WEBVTT

1
00:00:00.100 --> 00:00:16.640
[on-hold music] Welcome to the Rebooting show. I am Brian Morrissey. I am joined by Adam Singolda. Adam is the CEO of Taboola.

2
00:00:17.200 --> 00:00:31.580
Adam and I have been pre-preparing for this podcast for the last fifteen plus years because I've known Adam a long time, all the way back when Taboola was not actually a content ad network on web pages, but you guys started in video, I remember.

3
00:00:32.000 --> 00:00:32.699
Right. It was...

4
00:00:32.800 --> 00:00:52.510
The very beginning was a video recommendation engine, which only took me four years to realize that it's really hard to sell video recommendation software to publishers, and we then expanded to be kind of, you know, anything consumers want, may want to do next- Right...

5
00:00:52.520 --> 00:01:02.280
whether that's a piece of content or a piece of paid content. That was, that was the beginning. And I remember us talking and I, I, I think it was you and I. But it was...

6
00:01:02.600 --> 00:01:12.120
I was, I was saying like, "But wait, this isn't releve- Uh, uh, the, uh, this stuff doesn't look relevant." 'Cause relevancy was, like, everything. And you're like, "No, no, no. It's not, it's not relevancy.

7
00:01:12.440 --> 00:01:22.790
It's, like, whether people are going to take an action on it necessarily." Because think about how early days this was. Today with TikTok you have For You page- Yeah...

8
00:01:22.820 --> 00:01:26.930
which is, which has nothing to do about anything contextual. It's only about- Right...

9
00:01:26.940 --> 00:01:35.300
what I may want to do next, which TikTok has done a phenomenal job, sadly I gu- I suppose, predicting what I may wanna do next as a personalization engine.

10
00:01:35.560 --> 00:01:48.380
Back then, it was in the back of, you know, Google contextual ads and search engines. Yeah. So back then, contextual was, was technology and we tried to bring something that was more about you may like. Right.

11
00:01:48.960 --> 00:01:57.620
You know, and that was, that was new. Which is very common now. And I think the other thing that... Like, 'cause then it became you guys and Outbrain, right? It was like the Bloods and the Crips, you know?

12
00:01:57.740 --> 00:02:06.070
But of the web. You know, nothing like a good two Israeli company- [laughs] Exactly... have, ha-having a good time. Come on, Brian. This is, you know. I know. I get it. I get it, right? And you, you wanna have that.

13
00:02:06.180 --> 00:02:14.200
Like, you know, back then it was, like, s- what it, like, it was Tacoda versus... What was the other one? Anyway, in be- behavioral targeting. Yeah. Yeah, yeah. It was just- Behavioral targeting. Yeah. Yeah.

14
00:02:14.260 --> 00:02:32.260
But, like, I think what was really interesting then was that in some ways, not as big of a trick, but it was like Google in some ways in that Google took the web pages that were actually valued the least at the time, which were search results pages originally were remnant inventory.

15
00:02:32.290 --> 00:02:37.720
It was like, "V- yeah, just put some extra banner ads on there because they don't, like, monetize like a homepage or something."

16
00:02:38.340 --> 00:02:53.549
And in some ways you guys pulled a, a little bit of a similar thing by taking real estate that was thought of as a place for just, like, remnant ads, very low value, but that's when people make their decision about what to do next, and you can, like, funnel traffic- Right...

17
00:02:53.549 --> 00:03:02.380
in different directions. Do I have that right? There, there were, th-there were two counterintuitive kinda eureka moments. Go back to 2012.

18
00:03:02.440 --> 00:03:13.680
At least for Taboola, 2012 was the beginning of what we realized we're on becoming, we're onto something, and it started growing really, really fast from that moment.

19
00:03:13.690 --> 00:03:22.750
The first one was, as you said, when someone finish reading an... When they finish reading an article, they have this moment of next when they're about to do something. Yeah.

20
00:03:22.760 --> 00:03:32.690
They wanna read more thing, watch something, or make a decision. And, uh, you're right. The back then it was people were more in a viewability state of mind, so if it was top of the page it was good. Right.

21
00:03:32.700 --> 00:03:37.520
But nobody clicked anything at top of the page. And the bottom of the page, no one saw that, that's what people thought.

22
00:03:37.980 --> 00:03:46.240
But what they didn't realize that even though only 30, 40% of the people get to the bottom of the page, the engagement could be 10X higher. So that was number one.

23
00:03:46.300 --> 00:04:02.070
The second thing that was quite innovative for us was that advertising can be performing on publisher sites like it does on search engine or Meta or Facebook back then. We, we did not... It's funny to go back in time.

24
00:04:02.540 --> 00:04:10.570
Back then people talked about content marketing. There, it was all about- Yeah, yeah, yeah... you know, even Digiday, if you remember, you, you had... You made a good business- Absolutely...

25
00:04:10.670 --> 00:04:19.420
out of p-people pr- you know, that was you. That was you. Yeah. So but we wanted to, to show that actually, you know, if you can click...

26
00:04:19.820 --> 00:04:29.090
If you click on something bottom of the article through Taboola, you may actually make a decision and an advertiser could be measuring that experience like they do when they buy from Google and Facebook. That was new.

27
00:04:29.140 --> 00:04:40.910
Yeah. And, and those two things, you know, became a rocket ship for Taboola. We grew from $100,000 a month in 2012 to over $200 million in 2014 just for those two reasons. Right.

28
00:04:41.290 --> 00:04:51.380
And I think in some ways, and tell me if, if you disagree with this, like, it was kind of a attention arbitrage. And, and by that I mean that, like, the attention, like when you said, like, the top of the page was...

29
00:04:51.469 --> 00:04:57.900
And I'm... This is gonna move into, to the Taboola of today, this little walk down memory lane Brian. That's very good. We're having, you know, this, uh- [laughs] Yeah, a little old school.

30
00:04:58.060 --> 00:05:04.400
We- Two old guys sitting around. [laughs] Memory lane, memory lane matters, you know? No, but I do.

31
00:05:04.500 --> 00:05:18.610
You know, maybe it's 'cause I was, like, a, a, a, a history major, but I do think it does matter 'cause it informs what Taboola is today and where it goes next in that, like, it was taking the attention that was, like, the top of the page was, was being valued for the attention, not for the performance, right?

32
00:05:19.090 --> 00:05:26.230
And usually in arbitrage, performance is i- performance always goes to, like, arbitrage opportunities, right?

33
00:05:26.260 --> 00:05:42.560
And, like, the bottom of the page, and by this arbitrage opportunity I mean it, the bottom of the page was being mispriced because it was being measured based on things like viewability versus did it lead to some action that the marketer wants it to take.

34
00:05:42.740 --> 00:05:54.020
And, like, I think what you guys found out, and Outbrain too, I'm gonna give O-Outbrain a little credit, is that the bottom of the page actually performs well. I, I, I, I wouldn't underestimate...

35
00:05:54.240 --> 00:05:58.700
I, I think you're simplifying this, but, but- Well, I'm a journalist... which is, which is... That's good.

36
00:05:58.870 --> 00:06:22.010
[laughs] But, but you're, but you're oversimplifying one important point which is at the time, including companies in our space at large, it was very counterintuitiveThat you should have the advertiser, those who pay you for that placement, have a pixel on their page, pay on a per click basis and measure, measure a CPA, a-a-a, an acquisition cost.

37
00:06:22.570 --> 00:06:33.060
Back then, it was almost perceived as- It was like gauche. It was like, look, I covered, I covered direct marketing. I, I know it. Like, and, and like the, the- It was, it wasn't-...

38
00:06:33.090 --> 00:06:40.930
the dudes I talked to at the DMA convention were completely different than the media people I talked to. Do you know what I mean? [laughs] I, I, I... We proved...

39
00:06:40.990 --> 00:06:53.150
I mean, one of our, one of our first advertisers in 2012 was a big subscription publisher that g- I'll never forget, it was one of the first five advertisers to go ahead. And we proved that we can drive subscription.

40
00:06:53.390 --> 00:06:57.320
No one tried. No one even tried. Nobody... No one tried for years.

41
00:06:57.410 --> 00:07:10.870
It was only 2014, 2015, that the ecosystem said, "Wait a second, this bottom of article placement, these native ads," back then native ads, "is really not just a good location, but it can actually drive search-like performance."

42
00:07:10.880 --> 00:07:23.060
That was the mar- It took the market two years to catch up on what Taboola did, I would say, really well, because again, at the time, people didn't think of that location as a performance kind of- Right...

43
00:07:23.080 --> 00:07:34.100
advertising opportunity. Which became, over the last decade, the most important thing marketers ever want. Right. Performance marketing has eaten the, the internet, [laughs] you know.

44
00:07:34.150 --> 00:07:41.490
And, and obviously Google's business has been driven by performance because they have intent. You type in a search term, they return...

45
00:07:41.530 --> 00:07:49.990
I'm not saying, like, it's not, like, complicated, but in some ways it's like, okay, you ty- you type in that you're interested in something, and Google's like, "Hey, he- here, here are some ads for, for, for that thing."

46
00:07:50.410 --> 00:07:56.450
And it's all, you know, performance-based. They grew on performance. Maybe YouTube is a little bit... operates a little bit differently. Right.

47
00:07:56.490 --> 00:08:01.590
But Facebook, you know, the people f- who built the Google ad system moved over to Facebook and then they rebuilt the, the thing.

48
00:08:01.630 --> 00:08:16.050
We're seeing now into retail media, it's all performance, performance, performance, performance. And the problem for publishers is that in their DNA, they've never really fully embraced, I think, performance advertising.

49
00:08:16.130 --> 00:08:25.070
It was like, we'll run performance stuff, but that's, like, down the waterfall. Like, we want, we wanna have, like, CPM. We're, we're about brand. We're not about performance.

50
00:08:26.090 --> 00:08:32.370
I, I don't think, you know, we should hold, we should... You know, publishers, to their, you know, to, to... If, if I can represent publishers for a moment here. Absolutely. What?

51
00:08:33.010 --> 00:08:42.730
[laughs] And, and, and if I may, and I will. [laughs] It's, you know, it... To, to drive a performance adverti- a successful performance advertising business, you need tens- Yes... of thousands- Yeah... of advertisers.

52
00:08:43.230 --> 00:08:45.670
It is, it is not- I got you, yeah... a reasonable...

53
00:08:45.690 --> 00:08:54.830
It's not, it's not a reasonable, if an ambition, for any one company that sells itself to have tens of thousands of clients so I can choose the one that makes sense in that moment.

54
00:08:55.150 --> 00:09:03.310
For a company like Taboola, when we manage a network of 11,000 publishers, it makes sense. So I think publishers, you know, again, if, if I may- Well, it's a matching problem, right?

55
00:09:03.390 --> 00:09:10.770
I mean, like, the, the, the direct marketing works- Exactly... when you match up an offer with a context and a person. A granular... Right. Yeah.

56
00:09:10.810 --> 00:09:16.810
So, so, so that's why it made sense for publishers to say, "We're gonna sell to a few advertisers a lot of money at a high price."

57
00:09:17.260 --> 00:09:28.870
And by the way, what we don't sell, Taboola, you should be our performance advertising arm, and in many ways actually kind of like a foundation for revenue, right? Because we were...

58
00:09:28.910 --> 00:09:38.390
We became this predictable revenue for publishers. They can s- they could say, "This is, this is... We're gonna have this going up every year, and we can rely on Taboola.

59
00:09:38.399 --> 00:09:45.330
And at the same time, we're gonna sell premium experiences, high CPM to fewer advertisers upfront, new front, all those things." Yeah.

60
00:09:45.370 --> 00:09:53.630
So I think it was actually, you know, a good synergetic relationship for, for that reason. Yeah. And like, I know maybe this is part of our...

61
00:09:53.750 --> 00:10:02.200
But like, I mean, you, you give guarantees to a lot of, like, y- particularly the bigger publishers, and then, you know, to de-risk for them, which is always...

62
00:10:02.610 --> 00:10:09.330
For publishers, you know, look, ad tech was always, you know... Publishers have always looked to de-risk. They don't...

63
00:10:09.370 --> 00:10:17.070
They historically did not pay for, for the software, so instead they would do rev shares and things like this, and then go back and complain that they [laughs]

64
00:10:17.110 --> 00:10:25.490
that, that, that their partner which took on the risk was taking too much of, of the, the profits. But, like, you gotta- Right... you gotta take risk to, to make money, right?

65
00:10:25.650 --> 00:10:42.590
I was, I was always happy to take, you know, measured risks i-in a sense that if I knew, if I had, if I had conviction that something's gonna work for us and for them, and it could alleviate, you know, that risk for publisher to make a, a decision faster, I was always happy to do it.

66
00:10:42.630 --> 00:10:47.230
And I think, you know, when you run a startup, you know, you, you... it's all about...

67
00:10:47.360 --> 00:10:56.370
and, and when you scale a startup, it's all about managing risks and trying to be creative in, in, in your technology and product and also in the way you do business with partners.

68
00:10:56.410 --> 00:11:06.270
So I think, you know, for us, we were, again, I, I believe among the first companies, to my knowledge, that came up with this guaranteed model that said, you know, in exchange for working for a long time together, which was also my ambition.

69
00:11:06.430 --> 00:11:18.740
I didn't wanna be anyone's vendor. So, so for I was... You know, look, to, to the extent we signed with Yahoo a 30-year partnership, you know, which is I'm so proud of. But I always wanted forever partnerships. Yeah.

70
00:11:18.740 --> 00:11:28.690
You know, I wanted to, to work with publishers forever. Right. Yeah. So talk to... So I, I don't wanna get in... 'cause I mean, you guys have stayed in the performance, you know, place, space, right? We've expanded.

71
00:11:28.930 --> 00:11:29.650
Right. We, we, we...

72
00:11:29.990 --> 00:11:41.250
A, a year ago, we actually s- we expanded our performance view, point of view from kind of like native advertising to all types of performance formats and tech- you know, technology, and we named this Realize.

73
00:11:41.290 --> 00:11:53.230
So we came to market with Realize, and we said, "Taboola advertising platform has a name, Realize, a new name, and you can work with us almost as easily as you work with, you know, PMax of Google or, or Meta.

74
00:11:53.410 --> 00:11:59.810
You can upload any format you want." Yeah. You, you display vertical performance, native, whatever you got, we...

75
00:12:00.110 --> 00:12:11.110
Reels, whatever you have, and import that from your Google or Facebook," 'cause we all work with Google and Facebook. "And we, we can get going. You tell us your goal, uh, and we use AI to generate creatives for you.

76
00:12:11.570 --> 00:12:22.382
We, we advise you budgets and, and..." But it's all around-... performance. I want you to compare us to Meta- Sure... so we can create, you know, incremental value.

77
00:12:22.682 --> 00:12:36.652
We, I, I think that focus really helped us because in today's world, you know, es- especially as there's so much going on in the world, you know, so much volatility, to be, to, to do one thing really well is really hard already.

78
00:12:37.042 --> 00:12:45.922
Mm-hmm. And to try to have this fant- this full funnel fantasies, which I think you hear some companies speak about, you know, we'll do top of the funnel and bottom of the... I think it's too hard.

79
00:12:45.982 --> 00:12:54.172
It's different types of companies in, you know, merged in, in one or two or three, and, and I think the market is going towards that performance kind of dynamics- Yeah... in any case.

80
00:12:54.722 --> 00:13:04.182
So we, we want to be, and hope to be, kind of the largest performance advertising company outside of Google and Facebook. So explain that because, like, I mean, I could see...

81
00:13:04.502 --> 00:13:12.282
I can remember, like, us having, like, coffee one time at, I think it was at Gray Dog, and you were, like, saying to me, you were like, "What do publ- what do you hear from publishers that they need?"

82
00:13:12.502 --> 00:13:17.072
Like, 'cause I wanna, like, actually get... I... 'Cause it's this partnership thing. You're like, "Is it, is it analytics? Is..." You know?

83
00:13:17.762 --> 00:13:27.482
And I could see you guys going in a different direction where you said, "We wanna go literally up the page and up the funnel." Up the page, up the funnel, you know? We wanna- Up the page.

84
00:13:28.282 --> 00:13:32.662
Not up the funnel, up the page, yes. I wanted, and we do. I wanted all- But you know what I mean.

85
00:13:32.762 --> 00:13:41.812
I'm, I'm, I'm s- I'm using that as, like, a metaphor in that, like, the, in the traditional sense of things, going up the page means, like, you would have more, like, branding, and that, like, instead...

86
00:13:41.862 --> 00:13:51.752
But you j- you wanted to stay y- y- if- laser focused on performance rather than have, like, brand and performance and brandformance or whatever Yeah. From space, whatever.

87
00:13:51.882 --> 00:13:58.752
[laughs] No, there's no, there's no brandformance here. Okay. So, so, so, so we do wanna have access to all the placements a publisher may have. Sure.

88
00:13:58.762 --> 00:14:03.172
And they have great placements, and some of them may work to our performance advertisers.

89
00:14:03.462 --> 00:14:16.422
That is, that is something we are after via header bidding and other connections that we have, because Taboola is literally integrated on, you know, 600 million people a day via first party, you know, JavaScript code on the page.

90
00:14:16.522 --> 00:14:25.792
We have access to other placements. That is part of our strategy because it, we may find that this homepage unit can perform for Wayfair, which is a great advertiser. Right.

91
00:14:26.181 --> 00:14:39.402
So we wanna be able to show it then a- and now. But what we don't want to do, what we don't wanna fight, we don't wanna measure viewability, completion rates, and charge based on CPM. We, we don't.

92
00:14:39.442 --> 00:14:46.302
We wanna charge on CPC, measure CPA- Yeah... and know your objective. Yeah. So we're, we're, this is who we are. Right.

93
00:14:46.382 --> 00:14:57.002
This is how we do- Because if you get that, if you get that right, and, and I think this is what, what, what Google got really right, is, you know, they became a cost of sales, not marketing or advertising.

94
00:14:57.142 --> 00:15:06.502
Like, they're... Because, like, if your performance and you, you're saying, "Hey, I'm gonna spend a dollar and I'm gonna get $3 out of it." Just like, "I'm gonna spend as many dollars as you have $3 to give me back."

95
00:15:06.582 --> 00:15:15.022
Like, I will [laughs] you know, I'll just keep it always on bud- that's what Google always said, we'll have always on budget. So that is the, you know, the goal. I wanna... Like, when you look at, like, the...

96
00:15:15.722 --> 00:15:22.022
It's, like, the unit of it, like, whatever the unit of interaction is for Taboola, right, across, like, Realize.

97
00:15:22.562 --> 00:15:40.342
How much of that is your, what I would call, like, your legacy implementation, which is bottom of the page, you might like, and then, like, you know, some options there, and, a- and you send traffic out versus- It, it's st- it's still big, but remember, I'm sor- I don't know how much of it is public, but it's...

98
00:15:40.882 --> 00:15:49.702
So remember, over the last few years- This'll put the, this'll put the podcast on the map if you get an SEC- Yeah, I-... investigation on it. [laughs] No, no, no. I'm v- I'm very practiced. I'm very practiced. Exactly.

99
00:15:50.082 --> 00:15:57.102
[laughs] So, so, you know, I, I, I don't tweet without someone reviewing my tweets. I'm just joking. That I can, that I can actually do. [laughs] So, so,

100
00:15:58.062 --> 00:16:07.482
so look, the, the, over the last three years, we're very fortunate to be partnered with companies that are very divers- like, in many ways diversified Taboola completely and changed us.

101
00:16:07.982 --> 00:16:18.762
You know, Yahoo, which is, you know, homepage and mail and oth- you know, very different types of inventory, huge. Apple News, which is incredible. Our OEM partnerships, Samsung, Xiaomi.

102
00:16:19.462 --> 00:16:28.542
About a third of our traffic is in-app now, not mobile web, not desktop. And so we have, we've changed a lot. So I would say significant portion of our revenue

103
00:16:29.462 --> 00:16:39.521
is not what it was 10 years ago when it kind of, you know, you and I, uh- Right, right, right... talked. Yeah. Well, I ask because, like, and I know it comes up in the earnings, like, I mean, look, the webpage...

104
00:16:40.562 --> 00:16:47.422
And, and I wanna, like, talk about the future of the open web 'cause your business has always been, it's been a big funder of the open web, right?

105
00:16:47.542 --> 00:16:58.192
And, like, I had a breakfast this morning with a very large open web publisher, you know, an executive of that. He was like, "I don't even know, man. Like, I know what, what has been..."

106
00:16:58.322 --> 00:17:08.322
Because I'm asking everyone what's working, and he's like, "I know what worked, like, last year well. Do I know what's gonna work next year? Like, no. Do I have any idea the following year? Zero idea."

107
00:17:08.842 --> 00:17:19.342
And so I think there's a lot of questions about the open web, particularly on webpages. You know? The, the webpages are not okay. They're not getting a lot of love these days. [laughs] Is, is, is this a question?

108
00:17:19.402 --> 00:17:25.482
I can, I can tell you my, my, my thoughts. It, it's a statement and a reaction. It's a sta- it's a statement? Yeah. This is, this is a special- I'll, I'll-... implementation I do.

109
00:17:25.822 --> 00:17:33.432
[laughs] I, I wasn't sure if this was a- an opport- an in- an, an invitation for me to say something or more of a, of, more of a closure. It, it, it of course is an invitation for you to say something.

110
00:17:33.802 --> 00:17:39.332
First of all, we pay one and, we pay about $1.5 billion a year to publishers, so I, I'm very proud of that. Mm-hmm.

111
00:17:39.362 --> 00:17:48.462
I mean, I think we pay the best of the best of them and, and the alternative to our publishers is, you know, children spending time on TikTok and Facebooks, which is, which is disa- disaster.

112
00:17:48.982 --> 00:17:59.622
So I'm very proud of, you know, supporting trusted journalism. Always have been. Mm-hmm. But at that size now, it's, I don't know who pays more to, you know, very important, trusted publishers around the world.

113
00:17:59.662 --> 00:18:07.462
So it's, it's a, it's a moment of joy, you know, for us always when we finish the year and we look back and see- Mm-hmm... how much goodness, you know, we've done.

114
00:18:07.982 --> 00:18:16.322
So I, I think, you know, in terms of the web, I think the web w- here's what I think is gonna, like, we're gonna experience.

115
00:18:16.502 --> 00:18:30.194
Look, AI is a real thing and, and there's, it's only gonna con- the pace of change is massive.It's, it's, it's changing fast, and the, the size of the change are growing also. They're go- they're going up as well.

116
00:18:31.074 --> 00:18:46.514
I think we'll see the open web evolving in, in, in a way that smaller sites which enjoyed SEO and social traffic for a decade or two, those are at, at risk because I think search traffic will change dramatically.

117
00:18:46.574 --> 00:18:59.974
And, and it's gonna be very hard to get to be dependent on search in the future like we were in the past because Google will not be motivated as they used to be to drive that traffic just around the web.

118
00:19:00.334 --> 00:19:11.854
Gemini is an incredible business, and I think the advertising opportunity of Gemini for Google is th- way big, bigger and better than Google traditional blue link that we know.

119
00:19:12.114 --> 00:19:17.314
So they're gonna push Gemini as fast as and much as they can, and because of that, traffic dynamics will change.

120
00:19:17.674 --> 00:19:29.794
So I'm concerned about smaller sites which I like, blog sphere and, you know, those kind of like good voices. I think that's gonna be at risk, and they, they will probably reborn in, in a creator economy of- Yeah...

121
00:19:30.084 --> 00:19:38.784
sorts. They'll go to a social- So the mid and long tail seem like it's becoming like the cr- the creator economy is- The creator economy. Yeah. That, that's the future SEO, I think.

122
00:19:39.174 --> 00:19:52.434
So I look at Snap, Instagram, Facebook, TikTok. There's a huge opportunity for the mid, mid to small kind of creators that maybe had thousands, tens of thousands, maybe hundreds of thousands of, of page views a month.

123
00:19:52.474 --> 00:20:02.274
They have a real opportunity to, you know, to really get subscription and commerce revenue on as creators on the, on social. You know, Yahoo is, is, is having their creator network.

124
00:20:02.314 --> 00:20:11.294
So I, I think we're gonna see a lot of big platforms inviting creators i- in a way, in a new way to create and monetize and, and I think that's gonna be their future.

125
00:20:11.974 --> 00:20:30.834
When you look at the, the top of the market, specifically, you know, trusted publishers that have direct traffic, that have community around them and a brand around them, I think they have an opportunity, opportunity to grow much, much faster than they used to, and I'll explain.

126
00:20:30.914 --> 00:20:41.514
So first of all, I think the, the, again, the desire consumers will have to go to direct, direct-- By the way, I know this even myself. I go directly to ESPN.com a lot more than I used to.

127
00:20:41.554 --> 00:20:50.434
I go to the New York Post a lot more than I used to direct. I, I, I look for that direct relationship, and I'm never gonna replace... I, I love the Knicks. I watch the Knicks.

128
00:20:50.894 --> 00:21:04.154
There's no future in which I chat with ChatGPT about the Knicks. I, I may ask them when is the next game, but when it comes to what happened in the game, highlights of the game, I enjoy it. I don't want a short answer.

129
00:21:04.214 --> 00:21:16.094
Like, I want to get some, I wanna get some information. I, I watch with my... Like, me and my, my, my boy, we're watching ESPN highlights and things all the time. And there's, I-- it's not like I'm not a heavy AI user.

130
00:21:16.224 --> 00:21:17.774
It's just not the same experience for me.

131
00:21:18.374 --> 00:21:33.374
So I think there's, there's a lot of value in, in community and, and brand and trust, and the biggest value creation over the next few years, I think, will be when those trusted publishers adopt AI themselves in a much bigger way.

132
00:21:33.834 --> 00:21:44.074
Can you imagine going to publishers and suddenly everything is personalized, or you even have an LLM experience on the site? And, I know we, we can get into that because we have a- Yeah... product in market.

133
00:21:44.714 --> 00:21:51.374
But I think that's gonna be a big upside for engagement and revenue for those publishers.

134
00:21:51.434 --> 00:22:05.894
Okay, let's start with the Google point, 'cause I thought it was really interesting because, you know, having like, you know, covered Google from, from its early days and, and to becoming the hegemon that it became, you know, there was like a rough equilibrium.

135
00:22:05.914 --> 00:22:15.574
Like, it's never perfect. Nobody likes to be ruled over by a hegemon. There's not many popular hegemons out there, right? But, like, for this, you know, Google roughly, it was a difficult...

136
00:22:15.654 --> 00:22:25.234
I give it some, you know, I give it a break. A lot of people like to hate on Google these days, but, like, I give it a break in that it was trying to, to, to serve a bunch of different constituencies.

137
00:22:25.694 --> 00:22:31.394
And, you know, there was winners and losers all the time. Remember the Google dance and all- Yeah... those kind of things. Yeah.

138
00:22:31.454 --> 00:22:51.504
But its in- interests were roughly aligned with, and, and what I consider open web, it, what I mean by this is, like, web publishers, people who are publishing content to web pages, and Google was, if not the dominant, usually the dominant source of traffic to those web pages, and Google of- often monetized those web pages too.

139
00:22:51.614 --> 00:23:01.944
So, you know, it, it had an interest, right? Like, you know, it's not just a charity over there, as far as I know. That bargain is really breaking down. So wh- where do you see that developing?

140
00:23:02.014 --> 00:23:12.914
Do you see it breaking down further or, like, normalizing? Because Google wants to keep publishing alive, I think. Yeah. So, so again, the bigger publishers, their s- their

141
00:23:13.814 --> 00:23:25.514
search traffic is, call it thirty, forty- Yeah... percent. But that traffic is not the most valuable traffic. If you, if you talk to the publishers themselves and, and call them, I encourage you- Yeah...

142
00:23:25.534 --> 00:23:29.994
you'll see that the people coming from search tend to be the least money-making people.

143
00:23:30.814 --> 00:23:40.634
Like, if you compare a homepage person to a search person, a homepage person could be worth two to three X as search person because they read more. Yeah. They see different types, yeah.

144
00:23:40.694 --> 00:23:52.574
So, so even though it's thirty percent, thir- thirty, forty percent of traffic, it is not thirty, forty percent of revenue. Right. That's one important thing to state. Two, let's just say it goes down. How much down?

145
00:23:52.714 --> 00:24:01.754
Fifty percent down? They, they obviously need traffic to publishers because they need some dynamics with them for the content. They want to still train consumers to click on things.

146
00:24:02.174 --> 00:24:09.854
So there is value for Google and Gemini to keep clicking ecosystem healthy. But let's say it goes down fifty percent.

147
00:24:09.914 --> 00:24:20.734
So thirty percent of traffic becomes fifteen, and the publisher lost fifty percent of traffic, and let's say they lost, because of that, seven percent of revenue. One, it's not the end of the world.

148
00:24:20.774 --> 00:24:29.754
There, it's not the end of the world. Publishers have lost five, ten, fifteen percent of revenue for many other reasons over the last decade.Social, affiliate- Yeah... things came and went, right?

149
00:24:29.794 --> 00:24:40.254
So we've seen-- So first of all, not the end of the world, one, but it will be painful. 5, 10% of revenue loss is not insignificant, but that's what we're talking about. Two, what do you do

150
00:24:41.174 --> 00:24:47.754
with the 70, 80% of people that come to your site in other ways? I believe those people are worth

151
00:24:48.734 --> 00:25:00.594
a lot more if you apply AI and you serve them personalized content, and you give them LLM experiences, and you monetize that with a highly valuable performance ad.

152
00:25:01.134 --> 00:25:12.194
So what we don't understand is that we're in this junction now of old web, new web. Mm-hmm. In two, three years from now, I am convinced, convinced we're talking to publishers.

153
00:25:12.474 --> 00:25:23.094
We're going to our, we're going to USA Today, and we're having a five, 10-minute conversation about stuff we care about. What happened this morning with in the news? What happens in my town? What happened with my team?

154
00:25:23.414 --> 00:25:24.294
Can they win?

155
00:25:24.334 --> 00:25:42.594
And I'm, I'm talking back and forth, and I'm getting content related to my question, and this whole new supply, in the past, we called it page views, in the future, we call it queries, whatever we call it, but there's a whole new types of supply and LLM and time and engagement that will be born for trusted publishers.

156
00:25:43.314 --> 00:25:51.074
And, and in that supply, the monetization opportunity is something I can tell you from Deeper Dive like I've never seen before.

157
00:25:51.134 --> 00:25:58.774
So we now have a product in market which if you're listening to this podcast, and I hope you do, if you go to USA Today, you'll see Deeper Dive on the homepage.

158
00:25:59.514 --> 00:26:08.334
And on every other- And that's the AI, that's the AI search product that, that, that, that they wrote on with you guys. It's, it... Yeah, it's our LLM. Yeah. It is more like Gemini.

159
00:26:08.474 --> 00:26:14.494
Like if I, if I look at the world of LLM, you'll have Gemini, which is basically free LLM with ads.

160
00:26:14.534 --> 00:26:29.554
You'll have OpenAI, which is primarily SaaS, B2B to companies or subscription to consumers, but it's gonna be, I think, their vision would be more of a billion subscript- subscribers versus I don't think Sam Hel- Altman goes to sleep and dreams about CPM and CPC.

161
00:26:29.594 --> 00:26:36.674
I don't think that's, that's not... I don't think that's a top 10 dream of his. However, it is a top 10 t- you know, top 10 dream of Google.

162
00:26:36.714 --> 00:26:47.554
So Gemini will be really good at monetizing LLM and will educate marketers how to do it, and ChatGPT will never really be an advertising company. It will be a SaaS company. That's my prediction. I said it here.

163
00:26:47.614 --> 00:26:49.354
Oh, really? That's my opinion.

164
00:26:49.414 --> 00:27:03.314
And then Deeper Dive will be like Gemini for the open web, will be LLM free, the only free one with ads, and I think that revenue will be h- potentially way bigger than we've ever generated to the open web.

165
00:27:03.514 --> 00:27:16.754
So Google Search will become secondary to Gemini. Like the Google, this traditional Google Search experience that you, that you see today, like that you, that we o- often use. Is that the sort of prediction? I, so

166
00:27:17.814 --> 00:27:23.054
I'm-- I can tell you now with Deeper Dive, when you go to a publisher, the Deeper Dive is on the site. Yeah.

167
00:27:23.134 --> 00:27:35.834
If you start typing or you're clicking on a generated question based on what we saw you read before across the internet, which is one of our biggest advantages, 'cause we know what you really want around the web, you basically, what, what I know now, 'cause we only launched it in September.

168
00:27:36.544 --> 00:27:42.574
What I know now that I didn't know then is that you basically become a super, super user.

169
00:27:42.614 --> 00:27:59.954
The likelihood you will read something or the likelihood you will convert and take an action with an advertiser is, we haven't published it yet, but it's, I can tell you it is higher than I've never seen in my career to date, and I believe Google knows what I just said.

170
00:28:00.674 --> 00:28:10.674
So Google knows that Gemini performs probably a lot better than any other ad they've ever served on the web or on Search or on YouTube.

171
00:28:11.214 --> 00:28:21.094
So they probably can't wait for google.com as we know it to die as fast as possible so Gemini can take over. That's my guess. That's interesting.

172
00:28:21.414 --> 00:28:28.754
So yeah, if you go like on like do- de- deeper dive, and I think the big question is whether people are gonna use it.

173
00:28:28.874 --> 00:28:36.534
Are people gonna go to USA-- And I've asked, I've asked, I've asked the, all the Gannett- Well, I have the numbers... exact- I've asked all the Gannett. Okay. Well, Gannett, 'cause I've asked them this.

174
00:28:36.674 --> 00:28:42.814
I- I'm like, "Amazing. Congratulations for getting this. It's good to see, you know, innovation and product and stuff."

175
00:28:42.914 --> 00:29:00.534
And then I have a comma, and I say, "But is anyone really gonna use a search bar on a publisher website when the, an entire generation has been, you know, proving that nobody wants to use the, the site search on" It's like trained.

176
00:29:00.614 --> 00:29:07.313
Like, people who work for USA Today do not use the search engine on their site. Search is search, yeah, site search is, is dead for a long time.

177
00:29:07.334 --> 00:29:19.734
But no, no, I'm saying it's a consumer expectation issue is like, give me the data that says that, like, you're gonna be able to get people into that mode when they've been trained to...

178
00:29:19.774 --> 00:29:27.014
When they see a search bar on a publisher website, they've almost been trained, because they've been so horrific, to never use it.

179
00:29:27.114 --> 00:29:32.954
The, uh, I think that you and I are just, we live in the past because the question itself- I do. I'll speak for myself...

180
00:29:33.004 --> 00:29:39.094
the, the, the question is itself is archaic because consumers now that spend- Now we're getting into the good stuff. Thank you, Amir...

181
00:29:39.134 --> 00:29:48.384
because, because when 700 million people use everyday LLM, they don't, they're, they're, they weren't, they didn't live when site search even existed. They only- Okay...

182
00:29:48.394 --> 00:29:52.994
the only way they expect, the only experience they expect is conversation in free language.

183
00:29:53.514 --> 00:30:03.664
So when they go to USA Today or, you know, or India Today or, y- you know, m- many of our publishers around the world, we're, we're gonna announce some more soon, so you should, you should stay tuned.

184
00:30:03.794 --> 00:30:06.894
We launch almost every two weeks a publisher since launch since September.

185
00:30:07.634 --> 00:30:17.354
What happens is at first they actually don't even know what it is, and we look at things they've read around the web before, and we generate suggested questions for them. Yeah.

186
00:30:17.364 --> 00:30:22.634
So if I read about the Knicks on one site, and I go to a different site, that different site may ask me about the Knicks.

187
00:30:23.254 --> 00:30:30.074
That is something that even ChatGPT is not aware of because they don't know what people are reading about. They may license content, but they don't know that I love reading about the Knicks.

188
00:30:30.154 --> 00:30:39.150
Nobody gives them the information, but Taboola knows because Taboola is in many ways the internet.So now I, I was able to get you into Deeper Dive for the first time. The second time around, you're typing.

189
00:30:39.930 --> 00:30:54.360
You already know what it is. And I can tell you the usage, the users that are using it is, is in the double digit. Which, again, for someone that lived online for a very long time, one percent is considered high. Okay?

190
00:30:54.360 --> 00:31:03.990
Oh, wait, double digit percent. So over ten percent of users end up using it. Are us- are using, are using Deeper Dive. It is astonishing number. At scale, we're talking millions of people.

191
00:31:04.030 --> 00:31:14.230
So and, and, um, and those people that start talking... By the way, site search was never that number. So and because s- because it's not site search, it's conversation. It's LLM. Wait, do they ke- do they keep using it?

192
00:31:14.270 --> 00:31:23.000
'Cause there is a novelty factor. Like, I remember when mobile ads had like- It goes up...a thirteen percent, like, interaction rate. [chuckles] Yeah. But that, that... It goes up and up and up. But, but another point...

193
00:31:23.150 --> 00:31:34.970
So I, I'll give you another thing that you would appreciate. Using it is the, is... That's exciting. Great. What happens, and this is aggregated information across around the world of Deeper Dive publishers and users.

194
00:31:35.730 --> 00:31:43.310
Once they get into Deeper Dive, the, the likelihood for them to read something is again double digit. So it's performing.

195
00:31:43.410 --> 00:31:50.930
And then the, and then the likelihood they will click on an ad which will convert for the advertiser is very high.

196
00:31:51.590 --> 00:32:00.790
So you look at those numbers, again, it's- Mm-hmm...scaled, millions of people, and I believe Deeper Dive could become top five LLM companies in the world. Because I look at...

197
00:32:00.830 --> 00:32:08.190
I mean, OpenAI and Gemini will always be the biggest, but I look at the next in line, Perplexity and some other startups, I think we can be bigger than them.

198
00:32:08.250 --> 00:32:15.970
Because if we get ten, twenty percent of people, thirty percent of people to use LLM across the open web, remember, we reach six hundred million people every single day.

199
00:32:16.030 --> 00:32:25.950
So I'm so bullish on AI in the open web because I think that it, we can create a whole new experience for consumers on trusted publishers.

200
00:32:25.990 --> 00:32:43.959
And I will also say, when it comes to decisions that matter, and I don't know what you think about it, but if you're about to book a vacation for your family, and you're debating between different pla- you love Mexico, and you're debating between different places, there's a good chance you're going to ChatGPT and ask about some stuff.

201
00:32:44.790 --> 00:32:54.390
Like, you'll ask him questions about, you know, places to go. But there is no way you're making a decision just based on that. You're gonna watch videos on YouTube, reviews- Mm...to see if that's a good hotel.

202
00:32:54.470 --> 00:33:04.890
You're gonna maybe go to some review websites that, like, you know, they sent maybe some editorial person to that hotel to see what they thought. And you'll, you'll spend 30 minutes, an hour. Like, it's a big decision.

203
00:33:04.899 --> 00:33:15.670
You're about to spend thousands of dollars. This is where publishers, I think, win. It's, it's where trust matters. It's where decision that matters take place. Publishers are always part of it.

204
00:33:16.260 --> 00:33:24.430
And I think this is something that will always keep publishers, it's very important, no matter what AI model is coming up.

205
00:33:24.490 --> 00:33:38.360
And so is the idea here, and I would love any, any data that you have on this that hopefully does not violate anything with the SEC, is you then have a performance ad product because you, you, you have the answer.

206
00:33:38.420 --> 00:33:44.830
Anyone... I, I suggest you go and check out Deeper Dive. Just, you just go to usatoday.com. It's right, right on the homepage.

207
00:33:44.910 --> 00:33:54.730
And, and you have basically a, a Taboola performance ad, like, underneath the, the, like, two paragraph or so, like, response, right? Yeah.

208
00:33:54.770 --> 00:34:00.770
Well, it's very- And then you have, you have related lin- links to other, other articles in the USA Today network. Yeah. It's very light.

209
00:34:00.810 --> 00:34:09.990
I mean, we, it, it's, there's an ad at the top and, and there's an ad at the bottom, and that's it. And, and, and- But, but, but, but I think what I, I end up wondering is, like, look, everyone wants a search engine.

210
00:34:10.130 --> 00:34:18.190
I would love a search engine. If anyone wants to give me a search engine, it's a great business. I would love to have one. Look at Google. You can make every mistake possible.

211
00:34:18.200 --> 00:34:23.230
[chuckles] It doesn't matter if you have a search engine. That said, publishers would love their own search engine.

212
00:34:23.270 --> 00:34:31.350
They don't monetize the same because, you know, when I'm asking what happened to Savannah Guthrie, that from an intent basis is not very valuable.

213
00:34:31.490 --> 00:34:41.090
I'm not sure what, like, people can sell to me relating, related to this woman going missing. You guys take a different approach, and it seems like you're taking a different approach with this.

214
00:34:41.470 --> 00:34:49.200
How are you matching up the advertising, and how performant is that advertising? I- is it... Where is it between... Like, I mean, search is really performant.

215
00:34:49.930 --> 00:34:59.850
Again, we, we haven't published kind of like the advertising index. I can tell you here it is performing incredibly well for performance advertisers.

216
00:34:59.930 --> 00:35:11.890
We have demand for advertisers to be part of Deeper Dive in conversation with CMOs in ways I've just never seen before. It's performing really well, and I can tell you, remember the, the, the law of large numbers.

217
00:35:12.290 --> 00:35:17.729
You don't need, like, to really make a huge impact here. Yeah. We, you know, we need to just get...

218
00:35:17.770 --> 00:35:28.930
If we get significant CPMs effectively across these new types of traffic, that can create a revenue stream, a whole new revenue stream, a whole new journey. This can be a few decades of revenue for publishers.

219
00:35:29.450 --> 00:35:39.490
Search took 30... Search was a, is a 30-year business- Mm-hmm...that, you know, now is evolving into AI. And, and I don't know how long this will take.

220
00:35:39.510 --> 00:35:49.810
This could be another 30 years of revenue, and publishers can really play. As opposed to search, it makes so much sense for me to go to USA Today and speak about what happened today in the battery.

221
00:35:49.870 --> 00:36:00.890
Also, interestingly enough, I was at dinner with a company that kind of tracks CDN, you know, a-around the web. And, and they, and they spoke about what do they see people ask.

222
00:36:00.950 --> 00:36:11.550
They, when they look at the, kind of the crawlers from all different AI companies, what they said was that, you know, about 50% of crawls are last 24-hour questions. Yeah.

223
00:36:11.850 --> 00:36:23.550
People go to AI engines and ask about what happened yesterday, what happened today. This is a very important piece of information. And I'm, we've seen the same at Deeper Dive. 50%-plus of questions are recent stuff.

224
00:36:24.430 --> 00:36:25.710
Why, why am I saying it's important?

225
00:36:25.730 --> 00:36:42.582
Because it means publishers, who are the heartbeat of what happened today in the most trusted and curated and editorial way, are, are a very important leg on this AI tableOn their own site and even outside of it, in whatever deals will happen over time.

226
00:36:42.622 --> 00:36:52.922
So all the, all of it to say that consumers, when they use AI, they really wanna know what happened recently, sports, entertainment, news, local news, and that's something that publishers own- Yeah... will own.

227
00:36:53.082 --> 00:37:06.882
And the biggest gap is probably, if you're asking me, is culture. Because what happens is publishers really have to ask themselves, how do we... If we launch today versus maybe 100 years ago, how would we look like?

228
00:37:07.722 --> 00:37:19.282
And, and, and what type of data do we use? 'Cause, 'cause look, to have... I salute, you know, Mike Reed at USA Today for being first. Mm-hmm. You know, to say, "We're gonna, we're gonna do AI first and big. Big.

229
00:37:19.412 --> 00:37:29.242
We're gonna innovate big and fast." And you know, we-- he, he invited me to his all-hands, 10,000 employees. I was very excited. I know Mike for a long time, and I wa- you know, I came in and we talked about innovation.

230
00:37:29.302 --> 00:37:40.702
It was awesome. That was a cultural move. It wasn't a technology move. It, it... We used technology, but the culture enabled that shift to happen. And I think it will take...

231
00:37:40.762 --> 00:37:50.642
Some publishers will be early adopters and they'll say, "Yeah, AI all over the place. Let's empower the newsroom with AI. Let's give them data. Let's put it-- Let's personalize the homepage.

232
00:37:50.962 --> 00:37:57.292
Let's put deeper dive or different product-" Mm-hmm. "to, to..." But that, there's a lot that needs to happen here, you know, culturally.

233
00:37:57.292 --> 00:38:10.582
But when you see, when you see the ad- like, the, the leaps that are taking place in, in AI generally, right? And then you compare it with most publisher web experiences, okay? They're simply not keeping up.

234
00:38:10.782 --> 00:38:23.622
And like, I don't... I think consumer expectations will change, right? And when you go to a webpage, I landed on a webpage earlier, I don't know if they're, they're a tool. I don't know. But like, I, I won't say which.

235
00:38:23.702 --> 00:38:35.442
Actually, it was Variety. Like, I, I, I just like, I can't stand the Variety. Like, this experience is a loser in the market. It is going to lose every day of the week, and there is a absolute...

236
00:38:36.242 --> 00:38:49.382
The, the clock is ticking on this approach to taking a webpage and to having a floating player, and having another floating player, having a skin for Ford in Spanish. I'm not a Spanish speaker.

237
00:38:49.942 --> 00:39:03.262
In Spanish, having eight ad units for a 400-word story, having three different promotional units, having related articles, having then a content rec. It's a losing experience.

238
00:39:03.602 --> 00:39:08.942
I, I'm like, get me to an Instagram, get me to Apple News, get me anywhere but this freaking webpage.

239
00:39:09.082 --> 00:39:20.722
And I do not think I'm alone with that, and I don't think that the publisher product is keeping up with consumer expectations. I don't think that's a controversial statement, is it? It's, it's, it's not controversial.

240
00:39:20.782 --> 00:39:31.482
I think I always remind myself that, you know, publishers, and you, you know, you played a lot of time in that space, are doing God's work to- I know... keep people- I don't hate the player, I hate the game, et cetera.

241
00:39:31.512 --> 00:39:38.251
I know. I, I, I, I just, I just, I, I just, I just wanna remind you that- I know... we need those people to produce- I want them to live- I mean, I, I-...

242
00:39:38.251 --> 00:39:44.022
but I'm like, if you see a friend who's, like, drinking in the morning- I just don't-... you probably tell them, "Maybe you don't wanna do that."

243
00:39:44.422 --> 00:39:51.802
I, I just don't want my kids coming to me and making decisions based on a TikTok feed. So with all due respect, we do want those people writing content that matters. Right, so let's fix the product.

244
00:39:51.842 --> 00:40:02.542
And, and, and I think, I think we can create a much bigger future for the open web. I agree with that. And I believe the biggest upside is investing in direct-to-consumer relationships, so people come to you directly.

245
00:40:03.222 --> 00:40:16.922
You can interact with them on newsletter and notifications and apps and LLM and chats. And I think we can grow the revenue per person, ARPU, via AI in ways we have never seen before.

246
00:40:17.102 --> 00:40:29.582
I'm bullish on this deeper dive opportunity. Like you said, search business is the best business. This is better than search. Gemini is better than Google blooming. And, and deeper dive is Gemini.

247
00:40:30.302 --> 00:40:41.642
So for publishers to have a piece of that pie and to create double-digit user adoption- Yeah... rema- is, is, is big. Now, how fast can it happen? I don't know.

248
00:40:41.662 --> 00:40:48.882
It's very early stages for us and our partners, and I'm learning myself all the time. We're making mistakes. We're trying new things.

249
00:40:48.902 --> 00:41:00.042
But, but you know how you, we start this podcast by saying in 2012, nobody thought- Yeah... the bottom of article was a thing. Right. And nobody thought it could mean so much to performance advertisers.

250
00:41:00.422 --> 00:41:10.542
It's, it's way bigger than that eureka moment. Yeah. And so what do you think, like, in three years, like, w- a- are, are, are web pages gonna be the organizing principle for publishers?

251
00:41:10.642 --> 00:41:16.292
Like, that, that that's the main thing that they're shifting. 'Cause this is the part that I, I really honestly don't know about.

252
00:41:16.362 --> 00:41:27.952
Because I just think I see all the trends, and I know it's like something big is happening. I read the viral, like, you know, X article of the week that tells me I have- Yeah, yeah, yeah...

253
00:41:27.962 --> 00:41:35.282
18 months before, like, I'm, like, living in a cave- Yeah... or something. [laughs] Scrounging off- Yeah... the land in, in Prospect Park.

254
00:41:35.942 --> 00:41:46.532
But, like, it's hard for me to, to imagine that with all the changes going on, that we're gonna be going to web pages and hitting a back button. Or maybe we're hitting, maybe. [chuckles] Like, I just...

255
00:41:46.582 --> 00:41:59.182
It seems to me that the publisher product that they're ship- shipping has to change a lot faster. I agree. Okay. This is... That, that's why, that's why I think the biggest innovation ahead of us is culture.

256
00:41:59.752 --> 00:42:12.262
Because AI- Yeah... is a commodity. Anyone listening- Good point... to this podcast can today s- download Cursor, download Anthropic, download Llama open source and get going, have some fun.

257
00:42:12.842 --> 00:42:28.322
This is not the, the, the hard part is not the AI itself. It's the data. Do I know some- do you have something unique? And the distribution. And companies like us, and publishers, have both. They have both.

258
00:42:28.782 --> 00:42:36.922
So really it's about, can you build that experience to capture an opportunity to create monetization that is completely net new using AI?

259
00:42:36.982 --> 00:42:49.146
And, and that's why I think the biggest step ahead of us is-Can we change fast enough as, as an ecosystem? And those who do will, will benefit significantly from this. And I think we can play...

260
00:42:49.926 --> 00:43:03.336
This is not one of those... Look, it's to build an AI engine that can predict what I want to ask and can then monetize it with tens of thousands of performance search ads- Yeah...

261
00:43:03.346 --> 00:43:11.006
and then encourage me to continue conversing is hard. So you're doing text ads in this, right? Is it entirely- Yeah... text ads?

262
00:43:11.066 --> 00:43:15.886
You're not doing the classic, because, I, I mean, we have to get to that, otherwise I'll get like a million emails, right?

263
00:43:16.266 --> 00:43:29.596
Like, it, it's content rec ad net- ad networks are notorious in the popular i-i-i-imagination, let's just say, with quote unquote low quality. You hear words like clickbait, MFA, all the rest of it.

264
00:43:29.626 --> 00:43:37.586
We go through all these moral panics about... And there's some people who say, "I wish that we wouldn't have these things on our site. It's crack. I wish we weren't addicted to it," blah, blah, blah.

265
00:43:37.826 --> 00:43:44.325
You know all this, Ad. [laughs] You've heard this for a long time. I mean, I've... This is- [laughs] I, I, I heard it in, like, it was years ago you were talking about it. You're still...

266
00:43:44.666 --> 00:43:55.026
People still say this stuff to me, trust me. I'm not, like, if I did not bring this up, right? So, like, what is... And, and look, I say it all the time, publishers are, are in control.

267
00:43:55.086 --> 00:44:04.526
They can turn that knob any which way they want. They want the highest quality ads from, like, only, like, the best, that's fine.

268
00:44:04.546 --> 00:44:17.866
They're not, they're not gonna make as much money, and performance is performance, and every advertiser or every publisher, it seems to me, turns the knob towards revenue versus, you know, whatever sort of, you know, quote unquote quality.

269
00:44:18.006 --> 00:44:29.446
That's my sort of view on it. And I feel like, I mean, I, I, I can speak about, you know, the industry at large. I can speak about Taboola, and I'm very pr- we have 100 people full-time employee reviewing ads.

270
00:44:29.526 --> 00:44:41.176
We have an AI system we built that reviews ads. We have a public policy. I think we're more strict than our partners in terms of what we even accept or not. We don't have a knob issue. And look at who...

271
00:44:41.346 --> 00:44:54.916
And we work with the most incredible partners in, on Earth. I mean, Apple, Yahoo, Disney, NBC, CBS, Connecticut USA Today, McClatchy, Nexstar, Advanced Lidia, Le Figaro. I mean- Sure...

272
00:44:54.946 --> 00:45:06.086
unbelievable companies that have been around for, like, 100-plus years. What are we talking about? The, I mean, I, I just don't think anyone has, anyone has the patience for, you know, to, to take that level of risk.

273
00:45:06.186 --> 00:45:13.646
At least in my ecosystem, I'm aware of, you know, obviously, you know, I think over time things got better and we all improved and, you know. Yeah.

274
00:45:13.706 --> 00:45:25.166
But I mean, but, but as of now, you know, this is not, I think, something that I spend my time on. I spend my time on primarily on, you know, how do we 10X the open web?

275
00:45:25.666 --> 00:45:36.656
How do we grow the users that are on the site with high quality CPM, fa- you know, high CPM ads based, driven by performance, driven by, you know, CPC, CPA, all those good things. Right.

276
00:45:36.726 --> 00:45:49.046
That's, that's where I spend a lot of my time. A-and I also see, I look around, we're partnering with companies like Paramount and LG that is in, that are, they're in the TV business. Mm-hmm.

277
00:45:49.186 --> 00:46:04.086
So that we can connect our data and bring demand from TV in the form of performance ads on publisher sites to show the TV ads also drive performance. All of those events validate- Yeah...

278
00:46:04.116 --> 00:46:07.976
that, you know, that I think we're doing something in the right direction. So you guys have scaled.

279
00:46:08.006 --> 00:46:20.046
And, and I ask because, like, you know, it's one thing putting ads that could be interesting, like, on the, the bottom of, like, web pages, but Google, you talk about tens of thousands of aver- like, Google's got millions and millions of them.

280
00:46:20.055 --> 00:46:23.586
They- Yeah... they have, you know, a matching... Search is a lot of a matching issue.

281
00:46:23.726 --> 00:46:35.106
And, like, I remember, you know, Google wouldn't say how many advertisers they had before they went public, and they would always just tell me, like, over 100,000. It ended up being like 700,000 when they went public.

282
00:46:35.166 --> 00:46:46.736
[laughs] Like, and the, it's, they've got, they've got a lot, and so they're able to match really well. Is this a different sort of challenge, matching challenge, as you get into different surfaces- Yeah...

283
00:46:46.746 --> 00:46:48.246
that are not the web page? Yeah. I, I...

284
00:46:48.826 --> 00:47:01.145
So interestingly enough, I think the challenge and opportunities is not necessarily in that direction, 'cause this is, again, actually a search, this w- you're thinking blue link search, which is, by the way, definitely a, more advertisers is important.

285
00:47:01.446 --> 00:47:12.846
So yes, we want, as a company, for Taboola, I want to grow our net, we call it scaled advertisers, how many advertisers spend more than $100,000 a year. So we want that to grow as much as we can. So that's important.

286
00:47:12.886 --> 00:47:27.146
However, the biggest and interesting path I see ahead of us in the world of AI monetization, it's more about agentic AI in the sense that if I ask, let's say USA Today notices, Deeper Dive notices that I'm speaking a lot about the Knicks.

287
00:47:27.206 --> 00:47:35.686
Every, I come every day, I'm like, "Hey, what happened in the game, and why do you s- do you have a chance to win?" Compare this to, like, a year from ago. Like, I saw that in conversation, and I do it often.

288
00:47:36.446 --> 00:47:47.406
At some point, Deeper Dive should ask me, "Do you want me to do some... Do you want me to check games where I see prices are lower than usual and, and send this to you over email?" Mm-hmm.

289
00:47:47.656 --> 00:47:58.816
"And you ca- and do you want me to, do you want to give me $5,000 so I can buy you tickets for the rest of the year that are all in the $500 range versus two, uh, 1,000 because it's, it's a good...

290
00:47:58.846 --> 00:48:00.596
I'll look for those opportunities and I'll...

291
00:48:00.606 --> 00:48:08.786
Do you want to look for training to, to watch the Knicks when they play 76ers 'cause it's close and maybe it's more afford- And I'm starting, do you want a sub- do you want a subscription to the NBA app?

292
00:48:08.886 --> 00:48:16.606
I got a good deal for you. So at, do you wanna move this chat to your iMessage so we can talk more frequently on your phone versus on USA Today?" Mm-hmm.

293
00:48:16.926 --> 00:48:29.126
A lot of, there's gonna be a lot of asynchronous kind of interactions, m-multi days, and eventually I think the, the opportunity is that I ended up buying from Deeper Dive USA Today.

294
00:48:29.586 --> 00:48:38.566
Do you Deeper Dive, Deeper Dive says, "Do you want to buy this from Ticketmaster or..." And I support Apple Pay, and I give Apple Pay, and I pay it, and it's, and it, it's done.

295
00:48:39.366 --> 00:48:49.252
I think that's what Gemini is going to do.And I think they may-- this may be something that AI and search engines, you know, like we used to know, but in the form of AI, could grow into.

296
00:48:49.322 --> 00:48:59.382
Not only do we need more advertisers, it's gonna be longer journey and agentic such that it comes back to me, and it's feeding me information, and when I'm ready to buy, I'm buying it in the moment.

297
00:48:59.522 --> 00:49:08.702
I don't have to click and go to the advertiser page anymore. The funnel could be dead at, in the future. I'm just gonna make the decision. Okay.

298
00:49:09.062 --> 00:49:20.312
So that's a very different, like, ad system than the one that you have run before, right? I mean, that's, like, very different. That, that anyone has r- no one has done. I, I'm, I'm just, I, I suspect- Yeah...

299
00:49:20.312 --> 00:49:32.952
there's an opportunity. So I'll tell you why, I'll tell you why I'm saying that. We've never seen anything before that consumers spend so much time with. We've seen people spend 15 minutes a day on social.

300
00:49:33.062 --> 00:49:42.422
No one spent 15 minutes on google.com. Y- you search, you click, you leave. Yeah. But now with Gemini, and with DeepMind, you see people spending minutes,

301
00:49:43.322 --> 00:49:53.302
sometimes more than 10 minutes, conversing about something they care about. They ask, they read, they ask, they read. So if I give you so much time, would I trust you enough to also make a decision for me?

302
00:49:54.062 --> 00:50:02.942
So that's why I think this is a brand new opportunity for us. We have not done it yet, but I'm excited about, you know, where this could go from a monetization perspective. This is, again, I'm going back to trust.

303
00:50:03.632 --> 00:50:15.172
Do I trust someone like USA Today? Do I trust someone like, you know, NBC, CNBC? Oh, yeah. A lot more than I trust Perplexity, you know? All day long.

304
00:50:15.802 --> 00:50:25.642
I will give CNBC and NBC and USA Today my money all day long before I give Perplexity my money. So, so I think there's an opportunity for publishers to really capitalize on that trust- Mm-hmm...

305
00:50:25.702 --> 00:50:26.902
and offer me to make decisions.

306
00:50:27.362 --> 00:50:42.582
But this sounds, and this is the last thing, but, like, this, this sounds like this is something for, you know, that there's a g- a severe power law, that this is for, you know, just the top publishers, and I think that was getting into your, you know, w- bringing it back to, like, the open web.

307
00:50:42.622 --> 00:50:58.622
Like, I think of the open web, you know, it's been synonymous with, like, the long tail, and, like, there being these tens of thousands of, of websites, and that, that was both great and really challenging, I think, for the overall ecosystem because there was, they were up and down in the, quote unquote, "quality."

308
00:50:59.242 --> 00:51:10.442
And that kind of trust you're talking about, that's for people with real brands. Like- Or-... not people with plausible brands in a search results page. I, I think, I think you're right, though, that trust...

309
00:51:10.962 --> 00:51:15.952
But it could be a small local site. I don't think it has to be a huge volume site. Right. Yeah, yeah.

310
00:51:15.962 --> 00:51:25.482
This could be my lo- this, this could be a local site I love and I trust, and maybe, by the way, it's on TV, my TV- Board Panda was a very large site back in the day. Nothing against Board Panda.

311
00:51:25.622 --> 00:51:31.922
I don't know if they attend this. [laughs] Yeah, I'm, I'm not, I'm not, I'm, I'm, I'm not addressing any of that. Um- We will not address Board Panda.

312
00:51:32.222 --> 00:51:41.191
But, but, but, but we see today, you know, obviously we have some great businesses that are able to charge subscription, and there's, they may not be huge in traffic, right? Sub Stack has a lot of those- Yeah...

313
00:51:41.262 --> 00:51:50.402
very great voices. So I think it's, it's about, you're right, it's about trust and brand, not necessarily volume. Yeah. No, I think, I think, look, the open web, it, it has been amazing.

314
00:51:50.742 --> 00:51:56.442
There was a lot of arb publishing going on there. I would always, the, the Digiday pub- publishing summit was always strange.

315
00:51:56.502 --> 00:52:05.862
There would be people who are, like, from Vox Media, New York Times, and there was, there was arbitrage, but they were in a totally different, like, sector. They just happened to both be, quote unquote, "publishers."

316
00:52:05.942 --> 00:52:18.722
I think we're seeing some kind of similar thing. I mean, the, the sort of arb publisher of today is more likely a creator newsletter business, and, you know, they're doing a different form of arbing. Arb never dies.

317
00:52:19.502 --> 00:52:29.982
Look, I th- I think, I, I think, I th- I think what, what, what's, what's fun about this podcast is that anyone listening to this, basically, they, they get a first, you know, row view to our meetings over the last 15 years.

318
00:52:30.342 --> 00:52:38.531
It's basically me coming in, believing and being the optimist that there is always a path forward. No, I'm trying to be an optimist. I'm trying to be an optimist, Adam. Come join me. Come join me.

319
00:52:38.562 --> 00:52:55.682
I'm trying to be an optimist. I, I, I- Come on. I truly am. Anyway, I appreciate it. Thank you for taking time. It's always good to chat. It's great. Thank you. Thanks for having me. [outro music]