Most app ideas start with "I would use this". Mine start with a spreadsheet, because an app I would use can still be an app nobody searches for.
For the app that became Backmark I had three candidates: an antique identifier, a crochet pattern app and a settlement finder. I did the same thing for each one.
The scan
For every niche I ran six searches on the App Store and took the top 200 results of each, then removed duplicates and apps that only matched by accident (yarn games, for crochet). For every app left I looked up an estimate of its monthly revenue, its release date and its reviews.
That gives three questions I care about more than the total size of the market:
- How many apps actually make money, at least $5,000 a month?
- What share of all the apps is that?
- Are apps released this year among them, or is the money all with apps from 2015?
The third one matters most. A niche where only old apps earn is a niche where new apps don't get found.
What came out
Antiques: 272 apps, 11 of them earning $5,000 a month or more, about $794,000 a month between them. And of the 108 antique apps released in 2026, 4 were already earning.
Crochet, without the yarn games: only 2.2 percent of the apps made money.
Settlement finder looked better on paper, with 9 percent of apps earning, but the whole pie was around $303,000 a month, and only 1 of the 45 apps released in 2026 had broken through.
Antiques had the combination I was looking for: real money, spread over more than a couple of apps, and new apps getting in.
Then the reviews
Numbers pick the niche. Reviews pick the product. I read what people wrote about the antique apps that earn, and the pattern was clear: confident answers that turned out to be wrong, and paywalls that felt like traps. Backmark's two big decisions, honest confidence and an honest paywall, came straight from there.