An Interview with Eric Seufert about Apple, Facebook, and Mobile Advertising

Good morning,

In yesterday’s Article The Relentless Jeff Bezos, I called the Amazon founder “arguably the greatest CEO in tech history.” The only other contender, I think, is Steve Jobs. Being a truly great CEO is about more than capturing an opportunity and riding it, even if that opportunity is a trillion-dollar one; it also means thinking deeply about how to build a company that endures, and what was notable about both Jobs and Bezos was how much time they spent doing exactly that.

I’ve written about Apple and Amazon’s organizational designs on various occasions, including Apple’s Organizational Crossroads and The Amazon Tax. What is fascinating is that the two companies are polar opposites of each other: Apple is extremely centralized and focused, befitting its obsession with being the best, while Amazon is extremely decentralized and independent, befitting its obsession with experimentation. It’s why Apple is known for multiple groundbreaking products, while Amazon is known for multiple groundbreaking businesses. I think, though, the fact they are so drastically different speaks to why Bezos and Jobs rank so highly as CEOs: the only way you end up on the extreme end of the organizational structure axis is via clear intent and purpose from the leader.

Meanwhile, the vast majority of companies, even extremely successful ones, end up somewhere in the middle in terms of their organizational structure; they may build great products, particularly as long as their founders are leading the way, but in the very long run products are downstream from organizations. Meandering to the middle is a recipe for mediocrity and malaise in the very long run, and Bezos and Jobs, more than any tech leader that I have studied, worked tirelessly to attack this sort of decay at its root.

Also, there is this:

I guess I may be biased about Bezos! 🤣

On to the interview:

An Interview with Eric Seufert About Apple, Facebook, and Mobile Advertising

Eric Seufert is the creator of Mobile Dev Memo, where he writes about mobile advertising. Seufert speaks from experience: he worked for several mobile gaming startups when the app ecosystem was just taking off, and was later the Vice President of User Acquisition for Rovio, the developer of Angry Birds. Seufert then transitioned into building platforms for advertisers, and now works as a consultant in addition to his work at Mobile Dev Memo.

I have found Mobile Dev Memo to be an invaluable resource for understanding the mobile advertising space generally, and the ongoing dispute between Apple and Facebook specifically. Seufert both predicted this dispute, and has been covering its twists and turns on Mobile Dev Memo, as well as on Twitter, and I wanted to interview him for the Daily Update to really get an understanding of the issues at play. I thought this interview was absolutely fascinating, and I think you’ll agree. I am also making this Interview free-to-share if you wish.

As a reminder, the Daily Update Podcast is particularly useful for interviews, even if you are generally a reader; today’s interview is available now.

My questions are in bold; this interview has been lightly edited for clarity.

I’m here with Eric Seufert, who I consider basically the best analyst as far as mobile advertising on the internet right now. I’ve been a huge fan. How long have you been doing actual on the internet analysis as opposed to working on the industry?

Eric Seufert: Well, I’ve done them both simultaneously. I started the blog that became Mobile Dev Memo, I think in 2012, so going on nine years.

Very cool. And so what were you doing professionally at that time? What’s your background in this space?

ES: I started the blog almost as like a hedge because I joined a very early stage startup that felt very risky and I thought, okay, if this goes under, which it actually did, I have nothing to show for my time that I invested into this company so I should start blogging. That way there’s some artifacts of my knowledge that people can look at. I had joined Skype out of undergrad and then I moved, I was in Tallinn, Estonia working at Skype. After they got acquired by Microsoft, I moved to Finland and then started working in gaming and so I joined a mobile gaming startup that was actually doing location-based games, very similar to Pokemon Go, but back in like 2011 or 2012. So maybe a little bit too early…

There are no bad ideas, there’s just bad timing.

ES: Yeah, so I was working basically as the head of marketing at a mobile gaming studio.

Was that Rovio at the time or you were at Rovio after that?

ES: No, a couple of years later. So this was just a tiny startup that ended up not being successful. Although we had one game that was moderately commercially successful, but the company ended up shutting down. And then I joined a company called Wooga, a fairly large casual game developer based out of Berlin, recently exited to Playtika, which is a multi-billion dollar Israeli company that went public a couple of weeks ago. Then I went to Rovio as a VP of User Acquisition. Which in Rovio is a company makes Angry Birds, for those who don’t know.

How does user acquisition work for games? I think looking back at your career, it turned out you’re actually perfectly positioned to understand this entire ecosystem because games were very much at the forefront of figuring out these business models now which are used by other companies, particularly in DTC commerce, but it all started with games. Walk us through the way a mobile game business model works, particularly from the user acquisition perspective and the role that Facebook plays in that.

ES: That’s actually a really perfect way to set up the kind of current situation that we’re facing. The 2012 to 2015 era was a very specific moment for mobile games — you had the first cohort of very successful games, hitting their stride and garnering very large audiences — perfecting this business model that proliferated out from some of these companies. You see a company like Machine Zone (MZ), which made Game Of War, Mobile Strike, these were formative games in mobile gaming, and many of the heads of marketing for other gaming companies fanned out from there; King with Candy Crush, Supercell with Clash Of Clans, a lot of these sort of tent pole mobile games established this process, which is basically just buying users on Facebook. Estimating what’s known as an LTV, that’s broken out by multiple dimensions. Doing creative optimization and campaign optimization in a way that tries to get the costs below what this expected value is in a way that kind of triangulate like some segment of targeted users and then just running that at scale and spending as much money there as you can.

I think the way that I’ve always thought about mobile gaming, correct me if I’m wrong here, is it’s fascinating because of the fact that conversion happens on the same device where the ad happens. So you see a Facebook ad, you get converted, you download the game, then the game has the Facebook SDK in it so you can measure to what extent that user, both downloads the game, but then also how many purchases they made and can roll that back into your targeting. The old cliche about advertising is it works on 50% of people, the other 50% doesn’t work, we just don’t know which 50%. Whereas this flipped it on its head, where the level of knowledge of the success or failure of a campaign was way more knowable than ever before.

ES: Yeah, and that came to the market around like 2016, so Facebook rolled out these new campaign strategies that were based on what I call the events stream. Facebook’s got an SDK in basically every single app that exists and that SDK has total transparency into what users do in that app. So Facebook was receiving this stream of events, which comprise of fairly banal things like “completed level two” or very impactful representative things such as making a purchase and Facebook is ingesting this constant stream of events, and it’s attaching those events to the profiles of Facebook users. It basically is able to know what you spend money on in this backward looking time window, and then to show you ads for more things like that.

The way they have gotten so good at this is at this point that there’s not even really room in that dynamic for the advertiser to improve the system, this is totally online, “online” meaning in the machine learning sense it’s totally automated and just running autonomously. All you really have to do is feed it ad creatives and then Facebook takes those ad creatives and it just tests in this way that only it could given the access to data that it has and reach that it has on very specifically defined groups of people. So the more ad creative you give it, the more surface area it has to explore these very granularly defined groups and find the best ad creative for that group based on their proclivities to engage in apps like yours.

Is it fair to say that, because in this relationship Facebook’s ingesting all this data and they’re attaching to the profile, but you as a game developer, the royal you, you don’t really know who’s even seeing your ads, right? This data relationship is a one-way street, is that correct? Or to what extent can you know your customers, or does Facebook completely intermediate that?

ES: No, that’s totally opaque and you would prefer it that way because that would just add friction to the process. You want it to go and experiment. If it can come up with a meaningful combination of targeting dimensions to reach people that are relevant for you, you don’t want to slow that down by telling it “No, I think actually you should be targeting people that are of this age or that live in this city”. There’s no way for you to intuit what those dimensions are and the reason for that is because you are targeting, if I go into Facebook and I create a campaign and I want to target based on interests, that’s based on my assumptions about who would like this product. Facebook doesn’t need to use assumptions, it has historical data. It knows that this person made purchases in an app like yours or in a DTC web shop like yours, it doesn’t need to make any assumptions about anything, it has behavioral history. A lot of time I’ve been very shocked, just having worked for many gaming companies, at who the high monetizers are. There’s basically no pattern, the only surprise is that you’re surprised. It’s just groups of random people that probably have nothing in common if you put them in a room together.

This is very interesting because I think the point that I was thinking about and which you verified, is this idea that the data flow is a one-way street by-and-large, which I think actually assuages a lot of privacy concerns in that, yes, Facebook knows everything, but Facebook is a black box and actually from a privacy perspective, that’s kind of a good thing. It’s not like they’re spewing out that data back to you but the point you made on top of that, I think is so fascinating is it actually goes further than that in that there is this mistaken view of targeting. Apple doesn’t help with some of their advertisements on privacy and stuff like that of business saying, “Oh, we want to focus on, we want to get this Eric guy in Austin and we’re going to get a man in his thirties or forties or you look like you’re twenty.”

ES: Thank you, I’m thirty-six, thank you very much.

(laughs) It doesn’t even get even close to that, it’s just I want to acquire users and Facebook basically takes the ball and then runs with it.

ES: Right. And the thing is, that bank of data that Facebook has amassed, that’s not accessible to anybody. What is valuable is the filters that they apply to that and the machinery that it passes through and the outputs that come out of that, that are very valuable to Facebook because it’s machinery doesn’t look like anybody else’s machinery. I wrote a post about this like maybe a year or two ago that said, the title was Facebook doesn’t sell data. No one would want to buy it. The valuable data are the model coefficients, the targeting coefficients that wouldn’t be useful for anybody else, because nobody else has the different dimensions that they target that Facebook has. I’ve heard that Facebook has 100,000 different dimensions that it qualifies users against. Even if someone would hack into Facebook and pull out these model coefficients and hand them to Snap, Snap couldn’t do anything with them. Their advertising infrastructure doesn’t look like that, they don’t have those exact dimensions that they qualify users by. You couldn’t hand that to TikTok, they don’t have this exact same infrastructure, there’s nothing you could do with it.

Yeah. I wrote a piece back in 2018 where I called it a Data Factory and the reason was because I hate that term data refinery because there’s this point that the data in and of itself is completely worthless. You have to actually have the machinery and the infrastructure to transform that into something useful and I think it’s a really important point on two parts: It’s not just the fact that the value is in the infrastructure, but also it’s the fact that from the outside, all this data is just a whole bunch of crap. It’s all meaningless signals that is only useful and tangible in a way when it’s synthesized in the way that only Facebook can do it. To me, just to go back to the privacy point, that’s a comfort to me, that makes me feel better about it. Knowing the fact that all this stuff is worthless and no human — people have this vision of like, I always use this idea, I refer to the East German Stasi, they had those pictures of files and they’re going through your file. That’s not even feasible, and even if it was feasible, there’s literally nothing of interest there that just by the nature of what’s collected.

ES: Right, it’d be totally opaque and inscrutable, there’s nothing that you could do with that, it’s very much like purpose built for the, as you say, machinery.

It’s a bit of a sidetrack where we got to this point, but I just think it’s really important because I wrote about it this week about Apple and Facebook, that there are trade-offs here. What Facebook does is really valuable and unique and does enable small businesses or large businesses to compete in a way that was not possible previously. Apple’s broad point about user privacy, of course we want privacy, and so there’s a trade-off but to evaluate a trade-off, you have to actually know what the costs are. I get the frustration, I know we’re kind of on the same side here so this is maybe a less interesting debate than otherwise, but the costs of what Facebook is doing are so overstated in many respects just because people don’t really understand what they’re actually ingesting and doing with it.

ES: Yeah, we’re totally aligned there. What I’ve said about this is Apple claims to be just wanting to empower users with the ability to make a choice and now we’re kind of talking about the ad tracking pop-up. I want to give users a choice to decide, to empower them, to be able to decide if their data is collected for the purposes of ads personalization. Well, you can’t really make a credibly informed decision unless you know what that data being used for. And I posted this on Twitter, like 30 minutes ago, I’m sort of being pilloried of, but I posted a template of in an alternate universe, Apple used this as a pop-up text. And it says “Allow this app to personalize the advertising that you see, personalized advertising helps to keep your favorite apps free of charge.”

Obviously that’s very loaded, but so is the existing text. Apple is not giving people the knowledge about what this data is used to do and if people recognize that, wow, actually when people start opting out of tracking en masse a lot of free apps are going to really struggle to remain viable as businesses and hey, I like free apps, I like free content.

What is going to happen now going forward? So you’ve written a lot about the IDFA and you were writing about it before Apple did it. I think you said earlier that this was inevitable, it almost is surprising it took this long. And I’ve written about it, obviously, a fair bit on Stratechery that IDFA is a unique identifier on your phone. The reason why it matters is because every app has access to that, which allows Facebook to tie together your profile to the activity in sort of these different apps and now the user’s going to have to sort of opt into it, again, by going through this loaded text. I think there is a sense on Apple’s part, the impression I get, is this isn’t going to change anything, that it just going to be less targeted data, but I’m skeptical of that in part because what you said at the beginning, when you can measure ROI very closely, if you lose that capability or the ROI goes down, you’re just going to spend less. I mean, is there any way around that?

ES: Yeah, I think Apple is certainly trying to downplay the consequences of this because those consequences are going to be bad for consumers on net, but I’m also reticent to frame this issue from the angle of IDFA deprecation just because I think that understates the breadth and the impact of what Apple is administering to the mobile ecosystem. This isn’t just IDFA. If this was purely encapsulated with IDFA deprecation, it would mean mobile app campaigns are going to be less efficient. This we’ve learned in the last month and a half, that this applies to all of mobile advertising, Facebook has basically been clarifying Apple’s stance through its PR campaign. I was talking to someone today and I said, “It’s kind of like Plato’s allegory, prisoners learning to perceive by seeing the shadows on the wall of the cave.” We’re learning about what App Tracking Transparency (ATT) is through Facebook clarifying it via their PR campaign, because I think they’re getting a drip feed of more direct, pointed information from Apple in real time.

But basically what this is, ATT, is a totally new privacy policy that says basically the app developer should expose this opt-in prompt to the user and Apple has made it clear that there’s like very few circumstances under which you would not be forced to show this prompt. I think a lot of people misunderstood the wording to think that you didn’t have to show it, if you didn’t want to access the IDFA. Just equating this prompt with access to the IDFA meant, well, if I don’t show it, I just won’t access the IDFA and that’s it and I can track.

That is clearly what Facebook thought with their original iOS14 announcement.

ES: Exactly. And so Facebook said, originally, we’re not going to show the prompt because we’re just not going to access the IDFA, and they had to backtrack on that because Apple said, “No, no, no, you, if you collect any data that could be used for tracking, you must show the prompt.” Now, whether that’s the IDFA or fingerprinting relevant data with IP address and the phone model or even the email address, different sort of things that could be used as track identifiers, you must show the prompt, you must give people the option to opt out and if anybody has opted out, well, the practical consequence of that is you may not access the IDFA, but you also may not use any data to track them at all whatsoever in app campaigns or in app-to-web campaigns. So that means this applies sort to the totality of mobile advertising, and that’s why it’s so important. That’s why the IDFA deprecation thing understates it.

The thing that is always kind of weird to me about this is I think people underestimate and frankly I think Apple underestimates the degree to which Apple’s growth in services revenue and the growth in the App Store is because of Facebook and that Facebook and Apple have had this very symbiotic relationship where Facebook has done all of Apple’s dirty work, and Apple has harvested 30% on the backend just by virtue of owning the App Store. The question I have is, is it possible that Apple is shooting themselves in the foot here where their services revenue actually takes a meaningful hit because they’ve destroyed the engine driving it and they didn’t even realize it because they actually didn’t understand Facebook’s role in this?

ES: I don’t think so. I think Apple very much recognizes the role that it plays in the app ecosystem, which is the distribution engine for apps, and I think that’s what Apple doesn’t like. I think Apple sees that the App Store has basically become irrelevant as a point of content discovery. It’s basically this kind of frictional, annoying moment between clicking an ad and installing an app. Almost all discovery happens via ads or word of mouth and I think what Apple is doing in this, in recognizing the power that Facebook has in terms of influencing which apps become big, which apps are popular, how people are using their iPhones essentially.

In recognizing that, Apple is trying to regain control of that because if Apple cripples advertising, which it basically is doing, mobile advertising — this is all happening within the context of all this stuff that’s happening on the browser, which we don’t need to go down that rabbit hole, but Apple has been the instigator of all of that too. If Apple cripples mobile advertising, then the App Store becomes the primary discovery point for apps again, and Apple decides how people use our iPhones, Apple decides which apps are the most popular, and by the way, that’s a position that Apple used to occupy. 2012, 2013, Apple was king maker, if you got featured, your company valuation might increase by a hundred million dollars. It was really important to make that pilgrimage to Apple, go to Cupertino and beg for featuring, like “Please feature us, please give us the headline featuring because that would make such a big difference for our company”.

Then in that way, Apple got to influence what kind of apps got made and how you made them, so my sense here is that Apple wants to regain control. Now, I think there’s a broader three to five year arc that’s also happening, which is that maybe Apple recognizes that these one-to-one hardware dependent content platforms are becoming anachronistic, everything’s moving to the cloud. I don’t care about the App Store, I’ve got a device that connects to the internet, I can connect to any content platform, the App Store is just a middleman. Why do I need that? And if Apple maybe feels that way, then this would be the way to try to lengthen the useful life of this paradigm of hardware-based content platform.

What do you say to people that argue that the mobile approach to advertising was nobody was paying attention, there were no rules, they abused their position. This is finally an adult coming in the room and saying clean up your act, I think there’s a lot of folks that feel that way. Apple says this itself, “We’re just giving users a choice”. What’s your response to that argument?

ES: Well, it’s true. I wrote an article called The IDFA is the hydrocarbon of the mobile advertising ecosystem. It has been abused, it is creepy what certain ad tech companies do with this data, they collect a lot, and they don’t need to and it’s just unseemly. But Apple is defining privacy as what benefits Apple. If I asked one hundred people, “What does privacy mean to you?” You’re going to get one hundred different answers, and some are going to be totally stupid. No one really understands what privacy is. I think Apple is just saying, “Well, no, this is what privacy is, and by the way, hey, this benefits us.” We can move in a less extreme direction, I’d be happy to do that, that’s why I wrote that article about IDFA being this damaging by-product of abuse, which causes externalities but what they’re doing is they’re using privacy as a guise to change the rules in a way that benefits them, in a very big ecosystem. And also, by the way, hurts a lot of app developers that have made billions of dollars for Apple.

What would a better outcome be? What would be the way this ecosystem should work, where we get away from, to your point, the very real abuses that happen in mobile advertising, but also we don’t end up in a world where Apple controls everything? Is there a middle ground here and what might that look like?

ES: Apple in 2018 introduced this attribution framework called SKAdNetwork and then they rolled out the version 2.0 of that at WWDC when they announced App Tracking Transparency. It has a lot of limitations, but what it does is it basically allows you to attribute installs and attribute some kind of in-app context without revealing anything about the user, so it’s all aggregated at the campaign level. Now, if they just made that more robust, that would be a good solution because the way the attribution works now, it’s basically a hack that got invented by some companies that realize, hey, there’s no real way to attribute app campaigns on mobile and so we’ll create this intermediary step between clicking an ad, a bunch of data gets captured, you get forwarded to the app store, you install the app, that company has an SDK in the app and it takes the exact same snapshot and tries to reconcile those snapshots. That’s how mobile app-

Whereas for Apple it’s trivial for them to do it.

ES: Exactly, and they should have done that all along, but now what they’re doing is they’re just offering like this substandard replacement. They’re kind of doing the same thing on web too, the Private Click Measurement, they’re building the exact same framework on the web and for mobile as well as desktop to sort of aggregate all this data at the campaign level, which I think that’s a noble goal, but we need to be able to get more data than what Apple is giving you. With SKAdNetwork, they give you one post-back that can include up to one in-app events in it but that is moderated by this really convoluted timer logic that prevents you from really being able to capture enough signal there to understand the value that that user brought so you can optimize campaigns.

So you wrote earlier this week about the Apple Ads Attribution API, and that it’s actually more capable than the SKAdNetwork. Just to give context, SKAdNetwork is the API that’s available for third parties to use, to measure app conversions, whereas the Apple Ads Attribution API is to measure conversions from the actual advertisers that Apple has. They have them on the App Store, they have them on Apple News, basically on Apple properties. What’s the difference there? Do you think that Apple is going to maintain that difference? Is this an issue where they just got to one first or do you think they are actually privileging themselves and making their ads better by virtue of giving them more data?

ES: Well, right now they are just blatantly privileging themselves. They’ve blatantly privileged their own ad network by giving in their API that delivers reporting data for when you use their own ad network, you get actual creative level data with the post-back, so you get creative level data for that attributed conversion. You don’t get that at all with SKAdNetwork, that’s really important actually, because without that, you can’t optimize which ads perform better. That’s actually very, very important. Now I’ve heard rumblings that actually Apple is going to bring SKAdNetwork up to parity with this API, they just haven’t done it yet, but I think just the fact that they included it in their own API speaks to the value that they see in that. So why didn’t they include it in the first place?

This is just another example of Apple holding themselves to a different standard when it comes to “privacy”. I talked about the text of this pop-up, it’s really scary, they’re talking about tracking, they’re talking about across other apps and websites, the popup that get shown with the ATT prompt. Well, Apple has its own ads personalization setting in the Settings section of your iPhone in the setting screen, and it’s separate from the app tracking, the stalker behavior, they’re going to get your credit card info like that language is loaded. And then they have their own setting below, which is called Apple Ads and then you click through to that and then there’s “Hey, do you want your ads personalized or not”? Personalized seems a lot more friendly than this idea of an advertiser hiding in your bushes, watching your every movement, like tracking you through like the woods at night. That’s the sort of imagery that gets evoked with the wording they use in the current ATT pop-up, but they don’t hold themselves to that same wording.

How do you think this plays out going forward then? Facebook talked about next year, there’s going to be a difficult comp both because of the way advertising stormed back on the second half the year and also because of iOS 14. I thought they’d be a little stronger in saying there might be some difficulties here, how do you think this is going to play out both for Facebook but also for the advertisers that rely on the system as it is today?

ES: Well, I’ll start with the second group, I think this is existential for certain groups of advertisers. Any type of product, and here I’m talking about both app developers and websites that advertise on mobile. So any company that’s advertising to a niche audience and expecting that niche audience to monetize at a very high level individually is going to face real challenges here because Facebook simply won’t be able to do the kind of granular precision targeting that these companies have grown accustomed to. So I think, you’re going to see a lot of these companies just being unable to acquire users profitably, the unit economics won’t work out anymore.

I just wanted double down on that because to me, this is why I get frustrated about this specific debate, because to me the promise of the internet, I think the whole point of Stratechery or Mobile Dev Memo is you can run a business, you can reach just a specific audience and that’s one of the promises of the internet because the whole world’s your addressable market. Advertising is an essential part of any business and you need the advertising that matches that business model and that is Facebook. Apple referring to TV advertising, “We did advertising for decades and we didn’t have this tracking”. Well, you’re doing advertising for Procter & Gamble, do we want a world of nothing but Procter & Gamble’s? This is the fundamental issue at play.

ES: Yeah and I think that’s a very apt reference point there because I think what we’re going to see on mobile is just TV-esque ads, un-targeted, irrelevant ads. The reason people were happy to move into digital advertising is it’s targetable and measurable in a direct way and TV, you can’t really do any of that and that’s why people hate TV ads and they complain about them and see them as intrusive. So I think there’s different classes of advertisers that are going to face real challenges here and be heavily impaired by this. I think a bunch of different gaming verticals are especially susceptible to this and are vulnerable. I think a lot of DTC companies are very vulnerable to this.

The entire Shopify ecosystem.

ES: Well, actually it’s funny because no one really brings up Shopify when they approach this topic but my sense is, a lot of their customers are very heavily dependent on precision targeting from Facebook.

Absolutely. And you see this too, Facebook’s response is going to be emphasizing Facebook Shops because if the conversion event happens on Facebook’s platform, then of course they can track it perfectly so Shopify is just getting squeezed out of this, which is a real shame because that’s the exact sort of company and idea that to my mind, is the best part of the internet. This idea of unblocking all these individual entrepreneurs and it’s weird because Apple is actually helping Facebook when it comes to DTC.

ES: Absolutely. And Facebook, that’s kind of like the medium term play here. Facebook is just going to start subsuming all of this content into the Blue App, all this content interaction and commerce. They’ve already kind of done it, I think that’s one reason why Apple was all too happy to sort of instigate this privacy policy change that sort of disproportionately impacts Facebook because Facebook’s really been drifting out of its lane for like three years. They’ve got instant articles, that’s content interaction, they’ve got videos and they’ve got the whole new gaming initiative, that’s content interaction happening in the app. They’ve got Facebook Shopping, that’s commerce happening in the app. They’ve got Instant Games, that’s gaming happening in the app. They’ve subsumed all of this content interaction within the app and I think they’ve kind gotten away almost on a technicality because it’s not an app store. Apple famously has this rule, no app stores in the app store. Well, I don’t see an app store in here, do you? This isn’t an app. It’s kind of Netflixesque, there’s a content catalog and you explore it in this app and a lot of this content is published by third parties, but not an app store.

Fascinating. So you think that might’ve like triggered this response to a degree from Apple?

ES: Yeah, I think so. I think that, and just the fact that Facebook was a de facto distribution mechanism for mobile.

Interesting. In the long run, is Facebook just going to become even bigger? Is that kind of the outcome here, or do they face an existential threat as well?

ES: I don’t think they face an existential threat, but they’re going to face challenges. I think they expect ATT to be rolled out in March. Two, I think a lot of advertisers are going to pull back, they’re going to pull back because a) performance is going to degrade, but b) there’s just uncertainty. So everyone’s advertising models need to be recalibrated and they’re just going to pull back and spend as they do that, I think it could be pretty dramatic. But, at the same time, Facebook’s growing still at a pretty incredible clip, so my sense is that the growth offsets and they said they expected that to sort of straight line out through 2021. So my sense is that the growth offsets the revenue pain and they still are net growing, but I think it will be a challenge.

Do you see in the long run Apple really benefiting from this or what is App Store revenue look like in five years? Does Apple maybe have less App Store revenue, but more control over it? Is there way more subscriptions there would be otherwise but fewer apps downloaded, what do you think is the impact on the ecosystem?

ES: I think what Apple can do, my sense is that a lot of the motivating factors here have to do with very kind vague or nebulous notions of what Apple should be doing to protect the optics of its brand, and also its ability to control how people use the phones. I’m not really clear on what the path to more revenue is, except for when it comes to Apple’s ad network. This is clearly going to benefit Apple’s ad network, but not in a way that’s going to be transformative for its business, it’s going to be kind of a rounding error, it’s going to be non-material.

My sense is if there is a specific direct path to revenue growth here, it’s by them sort of bundling up a lot of these apps that existed as standalone experiences into iOS and making those subscription. They kind of did that with Apple Arcade, they’ve got a couple of health initiatives coming, they’re trying to connect all this to Apple Watch. So my sense is if they can control how you use your iPhone and they create a sort of multi-interface relevance there, then maybe there’s a path to growing revenue, but it didn’t seem like there is that sensibility baked into this privacy policy change.

Well, the great thing from Google’s perspective is they get to copy everything Apple does, which completely benefits them because Apple wants to kill cookies, Google’s like, “That’s fine. We have the browser, we have the App Store and we’re justified in doing blatantly anti-competitive things, because we’re just copying Apple.” The Apple-Google duopoly is really ruling everything.

ES: Yeah. Google’s silence on this issue has been really intriguing because I think they’ve got to feel like a little bit stuck. It’s like, “Well, okay, we’re clearly going to slow-copy Apple on any privacy initiative it does because we don’t want to allow Apple to create a competitive advantage around being seen as the privacy-centric mobile platform”. But at the same time, they run an ad network so they can’t come out and demonize targeted advertising because, well, that’s how they make some substantial portion of their money.

They’re being sued for antitrust, and this is so blatantly anti-competitive.

ES: Right, exactly. They’ve got a fine line to walk and I think they’ve hewed to a very shrewd approach, which is just like you said, kind of copy what Apple’s doing, but do it in a way that is not full manifestation of that. So if you look at what they’re doing on the browser with all the privacy initiatives there, it’s much more watered down than ITP, which is Apple’s privacy thing. I think they’re going to do the exact same thing on mobile so they’ll still be able to kind of like appease both groups. It’s like, “No, Android also protects your privacy”, but at the same time, it gives a lot of data to the ad networks, so don’t worry everyone, you’re all getting fed.

And then they’ll pass on billions of dollars to Apple in the background. That’s the thing that really rubs you the wrong way, it’s really rich to hear Apple go on and on about this — I guess all advertising is bad except for the money that goes directly into their pocket.

Anyhow, this was fantastic, Eric. This is really, really good. Everyone should check out and follow Mobile Dev Memo if you want to get really into the weeds of this, it’s my number one resource for following this sort of app stuff, it’s great. Thanks for coming on Eric, we definitely should do this again.

ES: Cheers, thank you so much for having me.

This Daily Update Interview is also available as a podcast. To receive it in your podcast player, visit Stratechery.

The Daily Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly.

Thanks for being a supporter, and have a great day!