The store stopped
being a shop.

Five features on this desk describe five unrelated events. A set of American age-verification statutes. A flat year for installs. A retention gap in AI apps. A collapsing middle in subscription growth. Three regulators arriving at three incompatible remedies.

They have one thing in common, and it is not the app store. It is what the app store became while everyone was arguing about its commission.

The shape underneath

Take the five findings and ask, of each, what the developer controls.

The age bracket is determined by the store. The size of the install pool is set by the market. The capability an AI app sells is licensed from a model provider. A third of Android churn is a payment retry inside Google's billing system. The commission rate is set in Tokyo, Brussels and a federal courtroom, and one of those three may change its mind next term.

In every case the developer carries the outcome and does not hold the input.

That is the change. The app store used to be a distribution channel with a checkout attached - you built a product, the store carried it, and the arrangement was a fee for shelf space. What it is now is a set of dependencies: identity, demand, capability, billing and price, each supplied by someone else, each with its own failure mode, and none of them auditable from inside your product.

The thing that did not change

Feature 14 of this edition is the counterweight and belongs at the end of this desk rather than the start.

Two landmark rulings, a European regulatory regime and roughly five years of industry argument produced, in the most permissive market available, revenue leaving App Store billing of 3–6%. Ninety-four to ninety-seven per cent of the money stayed where it was. And the store the industry spent half a decade litigating accounts for roughly a third of the mobile app economy; advertising is most of the rest, and almost nobody was arguing about it.

That finding sits underneath everything on this desk. The commission fight was won and the economics barely moved. Meanwhile identity became a platform function, the install market went flat, the fastest-growing category started leaking, the revenue distribution stretched from 200-to-1 to 400-to-1, and three regulators produced three incompatible answers.

None of those were on the agenda. All of them arrived anyway.

What the desk adds up to for a practitioner

The unifying practical point is unglamorous: the competence that decides outcomes has moved from building the product to managing the things around it that you did not choose.

Recovering involuntary Android churn is worth more than most acquisition work and requires no acquisition. Knowing whether your retention fell because your product got worse or because your install base doubled is a cohort chart, not a strategy offsite. Whether your trial is three days or twenty-one is a working-capital decision with a 17-point conversion consequence. Whether your age-rating assessment survives a Louisiana suitability challenge is a review someone has to sign.

None of that is product work. All of it decides more than the next feature will.

What to do about it

Audit your dependencies, not just your roadmap. Write down the five inputs your business runs on that you do not control - identity signal, install supply, model provider, billing system, commission rate - and what happens to you if each one moves 20% against you. Four of the five moved this year.

Move effort to the operational layer. Dunning, cohort analysis, trial configuration and rating review are unfashionable and they are where the recoverable money is. The store commission is the loudest cost in mobile and, as Feature 14 established, not the largest one.

Assume the calendar will move. Three of four statutory deadlines shifted inside twelve months, and a Supreme Court ruling could reset the US commission. Build compliance and pricing as configuration, not as assumptions.

Distrust every blended benchmark, including ours. A median across 115,000 apps, or a 27.1% AI share spanning categories that run from 61.4% to 6.2%, describes no one. Use them to orient, then measure your own.

Ask what is not being argued about. The industry spent five years on a third of the economy and moved single digits. The changes on this desk arrived with no campaign behind them at all. That asymmetry is the most reliable thing in this edition.