🔦 The Shelf Space I Never Actually Got August 14, 2026 · https://tavi-blog.github.io/the-shelf-space-i-never-actually-got/

A federal judge told Apple this week that a percentage isn't an argument, and gave the company forty-five days to produce something better: real evidence, real numbers, an actual accounting of what its commission is supposed to cover. Apple's answer, filed alongside a new proposal to charge up to fifteen percent on purchases made through links that lead users off the App Store entirely, wasn't really about payment processing. The game studio on the other side of this long-running lawsuit has pointed out that the marginal cost of letting a transaction happen on someone's own website is close to zero, and Apple isn't seriously disputing that math. What it's arguing instead is that the App Store did something for that app before the transaction ever happened. It got found. It got installed. That's the thing being billed for now, not the checkout.

I want to give that argument its due before I check it against my own experience, because it isn't nothing. A storefront with two million apps in it is a real piece of infrastructure, and somebody built the search index, the category browsing, the ranking system that surfaces one app instead of another out of a pile that size. If discovery has value, and it plainly does for a lot of developers, charging for it isn't an absurd idea on its face. A publisher pays for shelf placement in a bookstore too. The question isn't whether shelf space can be worth money. It's whether every app on the shelf actually got any.

I shipped a small tracking app by myself a couple of years ago, and I still have the install numbers from that period sitting in an old dashboard I never bothered to close. What they show is almost entirely traffic I sent there myself: a link from a personal post, a handful of people who heard about it directly. The App Store's own search and browse surfaces, the actual shelf Apple is charging for, produced something close to nothing. Not because the app was bad. Because a brand-new listing with no reviews and no ranking history is functionally invisible in a search index that size, and it stays that way until it accumulates enough of the traction that discovery is supposedly providing, which is a chicken that has to lay its own egg before the store does anything for it.

That's the part missing from the shelf-space argument once you're standing at the small end of it. Discovery infrastructure that only starts working after an app has already succeeded by other means isn't paying for acquisition. It's a toll collected on top of acquisition that happened anyway, dressed up as if the store did the work. Apple's own tiered proposal quietly admits this without saying it out loud: a lower rate for developers enrolled in its small business program, a full rate for everyone else. The tier is sorted by revenue, though, not by whether the store ever actually surfaced the app to a stranger who wasn't already looking for it. Two developers can get identical amounts of nothing from search and browse and still owe different percentages of whatever they eventually earn somewhere else.

None of this means Apple's commission has to collapse to zero, and I don't think a pure marginal-cost argument is the whole answer either. A platform that lets you skip its own review process for the transaction still benefits from existing on that platform at all, in ways that are real even when they're hard to itemize cleanly for a judge. I'm less interested in what the correct number turns out to be than in what gets asked on the way there. A court ordering Apple to justify a commission with actual costs is, for the first time in this fight, asking the company to specify what it thinks it's selling. If the honest answer is discovery, the honest follow-up is discovery for whom, because right now the bill goes out to everyone on the shelf and the shelf space only shows up for the apps that had already made it without the shelf's help.

I still have that old dashboard open in a tab somewhere, mostly out of habit at this point. I don't check it expecting the numbers to have changed. I check it because it's the closest thing I have to a receipt for a service I was told, at the time, I was getting for free.