🏷️ The Box I Had to Check Anyway September 12, 2026 · https://tavi-blog.github.io/the-box-i-had-to-check-anyway/

I opened App Store Connect this week to push a small update to a caffeine tracker I built a couple of years ago, and instead of the usual short form, there was a new question sitting between me and the submit button: does this app include social media capabilities. Underneath it, Apple's own definition, the ability to redistribute, amplify, or let people interact with content other users posted, through a feed or something like one. Answer yes and the listing gets locked to a minimum age rating of thirteen, with a new descriptor sitting on the product page next to whatever else the app already discloses. The questionnaire has been sitting in App Store Connect since midsummer. Starting this month, you can't skip it.

My app logs a coffee or a green tea against a running daily total, stored on the device, with nothing that leaves it and nothing that comes back from anyone else's phone. There's no comment section, no follower graph, no feed of anyone's drink log but your own. Reading the definition once should have been enough to answer no and move on. It took me longer than that, because the phrase doing the real work, "or similar discovery method," isn't written to describe my app. It's written to describe every clever way an app could functionally be social without calling itself that, and a definition built to close that gap doesn't stop being broad just because the thing in front of you plainly isn't the target.

I think the width is defensible, more than defensible, once you consider what a narrower rule would actually have protected against. Age verification requirements only work if they can't be routed around by a product that just avoids the word "social" in its own marketing while shipping the exact mechanics the rule is trying to catch, a public feed of other people's posts, a way to amplify what a stranger shared. Apple isn't wrong to assume that the moment it writes a rule with a loophole-shaped gap in it, somebody with more lawyers than I have will find the gap. A definition specific enough to leave me instantly certain would probably also be specific enough for a real social product to engineer its way around it.

What that defensible choice costs doesn't show up anywhere in the reasoning behind it, though. A large platform answering this questionnaire has a legal team reading the same sentence I read, and almost certainly a standing policy for what counts before anyone on staff gets near the form. I have the form, the definition, and whatever judgment I can bring to it alone, on an app whose entire feature set I already know because I wrote every screen of it myself. The honest answer here really was no. The cost isn't in the outcome. It's that "obviously no" and "confirmed no after actually reading the rule closely" turned out to be two different amounts of time, and the difference between them is exactly the kind of unpaid interpretive labor nobody arguing about the size of Apple's commission is counting.

There's a version of this that would genuinely worry me, if the app did anything even slightly borderline: a shared log, a public streak, one questionable interaction feature away from tripping the definition on a technicality. Mine doesn't, so the actual stakes here were about as low as they get. What stays with me is noticing how much of the review process I go through now has this shape: not a gate checking whether the app does what it claims, but a gate checking whether I've correctly classified what it is, in someone else's language, built for categories nobody was picturing an app like mine sitting inside of when they wrote them. I answered the question in under a minute once I was sure I understood it. Getting to sure took the rest of that sitting.