App Store Organic Traffic is Dying, What's next?

The Edge Mobile Apps Have Over Web Apps Is Shrinking

I’ve been talking about this with Aivars (@Aivars_Meijers) quite recently on our podcast.

And I think a big reason is that organic traffic just isn’t there in the same way anymore.

A flood of new apps is coming to the App Store because a new gold rush around apps has started thanks to vibe coding.

You can feel it in App Store review times too. I’m saying this from my own experience and from what tons of other app developers on X are saying.

What I mean by all of this is: you need to actively bring traffic to your mobile app in order to make it with your mobile app.

And by “bring traffic,” I don’t literally mean only paid ads.

I mean you pay for acquisition either with money or with your time.

Working with creators, bringing traffic through Google Ads or Apple Search Ads, or spending your own time making organic content.

All of that is acquisition.

And when you have to bring the traffic yourself anyway, the split on the App Store compared to the web suddenly doesn’t feel that great.

I’m saying this as a person for whom at least 50% of the traffic to VidNotes is organic.

So I’m not saying organic App Store traffic is dead.

It isn’t.

I’m saying it is no longer a viable strategy to build the whole business around the assumption that the App Store itself will bring you enough distribution.

And I think that’s why we’re seeing multiple indie hackers arriving at a similar conclusion and moving toward B2B SaaS.

I think this is one of the reasons why the web is becoming something more interesting again right now.

But there’s another side to this whole distribution problem that I think is becoming really interesting.

Software Is Becoming Content

Software itself is becoming content.

I heard this idea a couple of months ago. Maybe it was even last year. I don’t remember where I heard it first.

I heard Adam Lyttle talk about it again in one of his recent YouTube videos, and when I originally heard this idea, I don’t think I really understood it.

But now I do.

Who has been on X recently and hasn’t seen a one-shot AI-generated game demo?

Whenever a new model comes out, one of the first things you see is someone one-shotting a game with it.

People share great website designs made with AI.

People mix Higgsfield MCP with agents to create videos directly from their workflows.

People build websites with crazy 3D experiences.

People build small tools just because the demo itself is interesting.

And these things get tons of views on platforms like X and LinkedIn.

Of course, great designs of mobile apps and websites have been a thing on X and other platforms forever.

Whenever someone built an app with really good design and really nice UX, people shared it, and it could get a lot of views.

That part isn’t new.

What’s different now is how cheap and fast it has become to actually create the software.

You can do the same thing now with AI agents, experiments, small apps, games, websites, automations, or completely weird software ideas.

The software itself becomes the content.

You don’t necessarily have to write a post explaining an idea and then add some generic graphic to it.

You can build the idea.

The demo is the post.

And maybe this is another small advantage moving toward the web.

If I share some weird AI tool, game, or agent on X, you can click it and use it immediately.

The software itself becomes both the content and the landing page.

I think that’s something to really look out for.

It’s not necessarily a new business model by itself.

But it can become distribution for your services, for your SaaS, or even for your mobile apps.

If you’re building AI agents, instead of only talking about what agents can do, build something interesting with them and show it.

If you’re building software, the software itself can be the content.

And because building software has become so cheap, I think this changes the way we should think about content quite a lot.

That’s why Adam’s take clicked with me this time.

But distribution is only one part of it.

The other part is what kind of businesses actually make sense to build in this environment.

There are two models I’m thinking about a lot right now.

Open Source + Hosted

The first is B2B products.

Pieter Levels talked about this recently in a post that went quite viral on X: indie hackers are building more and more things, but the traffic and money aren’t growing (@levelsio).

And I think that’s true.

I think building another simple micro-SaaS and selling it mostly to other indie hackers is getting harder.

That’s the lane that is closing.

But Nevo David (@wickedguro) and Postiz prove a different model that I think is actually a really fair one.

Make the product open source.

You can take it, host it yourself, configure everything yourself, and basically own the whole thing.

Or you can just pay for the hosted version.

In the case of a product like Postiz, I would choose the hosted version.

I don’t want to play around with hosting it myself. I know how much pain that can become, so I don’t want to do it.

But I can.

And I really like that.

Postiz has already taken that model well into seven figures of ARR, with Nevo reporting $145K MRR in July.

That’s getting pretty close to $2M ARR.

That’s something you can’t ignore.

I think I first properly started thinking about this model after Theo (t3.gg) talked about open-source businesses in one of his videos a few months ago.

And there are many other examples of this model that I just can’t remember right now.

B2B SaaS + Services

And then there is another model I’m becoming much more interested in.

Not simple B2B SaaS.

B2B SaaS + services.

This model is very interesting to me right now.

Not a simple B2B SaaS where you give people software and say bye-bye, but B2B SaaS plus one-to-one ongoing services.

For example, setting up AI agents for businesses.

I think that’s a very valid business opportunity right now, but again, that’s not really for indie hackers.

Other indie hackers can do most of these things themselves.

But setting up Hermes agents, or doing way more complex agent orchestration and instrumentation contextualised around how a specific company operates, and basically helping make the organisation AI-powered — that’s something that’s very interesting to me.

I’ve been very deep into AI automation recently.

I’m automating myself out of many operations around my own mobile apps, and there are a lot of operations when you actually want your app to grow.

• Ads.

• Content.

• Acquisition.

• Analytics.

• Experiments.

• Support.

• Organisation.

Building the app is only one part of actually operating it.

And I think there’s a business model here where you don’t just sell software.

You provide the software, but you also provide the service around actually making it work for the business.

You help a company automate parts of its daily work.

You set up its agents.

You configure them around the way the company actually operates.

You connect those agents to its systems.

You keep improving the setup, keep making sure it’s running the way it should be running, and get paid a monthly retainer for doing that.

You don’t just give them another SaaS login and say:

Good luck.

I think this is very viable, but this isn’t directed mostly toward other indie hackers.

It works much better with businesses that aren’t building their own software solutions.

Or the problems you solve need to be contextualised enough and complex enough that it simply isn’t worth delegating an internal team to figure all of it out.

And that’s probably the part I’m becoming the most interested in right now.

Not just building another piece of software.

Building software around the actual operations of a business, and helping make sure it works there.