Why App Economy Disruption Won’t Happen As Fast As Everyone Thinks – Eric Seufert

David Barnard:

Welcome to the Sub Club Podcast, a show dedicated to the best practices for building and growing app businesses. We sit down with the entrepreneurs, investors, and builders behind the most successful apps in the world to learn from their successes and failures. Sub Club is brought to you by RevenueCat. Thousands of the world's best apps trust RevenueCat to power in app purchases, manage customers, and grow revenue across iOS, Android, and the web.

You can learn more at revenuecat.com. Let's get into the show. Hello, I'm your host, David Barnard. Today's conversation is shorter than usual and will be featured in RevenueCat's state of subscription apps report. Each episode in this series will explore one crucial topic and share actionable insights from top subscription app operators. With me today, Eric Seufert, media strategist, quantitative marketer, author, and investor. Eric currently shares his musings in the Mobile Dev Memo podcast, newsletter, and blog.

On the podcast, I talk with Eric about why app economy disruption won't happen as fast as everyone seems to think, how AI is just as useful for defending against copycats as creating them, and why the real barrier to app success is still distribution, not code. Hey, Eric, thanks so much for joining me on the podcast today.

Eric Seufert:

David, good to see you.

David Barnard:

Hardly a day has gone by in the past few months where I don't have someone asking me what's the future of the app economy, freaked out that AI is going to subsume all apps, that you're only going to have ChatGPT, that competition's crazy, that all manner of questions. So I've, gosh, six or seven of the top questions I get asked on a regular basis. And I wanted to chat through these with you.

I think of all the folks around the industry, you are kind of the foremost app economy expert, and I know you're also super into AI these days and following the industry very closely. So I thought it'd be great to have you on to get your perspective on all this. So let's just start with that first question. It's so much easier to build an app these days. People are vibe coding apps in an afternoon. What does a competition angle mean for the app economy?

Eric Seufert:

Well, first of all, thanks for having me. I always appreciate the opportunity to chat with you. Yeah. I mean, it's good timing because we had AppLovin's earnings two nights ago and the CEO talked about this. Right? So what happens in a world where the sort of barrier to entry for app publishing is far lower than it once was, right? Which, I mean, we're there now, but what happens as it just continuously descends? And I think what AppLovin CEO said, which is self-serving, is that the key sort of function of success is distribution. Right? And I've made the same point.

And I've been talking about this for a couple years on Mobile Dev Memo, but I made this point, I don't know, 2023, 2022. A lot of these generative tools are inflationary for the app economy and for advertising generally because as just you get more and more content being published, it gets harder to stand out. I mean, that's just a pretty fundamental premise. I mean, I don't think anybody would disagree with that. Right? And I wrote a post on LinkedIn about this a while back because I saw a bunch of people making predictions around the start of the new year about mini apps and how these would take off and it's so easy to start an app.

Everyone will just create the apps that they need and these mini apps will become the sort of dominant interaction paradigm. And they might. Look, they might, but you still have to get people to discover your stuff. And if there's a lot more stuff from which to stand out, you have to work harder to stand out. And that means advertising most likely.

And if you look back at the kind of history of these mechanisms that made publishing easier. Right? And in the piece that I talk about, I talk about the Dirigo type photography that made it easier to sort of get images in front of people or... But if you look at this sort of more recent history of just blogging. Right? The app store itself, when has user acquisition gotten easier over time since 2008? Has it ever? There's this long history of like, well, if it gets easier to publish stuff, it gets harder to get people to discover it.

David Barnard:

At a more fundamental level though, does AI change our relationship to app? So will somebody just say, "Hey, ChatGPT, I want to run a 5K" and ChatGPT is just going to coach them through running a 5K instead of downloading Runna or Strava or Couch to 5K or any of the hundreds of running apps out there?

Eric Seufert:

That's a really interesting question. I think it's certainly a very relevant question. I think it's TBD, but I think that there's a bigger question there, which is like, are you saying, "Hey, ChatGPT." Who are you saying hey to? Right? Because that's not necessarily settled science that ChatGPT just wins the category. Right? And so the question is do you just sort of... it turtles all the way down and up? Right? Where do you go for that information? Is it ChatGPT or does someone unseat ChatGPT as the go to verb for discovery? Right?

But my sense is, look, I mean, you remember all of Apple's promises two WDCs ago where they rolled out Apple intelligence, they talked about App Intents and none of that came to pass. Right? They talked about this overhauled Siri. None of that came to pass. Now, that's not because it's fundamentally impossible, that's just because Apple didn't do a good job on execution. But I don't know if you remember what the App Intent stuff, and they mapped out this whole series of interaction flows. And the reason I remember this is I remember writing about it, so I just looked it up before we started talking. Because I wrote a whole piece about, will this disrupt the app economy?

Because that's the same question people were asking. If you've got App Intents, do I need to interact with the app at all? If I need to do something in that app, I'll just ask Siri, it'll go in, you've got the App Intents, it'll hook into those and it'll do everything for me. And what I sort of hypothesized was when this comes to fruition, and it will at some point, it was like, yeah, maybe the category winners just run away with everything. Because I'm not going to say, "Hey, Siri, open up one of my ride-sharing apps and book me a ride to the movies." I'm going to say, "Open up Uber," because that's a dominant one.

And so that could become the new dynamic that we play in. But where I don't think we're going to end up is some sort of agent that orchestrates everything for us and there's no UI in any app because it's all just APIs or just MCP interfaces. I don't think we're going to end up there. And to your question about with Strava, my sense is the sort of chatbot interface that we see now is limited in a lot of really fundamental ways. And at some point people just prefer things that are more tactile and more interactive. Now maybe ChatGPT could replicate that, but that feels like a standalone interaction use case that you'd have to construct.

And I don't know that, certainly not in the current interface, but I mean, this stuff changes all the time, but could they just have an interface for everything? And learning how to run or whatever tracking you're running is one, and cooking is another and whatever. My sense is that probably ends up not being super meaningful for them. In the same way, why didn't Facebook do that? Because ultimately you're going to have your core business and that's going to dominate everything. And so building out all these use cases of specific functionality are not going to be valuable enough for you to even do.

David Barnard:

And I think people, the kind of AI takes over everything Bulls are underestimating the productization that takes place that all the people at Strava who've been building this product over years and have that knowledge and have new features they're continuing to add on in a weekly, monthly cadence, you don't get that. And so at some point ChatGPT or some LLMs, a better product manager than human product managers, and it can just like build the perfect interface. But I think we're a long way from that being true. But then what about like vibe coding? Are people just going to vibe code their own 5K app?

Eric Seufert:

No. I mean, that's going to be a hobbyist effort. I mean, the thing is, coding has never really been the core limitation of getting an app done. Right? And I mean, you'll see this. I mean, yeah, you could probably vibe code like a recipe app or something that works just as well as anything you get off the internet, but even then, probably not. Right? Because it's a specialization and it's managing the edge cases and it's having gotten user feedback. I mean, if you're just vibe coding something from scratch, it's not going to be very good.

I mean, you're going in there with just your intuition versus like an app that's successful in the app store is actually operating off of like this long history of feedback. Right? And that's, it's been improved over time. Right? And that's why there's a survivorship bias kind of thing there. Right? And so look, I mean, this was a question when a lot of no-code tools were released to allow people to publish a game in an afternoon. Right?

I mean, it's essentially no different from using, probably actually easier with the no-code tools to make a very simple, superficial, hyper-casual game than it is to do it with ChatGPT because then you're still dealing with XCode and you publish, you have to get a DUNS number as you know, you have to get an Apple developer ID and all that kind of stuff. And I mean, you still have to do that with a build box or something, but still, it's a visual interface for doing that. And that didn't really... I mean, none of those games were commercially successful.

And here's a good point that I see made a lot, right? It's not mine. Name one successful company where their product was built by Lovable? Name a single one? Now, maybe there's a bunch of people using their own recipe app that they built with Lovable. Can you name a single company? And Lovable apparently, according to them, has a lot of ARR. I don't know what words they're substituting in for that acronym, but they say it's ARR. But name me a company that is successful that is built off of Lovable? And I don't think I could name a single one.

David Barnard:

And we're still early, so it'll be interesting to see. Because I do think there is potential for somebody with good intuition and a good idea to go into Lovable and build something that becomes a great app that a lot of people use. So I don't think that's outside of the realm of possibility in the future, but to your point, so far we haven't seen any real breakouts.

What we've seen is people who are already building apps and already had that product intuition and do have some amount of code experience, they're the ones really actually succeeding. And then to your earlier point, it's also the ones who figure out distribution. It's not just about the code. And so yeah, you can pull something together in Lovable, but how are you going to actually get it in front of people? And that's still such a huge barrier to entry.

Eric Seufert:

I do push back and maybe someone would say like, "Okay, Boomer to this," but I just pushed back fundamentally on the idea of vibe coding, like an enterprise grade or a scalable product. I just don't think that's possible. And I don't think that's ever going to change actually. I mean, you just encounter so many dependencies with writing code and it has to be done really thoughtfully. I mean, there's a reason if you talk to a senior software engineer, if they publish 200 lines of code in a day, that's a big day. Right? The biggest part of the job is not writing the code.

It's actually thinking through the implications on the code base of what you're publishing. Is it scalable? Is it sustainable? Is it going to break something else? Is this going to have to be rewritten in three years? Is it extensible? And I don't vibe code in the sense that I don't go into cursor and tell it to build me an app. That would not be a scalable thing. That could build you with... And you hear this said a lot. I mean, this is again, this is not some novel insight, but you can build an MVP really quickly with these tools.

To build something that is stable, that is scalable, that is extensible, that is really difficult. And so I use... Now don't get me wrong, I mean, I use AI tools all the time for coding. But I want to see and touch every single line of code that goes into a big code base. And I maintain a couple. And what I do is I'll have it write a function for me. Right? I'll have it review code, right? But if you talk to anybody who actually writes software, ask them what the most fulfilling part of their job is.

It's deleting code. It's getting rid of code. It's getting rid of old junk and knowing that you can get rid of it without breaking something. And that, refactoring in that way, what you see a lot of people doing with vibe coding is they'll just start the whole code base over. That clearly doesn't scale for a large enough code base. A, it's going to be more expensive than just having a software engineer working on it, but you're just going to be burning tokens all the time. But B, then you're just going to have to do that again and again and again, and you're not going to have the consistency of a code base because some things are going to change each time you do that.

David Barnard:

If you're going to build a meaningful app, these tools as they exist today, and they will get better, but they're not going to support that kind of scalability. They're not going to support that kind of uptime. Yeah. And it just takes a lot more than is still assumed in the AI hype cycle as of today to actually build that kind of great stuff. So but things are moving quick, so we'll see how much better things get.

As things do move quicker and quicker though, the one thing I have heard a lot of people asking about is copycats, is that a skilled developer using AI can move quicker. And so ripping off an app, the entire interaction paradigms, I mean, some are even just ripping off style and assets from the apps and stuff like that. How do you think about copycats moving forward, that if you have a success within a few days, there's going to be tens of competitors for any app that goes remotely viable?

Eric Seufert:

A couple of things, right? So one is that's where true design intuition tends to save you and tends to differentiate you. And like you said, I mean, people are just copycatting these features or whatever. And so like, well, then people are just piggybacking off of your true, your sort of inherent design instinct or a skill that you've built up over a long time. It doesn't necessarily like an instinct. They can't build new features that way, right? I think that ends up differentiating quality experiences from the mass, right?

If someone's just copying what you've done pixel by pixel, there's not much you can do to defend against that. You see this a lot now with creatives, right? I mean, this has been an issue for years, right? If you have a winning creative, it's copied the next day. And so what do you do? You've got to build a process to always have a winning creative. That just shrinks your lead time. And that's just a fact of nature. And that didn't stop user acquisition. In fact, it just made it a lot more important to have a scalable process. Right?

An actual repeatable process for producing winners. And that, what that ended up doing was just accelerating the velocity of stuff that you need. But I think it also made it a lot more process driven, which is not possible to replicate. If you're a solo dev, you can't really replicate a process because you don't see it. You can replicate the ad, but you can't replicate the process, which means you'll always be eating the crumbs. And that's where, again, you see the primacy of distribution, not production.

David Barnard:

I have been really encouraged to see Apple taking a hard stance on copycats. They updated the rules of the App Store recently saying copycats will be banned. I've heard from several developers who had blatant rip off kind of apps where Apple very quickly took them down. So I do think that there will be some onus on the platforms.

And this is where sometimes the platforms can be helpful in being able to police some of this stuff. I mean, it's not going to be perfect and folks don't always just completely rip you off. They take aspects of the idea. But yeah, it's nice to see the platforms getting involved in that way as well. And that provides a little bit of defensibility, I think, in a healthy way, not preventing competition, but preventing just being ripped off.

Eric Seufert:

There's another piece to this too, which is the AI stuff, it's not one-sided. It's not just the copycats and the scammers and the fraudsters that benefit from AI. I mean, the other side of that is using AI to defend yourself against that kind of stuff.

So you can use AI to scan the app store every day, look for copycat apps. I mean, we've built stuff at [inaudible 00:15:40] to do this. We look at the app store every day. We're looking at stuff that's rising virally. We're scanning it, we're looking at the ads that they run. And so we have defense mechanisms for this. So it's not just one-sided that the nefarious use case of AI always wins.

David Barnard:

The last thing I wanted to touch on was wallet share and willingness to pay. As software gets easier and easier to produce, and it does seem like it's just going to continue to accelerate from here, is this deflationary on the entire software ecosystem? Apps are cheaper and cheaper to build, so developers can charge less and less for them. Do app prices go down? As AI apps like ChatGPT and Claude and Gemini and others take wallet share, are people going to pay less for the other apps? Where do you see things going economically for apps?

Eric Seufert:

There's two things here, right? So one is just, I mean, if you kind of take a much longer sort of term view, I mean, you could say, "Well, there's a lot of welfare effects from AI and it's going to drive the cost down a lot of ways and it'll give people more free time." And so there's just more need to use more apps, but you've got fixed income, but what does discretionary income mean in the world dominated by AI? And that's your kind of thinking very long term. If you kind of project out the next couple of years, what I think happens is people have fixed discretionary income, they've only got so many apps they can support.

Yeah, maybe you use price as a differentiator and you just race to the bottom, but what does that mean? It means ads. It means that's how you monetize. And so everything's an ad network on steroids. Right? So now you've got everything running ads. Well, okay, well, who's buying these ads? Well, all the companies that didn't exist before, but can via the sort of miracle of AI launch a product and they've got to get distribution, they've got to get eyeballs on their stuff. Right?

And so they're the ones buying all these brand new ad units. Right? And so you've got this ads economy that's just turbocharged as a result of this because you can't charge because if you charge, your competitors undercut you and so then you'll get no users. Right? And so therefore the only way to make any money is to show ads. And but won't CPMs collapse if that's the case? No, because there's a bunch of buyers now because there's way more software that's trying to get eyeballs.

David Barnard:

It's going to be fascinating to see how all this plays out in the next few years. But I will say, I think the current kind of AI doomerism that there will not be apps in a year, it is just not true. And I think it's very overhyped. As an app company today, you should be keeping a very close eye on this. And like you said, running app store queries on a daily basis to look at defensibility.

You should be leveraging AI to move faster yourself. You should be watching your back. There's all these things you should be doing, but freaking out and saying, "There won't be software here, let me go start a farm," I think is very premature. I think those of us deep into the AI side of things and on Twitter and that sort of thing are exposed to it in a different way than the rest of the world is and are a little ahead of our skis on some of these predictions.

Eric Seufert:

All these frameworks for building apps very quickly have existed for a really long time. Copeland's existed for a long time, right? Unity 3D for publishing games has existed for a long time. And I mean, you never fully took the bar down to zero and even where you made it very, very easy, you invite the hobbyist class in who was just not really interested in trying to build a business out of it. So you'll see disruptive change, but it's going to happen a lot slower than people think.

David Barnard:

That's a great way to sum it up. There will be disruptive change, but it's slower than the current estimates. So yeah. Well, Eric, it was so much fun chatting through this with you. Anything you wanted to share with the audience as we wrap up? I know you've been blogging and podcasting up a storm over at Mobile Dev Memo.

Eric Seufert:

Yeah. No, not really. Nothing specific, but you can always find me at Mobile Dev Memo on X.

David Barnard:

Awesome. Thanks so much, Eric.

Eric Seufert:

Thank you.

David Barnard:

Thanks so much for listening. If you have a minute, please leave a review in your favorite podcast player. You can also stop by chat.subclub.com to join our private community.