Web dev at the end of the world, from Hveragerði, Iceland

Addendum on loose coupling and the iOS App Store

In yesterday’s doodle, I left out a note on why I positioned the iOS App Store in the upper right hand corner.

The greatest contrast between services like Facebook and the iOS App Store isn’t that app providers have access to greater capabilities than your average social media content provider but that the App Store is overall a liability for Apple, although I’m not sure that Apple recognises it as such. Even if it did, it’s stuck with the App Store as getting rid of it would be next to impossible without damaging the platform.

A tightly integrated App Store that is the sole method for distributing an app to a platform offers two benefits for the platform owner.

  • It lets the platform extract rent from app providers. In practice this is less than effective. Big providers like Amazon and Netflix have demonstrated that they can easily work around Apple’s rent-seeking if they want to. Many mid-sized providers who don’t have the standing to negotiate better deals from their own providers—like Kobo—have been particularly hard hit by Apple’s rent-seeking.

  • It lets Apple enforce the strong enough base level of security and safety that has let iOS become a mass market platform without a major malware problem. (Android, whose App Store can more easily be bypassed and is less policed to begin with, has not accomplished this.) Apple can’t change the fundamentals of how the App Store works without putting the mass market viability of iOS at risk.

The third advantage is a percieved advantage only. It is largely a faith-based ideological belief with no economic foundation:

  • An App Store gateway lets Apple turn iOS into an ideological, aesthetic, and cultural ‘gated community’. It lets Apple enforce U.S. values throughout the app store. This is fine if your only market is the U.S. but disastrous if you happen to be selling to the U.S., Norway, Japan, and India. The problem with this decision isn’t in deciding which value set should dominate but that in doing this Apple has inserted itself in the middle of a problem that cannot be solved. The world will both be diverse and the internet will give equal access to all of that diversity. This is a profound—revolutionary—change in human culture and by deciding to use the App Store for cultural propaganda, Apple has inserted itself on the losing side of this battle. Once it has put itself in the role of planetary cultural police, it no longer has the option to stop without incurring brand-destroying moral panic.

Apple’s first big issue with the App Store is that the benefits come attached to major liabilities that largely undermine the value they are getting from it.

There’s more:

  • Because apps are expensive to make (i.e. highly differentiated in general) and the tight coupling limits an app developer’s ROI (and makes their business more fragile) those app developers have a disincentive from innovating or taking risks. Hence all of the copycat apps that are filling the App Store. To reuse the terms I used in the post yesterday: like publishers, mobile app developers are underusing the platform’s scope for differentiation. This drives the overall value of the platform down and undermine’s Apple’s positioning of iOS as a general purpose computing platform.

  • Because of the innovation disincentive, Apple can’t operate on the principle that if they provide the hardware and the OS capabilities, developers will build tools that turn those capabilities into value. The odds of a new iOS ‘killer app’ have dropped dramatically. If Apple wants the iPad to be taken seriously as a general purpose computing device they have to build the apps themselves. This limits their ability to manoevre and puts excessive pressure on their internal development teams.

  • As I outlined above, Apple can’t open up the App Store without incurring some sort of major penalty. A whitelist of allowed apps is the only reliable security model for the large scale distribution of locally installed, packaged executables. IOW: if you have to have local apps, and if you want everybody to be able to use them safely, you pretty much have to have an App Store gateway.

  • Even if it did find a way to open the platform up, there’s a good chance that Apple has no appetite for becoming dependent, strategically, on big app developers, even if that would be best for the platform.

The App Store is a major liability and Apple is kind of screwed on this. Unless they change something, they will be forced to implement more and more productivity and business features themselves to stay competitive.

One possibility would be to solve this problem the same way that Google has with Android.

No, I don’t mean that they should open their app store up be be more like Android’s. That app store has many of the same problems as iOS’s. What Google has done is iterate on and generally improved the web as an app platform with Mozilla, Microsoft, and Opera all following suit.

The web has a security model of its own that bypasses many of the limitations of locally installed executables (largely by policing the capabilities of the apps instead of policing the sourcing of the app as app stores do). There’s already a precedent for Apple not to police the contents of the browser so it gets them out of the cultural police game. It’s an incredibly loosely coupled platform which lets developers choose the business model and strategy that they want and gives them the flexibility to adapt when circumstances change. It has marginally fewer capabilities but the gap has narrowed dramatically over the past two years.

To put it plainly: Google, Mozilla, and Microsoft are iterating the web faster together than they are iterating the API capabilities of their app platforms individually.

(Mozilla having given up on their mobile OS aspirations entirely.)

The web was and is behind native apps. But the API capabilities of app platform have overshot the requirements of most apps substantially. Progressive web apps are already good enough for many apps, are cheaper to make cross-platform, and degrade gracefully in legacy platforms. They combine reach and flexibility in ways that native apps can’t match.

The downside to supporting Progressive Web Apps is that they are aggressively cross platform. A killer progressive web app adds value to all supporting platforms, which undermines the competitive value of supporting PWAs in general.

But most modern killer apps are industry-specific and highly specialised. The incentives of most app developers mean that they are unlikely to create apps for a single platform anyway unless their particular industry has standardised on a single OS throughout. Many developers are currently forced to build custom apps for every platform, which instantly destroys the long term viability of many companies.

The alternative for them is to build just one app: a Progressive Web App that will work like a native app everywhere except on iOS, where it will offer a functional but degraded experience.

The choice for Apple isn’t between having exclusive access to exciting new apps on the one hand and sharing those apps with other platforms on the other. The choice is between being the worst platform for the business apps everybody is using or being just as good as the rest, while continuing to differentiate on hardware and User Experience.

I strongly believe that if Apple does not decide to go all in on Progressive Web Apps at some point over the next twelve months, this will be seen as the moment when the iPhone turned into post-iPhone Nokia.

You can also find me on Mastodon and Bluesky