← Behind the Numbers: an EMARKETER Podcast
Behind the Numbers: an EMARKETER Podcast
What If the EU Bans Infinite Scroll, Facebook Becomes Cool, and OpenAI Makes $100 Billion in Ads? | Behind the Numbers
00:34:08
Loading summary
Transcript69 lines
- [00:00]
A
Foreign.
- [00:05]
B
Hey, gang, it's Friday, July 24th. Max Rahal, Nate and listeners, welcome to behind the Numbers and E. Marketing Podcast. I'm Marcus and join me for today's conversation. We have Calling Philly home, principal social media analyst Max Willins. Yep, there he is. Maryland staple. I don't. I think that's where he is. Director of reports editing. It's where he's from at least. Rahul Charter.
- [00:25]
C
Hey, Marcus, is that where you are now? Yes, that is correct.
- [00:28]
B
New York based principal AI analyst Nate Elliott.
- [00:33]
D
Good day. Fine sir.
- [00:34]
B
Hello.
- [00:35]
C
Hello.
- [00:35]
B
Joins us from our New York studio today. We start, of course, with the fact of the day. Okay, so where on earth was the hottest temperature ever recorded? So I went to Paris a few days ago and it was 37 degrees Celsius, 99 degrees Fahrenheit, and it was brutal. However, that's not even remotely close to the hottest temperature ever recorded. Can you guys guess the country city, if you're feeling really ambitious, with the hottest temperature ever recorded on planet Earth.
- [01:10]
D
My apartment in Queens last week.
- [01:13]
B
Yes, that's correct. It's not that, but it probably could have been.
- [01:17]
A
It felt like Bakersfield, California.
- [01:22]
B
It is.
- [01:23]
A
In.
- [01:23]
B
Oh actually, is Death Valley in California?
- [01:26]
D
It is.
- [01:26]
A
It is.
- [01:27]
B
Then it is California. It's. Death Valley is the hottest place. The US still holds the highest world's highest officially recognized air temperature at 57.6 Celsius. Call it 57 Celsius. That's 134 degrees Fahrenheit recorded in Death Valley in 1913, according to the World Meteorological Organization. I think I said that word wrong. That's too hot.
- [01:54]
D
That is. I have questions.
- [01:56]
B
Let's do it. What's your question?
- [01:57]
D
How much do we trust the thermometer from 1913
- [02:02]
B
if it said 57 Celsius? Probably not at all. I don't know how that sounds.
- [02:08]
D
It sounds like a broken thermometer to me.
- [02:10]
B
Humanly possible. Somehow 300 people called death Valley home year round. National Park Service staff, hospitality workers at the oasis Death Valley, and members of indigenous Timbisha Shoshone tribe primarily residing in a place some reason is called Furnace Creek. Jesus. Nine of the world's 20 hottest national records were set in the Middle east with all nine exceeding 50 Celsius. So it goes. USA's Death Valley. Then Tunisia, Kuwait, Israel, Iraq, et cetera, et cetera. I was wondering why it's called Death Valley. I assumed it's because it's so hot that people. It just cooks people. That's not true. According to national park service. Death Valley was named by pioneers lost there during the winter of 1849-1850. Only one person is believed to have died, but they did fear that the valley would be their grave. After being rescued by scouts William Lewis Manley and John Rogers, one man looked back whilst leaving and said, goodbye, Death Valley.
- [03:14]
D
And the name stuck because of.
- [03:17]
A
So maybe it was the cold. It's not because of how hot it was?
- [03:20]
B
Well, yeah, it could have been because of that. Yeah, I assumed because it's so hot there it was because of the heat. But yeah, possibly. Anyway, on that cheery note, today's real topic. We make some very specific but highly unlikely what IF predictions for 2026. Okay, here's how this episode works. Max will go first. He gets about 30 seconds or so to pitch a what if prediction he thinks will come true this year. We discuss Rahul goes next and then we end with Nate. Max, what is your what if?
- [03:56]
A
So my what if is pretty simple. What if Infinite scroll goes away?
- [04:01]
B
Oh, okay. Tell us why this came to mind.
- [04:05]
A
So this has been on my mind off and on for the last 12 months as these lawsuits that Metta and other social platforms are facing in the United States. But it was sort of pushed to the front of my consciousness a couple of weeks ago when the EU found that METTA is in violation of the EU's Digital Services act specifically for failing to consider the risks associated with design features like autoplaying video, like ceaseless push notifications, and like Infinite Scroll. This is a preliminary finding that they put out and there's a lot that's going to happen between now and when a final ruling is issued. But if the EU does decide that Infinite scroll has to go away, Metafase is a fine of up to 6% of its global turnover if it fails to comply with these rules. For those of you keeping score, that would have been about $12 billion last year and is likely to be quite a bit higher than that this year. And if that happens, potentially this opens up one of the biggest and most fascinating vacuums in really all of media. So that's why I brought it to the table this afternoon.
- [05:20]
B
Well, let's touch on that for a second. Explain the vacuum.
- [05:24]
A
So yeah, the vacuum is specifically related to a couple of different things. One is just the enormous amount of time that we spend on social media, right? So currently one of the, or the second largest sort of claimant of our time every day outside of time spent in front of the television, and we spend a ton of that time watching social video so social video has sort of transformed into one of the kind of main sources of media intake for us and really just people around the world. And if the EU were to ban infinite scrolling, it seems pretty clear that that would significantly curtail the amount of time people spend watching these kinds of videos. And what that would force, I think is a pretty hard pivot from lots of different social networks into things that would keep people on their platforms without running afoul of EU and I guess in this scenario, global regulations. So I think in some cases, like for TikTok, that would be a really, really hard pivot into shopping. They've already made TikTok shop kind of a core component of their experience, but they would I think, find ways to sort of make it even more central. I think for something like Instagram, it might mean trying to lean more into the kind of messaging features that they've been trying to sprinkle more onto what they onto their user experience. For Facebook, I think there's a lot of different directions that meta could try to take things, right? Like they could try to drive more and more activity around groups. They could also pivot into shopping by making Marketplace a more visible component of
- [07:11]
D
the app or farmville.
- [07:14]
A
Hey, Zynga investors, all six of them that are left definitely be very excited by the re emergence of a non Infinite scroll setting. So to me it's just a really interesting thing to think about. And then there's also the sort of most dismal possibility also, which is that meta basically just takes the mountains and mountains of user generated content that gets uploaded to them every day and uses AI to essentially stitch them all together into big blocks that are effectively imagine 12 swipes, or maybe more like 50 swipes compressed into a 22 minute block of ooze that you just stare at on your phone. And so you're not scrolling, but you're getting the same amount of media just delivered to you in a slightly different format. And if I'm being perfectly honest, I think that is probably what would happen. But that's one of the many possible scenarios that I could see unfolding here.
- [08:19]
D
Is it possible they just take the link, pay the money and keep raking in the cash that they generate from Infinite Scroll? I mean, someone over there is doing the math on this. How much cost of dinners more money do we make because of Infinite Scroll? And do we make enough that just paying the fines is worth it?
- [08:38]
A
So this was actually something that I did think about quite a bit because as you say, Nate, $12 billion is a lot of money for any company on earth. Except maybe Meta. Right. Like as you say, their margins are so healthy and so strong that it is certain that, you know, would this, were this to come to pass and EU to drop the hammer down on them, that they would just say fine, you know, if.
- [09:09]
D
No pun intended.
- [09:11]
A
Exactly. Because the reality is that they, the entire, the core of the business is just sort of that every swipe up inside of Instagram or Facebook, it brings more ads, it brings more opportunity to shop, it brings more opportunity to discover stuff, consume content. And it's very hard to imagine a sort of easy replacement to that user experience. So as you say, it's possible that they may just say we're just going to keep going and they can't all, not every market can find us $12 billion a year, so we'll just take our chances. So that's definitely a non trivial possibility.
- [10:04]
D
Do you have a sense or a guess as to how much revenue they would lose if Infinite scroll actually went away?
- [10:12]
A
You know, I have asked our, we're having sort of preliminary discussions on the forecasting team about building a forecast that examines what that would look like and what kind of a hit would push their bottom line. But we haven't crunched the numbers yet. But watch this space because again, like this is a preliminary finding and there's more to come. And then there's also, frankly, like I mentioned, one of the reasons that this has sort of been on my mind is that there's a trial in New Mexico that met a loss earlier this year. And one of the phases of the trials, the second phase of the verdict essentially is a judge is going to decide on what sorts of remedies Meta is going to have to undertake to make the suit right. And getting rid of Infinite scroll is one of the many things on a long list of possible remedies that the judge could land on. So it's possible that, you know, again, the EU is a, is a valued market for matter, but it's, you know, not nearly as valuable as the United States is. So any forecast that we build would have to sort of have, I think both. There'd be a US specific one, we'd probably build a global one. But it's a huge and interesting thing to know, try to tease out.
- [11:32]
C
Hey, one question that just kind of jumped out to me is that, you know, platforms have proven to be sort of masters at dark patterns and you know, while the intent is there, I think historically we've seen regulators tend to lag behind business. You know, Is the spirit is this, you know, they may want to follow these regular adhere to the letter of the law, but do they really have an incentive to follow the spirit of the law? And you know, the like follow up question of that is, are they just thinking of new ways to keep people kind of tapped into the platform that we haven't even considered yet?
- [12:04]
A
Oh yeah. I mean, that was partly why I closed my spiel with the possibility that Meta will come up with introduce something called Meta story time. And it will just be all the stuff that would normally get crammed into a normal 35 minutes of scrolling. It'll just be smushed into one AI formulated block that people can stare at and that will technically fulfill or get them out of the crosshairs of regulators. But it is effectively doing the same thing of just putting people in front of this lazy river of, you know, synapse frying short form content. It's just going to be packaged in a different way. So I think there's lots of ways that they could kind of, as you say, like fulfill the, you know, letter of the law, but maybe not the spirit of it.
- [13:08]
C
One other question I had for you too is, you know, Minda Smiley, your colleague on the social media desk, has written a lot about the blurring of lines now between streamers and social media platforms and YouTube. Like everybody was just trying to get eyeballs, you know. Do you think this change that might be more relevant to mobile usage could hasten the movement of these platforms into the living room onto the television sets?
- [13:33]
A
It's definitely possible. I mean, I think that a lot of, there's kind of a, one of the reasons, I mean it's, it's not hard to understand why, why TikTok and Instagram in particular want to get themselves onto television screens. But I think that is kind of a fundamental mismatch between the form of mobile vertical video and the sort of twitchiness that comes from being able to change the channel with just a swipe. And the stuff that you get when you sink into your couch to watch old fashioned television or long form content, or even just like medium length content, which is such a huge portion of what's available on YouTube. And so I think that it's possible that it could kind of drive more viewing on the biggest screen in the house. But I think that the people have gotten sort of sufficiently acclimated to the idea of content that they can just fire up at a moment's notice whenever they have any kind of downtime at all. And until people figure out a way to carry their televisions with them. I don't think that that's going to be much of an option.
- [14:44]
B
That's the problem with this episode is that each of these segments could be its own show and I should have maybe broken them out. Great start for, for the episode though, Max, is what happens. What if the EU bans Infinite Scroll? We have to move on, unfortunately. Rahul, what is yours?
- [15:01]
C
Mine is what if Facebook became cool again? Okay, I can hear the just outright skepticism in the sharp intakes of breath, but just hear me out. You know, I, I think Facebook seems acutely aware of the, you know, the reputational problem they have now is just being a way, you know, the shift from being a way to like actually engage with your real life social circles to being coming up repository of AI slop and, and boomer humor and, and just endless ads, you know. But you know, I think over the last 18 months they seem like they've tried to make some moves to, you know, recapture some of their original magic. I think one of the, one of the things they've really doubled down on is trying to entice creators to the platform. They've done that through stre payouts to creators. They established a program that guarantees payouts to, to creators who have, you know, sizable followings on other platforms. And they've also like, supposedly changed the algo to help small creators build new audiences. That's according to information. They've also kind of doubled down on, you know, trying to make the platform itself more appealing from a UX perspective, I think. And they've done that through things like the friends tab, which is supposed to be really, I think, what you could describe as the original version of Facebook, you know, which is really just people you have connections to and not leaning as heavily on the algo to dominate your feed. They kind of try to resurface the poke, if anybody even remembers that. And you know, they've said they've been trying to down, downrank unoriginal content, low effort reposts to try and like, you know, improve I think the quality of the content that people are engaging, engaging with on Facebook.
- [16:38]
D
I love that you led with their increasing payouts to, to creators who are popular elsewhere because nothing says cool like attempting to buy the affection of cool people.
- [16:49]
C
I mean, that is true, but you know, I think from the creator perspective. And Max, you know, correct me if I'm wrong because you're the expert here but like, you know, my sense is creators too are facing a lot of challenges and diversification of revenue seems appealing. I think, you know, the problem with, with all these programs are kind of pumping out to entice creators is like, is there, you know, are they going to be long lasting? Is it just like a test effort? They're going to, you know, cash their checks and then in three months they're going to be out? Like, you know, that's. I think that's a question that remains to be resolved in terms of all the enticements they're doling out the creators for sure.
- [17:20]
A
Yeah. I mean, there's now a long, long stretch of examples of Meta rolling something out to try to entice different pockets of the creative or media class and then abandoning them after they don't drive the results that they wanted. And I think that creators are sensible enough that, as you say, they'll take the check, but I don't know that it's going to necessarily get them to stick around. I was, though, when you brought this up, Rahul, trying to think through what it would mean for Facebook to become cool again. And it really did lead me down this weird path of trying to wonder whether that would get us closer to a super app. Because this is a thing that has come up as a topic of conversation in and around Social for the last 15 years because they're very prominent and kind of established in China. And Meta has made a couple runs at this. First they tried to create currency with Libra, which they thankfully abandoned. TikTok has definitely signaled that they're interested in doing this. Elon Musk is obviously obsessed with this idea. And I think that if you could somehow kind of magically get everybody convinced of the idea that Facebook is the place to be all the time, it has the most raw material needed to move in that direction. Right. It's a place you can do shopping. It's a place you can forge direct connections with people in your community. It's a place, it has really robust kind of messaging capability and infrastructure. And, you know, once they sort of continue moving down the path of facilitating more transactions and also this, this agentic surface that they keep trying to sort of shoehorn into their surface like, then we're talking about the first kind of ply or viable super app in the United States being Facebook, which is a crazy thing to contemplate, but that was sort of where my mind went when I first saw this as a topic of conversation. I don't know what you guys think about that.
- [19:43]
D
I think ChatGPT wants to have something to say about that, but I'm not sure either chatgpt or Facebook Meta will actually get there.
- [19:52]
A
No, I completely agree with you, Rahul.
- [19:54]
D
You mentioned bringing back the poke or trying to bring back the poke, which seems like a uniquely unlikely thing to happen. But I mean, can you see this happening? Can they come roaring back riding on a wave of mid aughts reminiscence?
- [20:12]
C
I think that the only the moonshot I can see bringing this full circle for them would be if Alphas start ironically embracing Facebook and the ironic use translated into real use. I mean, I think Max already kind of touched on some stuff. I think Facebook's really been relegated to this very utilitarian app where it's like Marketplace. You know, my wife is on the mom's group. You know, there was a buy nothing group that we use. It's not really a place that people go, I think, to consume content as much anymore. You know, Facebook also did redesign their core big blue app to, you know, make it look more like Instagram. They're, you know, you can watch reels on Facebook. They're trying to make it happen.
- [20:53]
B
I wonder though if you made the joke about Alphas, the youngest generation, but I wonder if there is a similar to how people have gone back to the dumb phone, so to speak, you know, that yearning for how something used to be and for a simpler way of engaging, whether it was with a phone, whether it is social media. So maybe there is something there. I would have thought this was an insane take until had we not done the episode. Max, me, you and Ethan last week on Time Spent and there was a piece in there, Oscar, who is one of the senior directors of our forecasting team, had said, quote, well, this is from one of the analysis that he wrote about how time spent with social media is changing. He said in our Q4 2025 update, we reversed Facebook's engagement declines to show growth for both 2025 and 2026. And in the Q2 2026 update, we are showing stronger growth in time spent on the platform. Time spent was up 4.3% year over year in 2025. We at 3.1% in 2026, up from our previous prediction of 1.6%. So you said Facebook's daily average time spent figures will settle in about 34 minutes a day this year. So they're not huge growth numbers, but they're growth numbers. And so it does seem like there has been a reversal in engagement, at least. I don't know about coolness. Harder to measure. Let's move to our final one. Nate, what is your what if?
- [22:22]
D
My what if is what if? OpenAI actually does make $100 billion in ad revenue in 2030.
- [22:30]
B
Okay, so this is coming from their claim from earlier this year. Correct. That they were going to make about two and a half billion in ad revenue from AI this year and 100 by 2030. 100 billion, correct.
- [22:43]
D
A hundred billion by 2030? Yeah. That's reportedly what they told investors when they went out and did their last private. Their most recent private round of financing.
- [22:53]
B
Okay, so tell us more.
- [22:54]
D
Well, to be clear, we don't think that's going to be anywhere near true. And if you look at our forecast, which hopefully is on the screen right now, you'll see that while we're very bullish on AI advertising overall, we think that U.S. marketers will spend $68 billion buying ads on AI interactions and responses by 2030. While we're bullish on AI advertising overall, we think most of what's happening in AI advertising will happen through more traditional search channels. Google's doing a great job of convincing people they don't need to go to ChatGPT or anywhere else to get AI responses. They can just Google something, and if they want an AI overview, they'll get one. If they want to dip into AI mode, they can do that. And so while we're bullish on AI advertising, we actually think that the total market for in chatbot advertising in 2030 will be about $5.4 billion. So we don't think they're going to come anywhere close to that. But to create that forecast, we had to sit down and think about what are the assumptions that OpenAI had to make in order to get to that number and to be able to answer people's questions about that number. And that's the interesting part of the what if for me, because to get to 100 billion, it means certain things would have already happened. For instance, AI would have already started to significantly cannibalize the amount of time and the number of queries people are spending with traditional search. It means that ChatGPT would have to reverse the current trend in which they are losing large amounts of AI market share every month and every quarter and go back to where they were a couple of years ago and be as dominant in consumer AI as Google is right now in search. And it also means the thing that would have to happen is we'd have to have already gotten a huge ad load inside of AI that basically every turn, every response of every commercially pointed AI interaction would have to contain at least one ad. So those are the things that would have already had to have happened to get there. And then the other side of the what if is what does that mean would happen going forward? Because it means not just a growth in ad revenue, it means a reordering of ad revenue. It means that maybe Max's prediction that Infinite scroll goes away. And so lots of companies have lots of ad budget they would have thrown into Meta, for instance, are looking for new places to spend that money. And there's a reordering of whether money's spent on Meta's platforms or whether it's spent with ChatGPT. If this requires AI cannibalizing search, it means that there's a reordering of the amount of money spent on Google advertising and that a lot of that's going to ChatGPT. So it's fascinating to me both in terms of the conditions that are necessary to get there, none of which we think will happen here at eMarketer, but also in terms of what it means would happen from that point forward in terms of the shape of the industry.
- [25:58]
A
Yeah, I feel like the first question that I had when we looked at this was is there any private consumer facing journalism left? Right. Because one of the things that this presumes I think is a kind of full fledged embrace of AI and ChatGPT in particular is kind of the front door to all kinds of information, right? Like, and in that scenario that basically means that all of the content and information that AI or a lot of the sort of news content that, you know, AI generates has to come from, from somewhere. But I, based on what I've started to see in the sort of emerging market of content marketplaces and stuff, we're looking at something akin to what happened to the music business when Spotify showed up. Where we're basically going from the news industry already went from analog dollars to digital dimes, and then this is then a move into AI micro pennies.
- [27:04]
B
And
- [27:06]
A
the largest news organizations, especially the ones that have institutional backing, like the BBC for example, are going to find a way to make those scraps work. But every local news publisher, every mid sized, mostly ad supported journalism endeavor is almost certainly going to find those economics unsustainable. And so that was one of the things that my mind flashed to immediately when we came up with this hypothetical.
- [27:39]
D
I mean, it's a huge concern and we're already seeing reports of just enormous drop offs in traffic to a lot of small and mid sized publishers, as you say, some of the larger ones I think will Find a way to survive. You mentioned BBC. They have a license fee to fall back on as well as ad based international opportunities. But without the licensing fee for most properties, they're going to have to find a way to make it work between advertising and subs. But the larger players should, most of them should be able to find a way to do that. Ahrefs, the search and AI visibility reporting platform, published some data recently. And if you do the math on the percentage of queries people are submitting to both ChatGPT and Google, and the amount of website traffic that's coming from ChatGPT and Google, it works out to something like a Google searches, I think 23 or 24 times more likely to send a visit to a website than a ChatGPT prompt or conversation. So we worry about Google 0, and that's a fair worry. I mean, even before AI, only about maybe half of Google searches led to a click out of the Google ecosystem. And AI overviews and AI mode has driven that number down further over the past couple of years. But the real concern is the standalone chatbots, because when you look at 23 or 24x in terms of Google being more likely, I mean, we're worried about Google traffic, but they're 23 or 24 times more likely to send a visit to a website after a query than ChatGPT is after a prompt, then Max, the concern you bring up is real and it's existential and it's a little bit frightening.
- [29:28]
C
Nate, I'm curious, what do you think 100 billion in revenue looks like from a user experience perspective? I think one of the codes, I'm not sure OpenAI's crack yet is like, how do you serve it? In a way, they're additive and not interruptive. How many are too many ads? And then, you know, what happens to user trust if they feel like they're, you know, getting non objective results from the chatbot?
- [29:50]
D
Yeah, those are all incredibly important questions. And really weirdly, they're questions that OpenAI doesn't seem to be trying to answer right now. Yeah, they said they were launching an ad trial. Their ad trial, as far as I can tell, consisted of one single creative format and one single ad placement. And within weeks they were bragging about the annual run rate and talking about expanding into other markets. It's not an ad trial, it's a full fledged ad launch. And they skipped over the part where they test and learn and improve the offering. And what people forget about traditional search is it took several years for the search engines to figure out which ad format was going to work for everyone involved, by which I mean the consumer and the advertisers and the search platforms. It took Facebook even longer. Facebook ads were terrible for the first five or seven years that they were sold. People forget that when Facebook IPO'd, they had been selling advertising on a large scale for, I think, four years, and their average CPM was less than 40 cents a thousand. The ads weren't working very well. It was, you know, the IPO was based on the hope that they might figure it out. And they did figure it out, but it took them five or six or seven years to figure it out. OpenAI hasn't really tried to start figuring it out, as far as we can tell. And we keep hearing and seeing things that say that maybe they're moving in that direction. I got a call from a reporter recently about a job posting that OpenAI had published, looking for someone to deal with multimedia advertising. And that would at least indicate that they're thinking about trying different creative formats. But if you're not testing different ad formats, different creative formats, different placements, then you can't possibly figure out what the best option's going to be, what's going to be most effective for advertisers, Most profitable for OpenAI, and most tolerable for users. And that's the trifecta that you need to figure out. And I really hope they start working on that sooner than later, because if they don't, then what you're going to get to. Your question, Rahul, is you're going to get one format in one placement, and you're going to get bajillions of them. And like I said, I think the $100 billion in 2030 assumption requires there to be ads in every turn of every conversation that could possibly be viewed as commercial in nature. And I don't think that's going to be a great experience. And I think there are other options if they test and learn and try to find those other options, but we don't see that happening right now.
- [32:28]
B
Yeah, yeah. 100 billion in revenue by 2030 assumes. Yeah, Ads Everywhere also assumes 3 billion weekly users. The projection they made, they said OpenAI was saying they're going to need about 3 billion weekly users by 2030. And just at the time, they had 900 million.
- [32:46]
D
Yeah, they have 900 million right now. A year ago, they were adding at least 100 million new weekly users every month. As we record this in mid to late July, they've been sitting at 900 million without having announced 1 billion for almost six months. It's been five months and change at this point. So that growth rate has slowed tremendously. And if they, even if they hit, you know, a billion today or in between when we record this and when we post it, that means that by 2030 they'd still be at maybe half of that 3 billion number if they continued at that same pace of adding new users.
- [33:27]
B
Yeah, gents, great takes. You have to leave there unfortunately. Thank you so much for hanging out with me today. Thank you. First to Max, always a pleasure.
- [33:34]
A
Marcus, thank you.
- [33:35]
B
And Sarah hall, thanks for having me. Yes, indeed. And tonight, thanks. Thank you, guys. Thank you to production crew Danny Luigi. I think I heard answer the background too. Thanks to everyone for listening into Pine, the Disney Marketer podcast. I'll be back on Monday talking to Ross and Danny Konstantinovich all about the three big questions surrounding Netflix. But before that, happiest of weekends,