldn't be a problem. Scale wouldn't be a problem. Joining meant I'd be able to focus entirely on the product. By joining forces, we'd scale our impact many magnitudes further than continuing as a standalone tool. 8:26 AM ✓

TA	Before we get into how it bundles up - walk me through the actual loop for how feedback happens inside JPD. Step 1, 2, 3 - what's the opinionated POV? 8:27 AM ✓

MB

Here's the model end-to-end: It starts with capture. Feedback comes in from everywhere - Slack, Zoom, JSM, Zendesk, Teams, Gmail, custom APIs. We pull it all into one place. Then AI does the heavy lifting. It extracts insights, summarises the customer voice, surfaces the patterns: That feeds into voice-of-customer dashboards and JPD ideas. Now here's our opinionated bit. In a JPD idea, you can write a release note - even before you ship the feature. Apple-style. The release note lets you get to the essence of the outcome you want to achieve. Each idea is linked to Jira. When the work items move to done, the idea moves to done, and the release note can auto-publish to a public changelog. That’s where it closes: each release note is linked back to every piece of feedback that informed it. Slack, Zendesk, JSM, HubSpot, wherever it came from. In 1 click, you can fire automated communication back to all of them. Maybe you shipped one release note, but you had 150 feedback items linked to it. One click, and you're back to all those folks: ”Hey, we heard you. 2 months ago you said this. Look at the release note, tell us what you think.” When you do that systematically, you create a virtuous cycle. Customers turn into co-creators. They want to give you even more feedback. You build trust. 8:33 AM ✓

TA	Love it. So the loop also closes back to the internal teams who logged the feedback, right? Sales, support, etc? 8:34 AM ✓

MB

Exactly. That's the 2nd-order effect. In B2B, most feedback comes from internal teams - support, CSM, sales - because they spend the most time talking to customers. So the loop has to close back to them too, not just the end customer. Once it does, the math changes. Sales gets a notification straight in Salesforce: “this deal was blocked because the integration hadn't shipped. Well, it just shipped. Here's a reason to reach out.” CSM gets the same on retention. Gets them closer to their KPIs. Without that loop, sales just stops giving feedback. They're busy people with their own KPIs, and if it disappears into a black box they don't bother. The loop isn't a nice-to-have. Without it the whole system breaks. 8:40 AM ✓

TA

So the loop literally helps the sales team hit quota. Love that. Tell me about this new ‘Product Collection’ I keep seeing. When I was Atlassian we spent a lot of time thinking about the best way to bundle / combine / split / position products. Would love to know what the latest thinking is. 8:41 AM ✓

MB

The Product Collection is our bundle for product teams. It includes JPD, the new Feedback App, and Rovo - and there might be other apps later down the line. Rovo enables these next-gen use cases we’re obsessing over atm like agentic roadmapping. It positions us head-to-head with the roadmap-plus-feedback players in the category. The Feedback App brings the qualitative side - what people actually say about your product. JPD brings the roadmap. Rovo brings AI agents on top. Imo the other piece that needs to be in there longer-term is product analytics - the quantitative side, what people are doing with the product. You need both to make better product decisions. For now we're integrating with Pendo and other product analytics tools. 8:48 AM ✓

TA	So how is the new Feedback app priced? 8:49 PM ✓

MB

We're experimenting a lot because AI pricing is a totally new frontier. How much does it cost us in tokens to serve a customer? What's their willingness to pay for the value those tokens bring? But we’re excited because it might end up being Atlassian’s first swing at usage-based pricing. At Cycle, we removed seats and went 100% usage-based, with AI credits as the unit. The key insight is the value chain - the cost to us in tokens, the price we charge, and the actual value the customer gets (PM time saved, faster decisions, better prioritisation). When that chain has 100x between cost and value, you have a lot of room in the middle for healthy margin while delivering incredible ROI to the customer. When we made the switch at Cycle, engagement and ACV really skyrocketed. 8:56 AM

MB

The other key thing - the Feedback App isn't a copilot, it's an autopilot. So tokens are predictable. If a team has 3 million feedback items per year, that's 3 million credits minimum. Every piece gets processed by AI, extracts insights, links them to ideas. You can predict revenue fairly well when it's autopilot. With copilot, you can't estimate how many times people are going to ask questions to your AI. Autopilot is event-based, not user-input each time. 8:59 AM ✓

TA	What about distribution? JPD is past 20,000+ customers off the back of cross-flow into Jira and the wider Atlassian ecosystem. What’s been your first taste of that playbook as a founder coming in from the outside? 9:00 PM ✓

MB

I have crazy anecdote on this. We did a little Pendo form. “Hey, we're building the Feedback App. Who's interested in joining the early access program?” We built it in 15 minutes. Sent it just to JPD users, not even all of Jira. In a few days, we had 7,000 teams asking to talk to us. For a few weeks, we ran 100 customer interviews a week. This is my colleague’s actual calendar from those weeks - back-to-back Early Adopter Program meetings! 9:06 AM ✓

MB	As a startup, we'd been struggling to get 5 or 6 interviews per week. We joked with the team: why are people complaining so much about distribution? Just send a Pendo form, get 7,000 signups. Easy. 9:08 AM ✓

TA

<Quietly devastating to anyone in startup land trying to book user interviews 😂> Let’s go contrarian for a second. Everyone's talking about AI-native product discovery, the solo PM-builder, the one-person billion-dollar company. What's your take on all this as someone who’s building for builders? 9:09 AM ✓

MB

I actually think there's a broad misconception. What folks are building is mostly prototypes, not working software. We don't hear enough about the difference between a localhost demo and software that actually ships - every edge case accounted for, multiple users, concurrent sources of truth. What’s changed most is the ability to prototype fast. PMs can test 10 ideas before landing on the one that gets shipped. But the separation of roles will remain. Designers, PMs, engineers all do more, especially on prototyping. The handoff persists at scale. 9:15 AM ✓

TA	Last big one. Anything you'd do differently? 9:16 AM ✓

MB

Post-acquisition, we shut down the product. That was a good decision. The right call, for sure. But that meant we also shut down all the AI pipelines that were running. We'd had a continuous R&D engine - what works, what doesn't - and we shut everything down. That created a gap of several months without AI R&D. In this world, several months is a lot. Then there's inertia in restarting. I'd have kept at least 10 customers. The cutting-edge teams really pushing AI feedback workflows to the limit. There's a concept I really like: the future already exists, it's just not evenly distributed. We had a couple of teams that were representative of that. They're building the future - they're doing what Atlassian customers will do in a few years. We should have found a way to keep them, keep the AI pipelines running, and not break the momentum. 9:22 AM ✓

TA	That's a great one to land on. Mehdi - thanks for doing this mate! 9:23 AM ✓

MB	Pleasure. Ping me anytime if you need more! And for the product teams reading this - check out our new Feedback App and Product Collection here :) 9:24 AM ✓

Want more interviews? Let us know by voting and replying to this email!

Whenever you're ready, there are 3 ways we can help you:

1. StrategyHub

Our flagship course on how to use free internet data to make better strategic decisions. Contains 5 years of strategy expertise, proven methods, and actionable tactics to accelerate your career with modern-day strategy skills.

2. Partnerships

We have a growing audience of 55,000+ strategists from top companies like Google, Meta, Atlassian, Stripe, and Netflix. Apply to feature your business in front of Strategy Breakdowns readers.

3. Tech stack

One of the most common questions we get asked is: “What tools do you use to run Strategy Breakdowns?” So, we’ve open-sourced our tech stack to give you an inside-look at exactly what tools we’re using to power each corner of this operation.

Think about the AI tools we use every day.

ChatGPT lives in its own tab. Claude has its own desktop app. Cursor sits alongside your IDE. Devin. Lindy. Always the same pattern.

The agents that promise to "work alongside you" tend to open in a private window, just you and the model. 1 human. 1 assistant.

This is a choice the entire industry made by default after ChatGPT, and almost nobody has questioned it.

Viktor is one of the only AI companies betting against it, and they’ve scaled to $15m ARR in just 10 weeks (!!) by doing so.

Today's breakdown is about what I'm calling ‘multiplayer AI’. 

→ How Viktor went team-native by killing their own web app in two days

→ Why 13,000+ teams hired Viktor as a colleague rather than a tool

→ Why I think the rest of the AI stack is quietly going to follow.

I've been running Viktor inside the Strategy Breakdowns Slack workspace for a while now, and the thing that keeps surprising me isn't what it does when I ask.

It's what it does when I don't.

Shoutout to to the newest member of the Strategy Breakdowns team: Karen. Viktor is fired up.

Let’s get into it.

— Tom

P.S. Huge thanks to Antoni and the Viktor team for the candid behind-the-scenes context that made this one possible.

Don't SOC-block your best engineer

Your software needs to be compliant to win deals. But you also need your engineers focused on building your product — NOT pulling SOC 2 evidence.

The Vanta Agent is the sharpest GRC engineer you've never had to hire, working tirelessly across the platform to draft policies, complete questionnaires, and flag issues before they escalate.

Fast-moving companies like Ramp and Cursor use Vanta to get and stay compliant, simplify their audit process, and unblock deals — so teams can get back to building.

Own, store, and query your data in one place

PostHog is launching a managed DuckDB warehouse that's fully integrated with your product analytics. 

Get direct database access, built-in SQL editor, modelling, and visualisation all-in-one, instead of a fragmented stack. 

_{Thank you for supporting our sponsors, who keep this newsletter free.}