Ranking assumes effort is a signal
Search ranking, recommendation and reputation all lean on the same buried assumption: that producing a listing costs something, so the act of producing one carries information. Reviews assume a seller cares about their reputation because building it was expensive.
When a listing can be produced in minutes, none of that holds. The signal that ranking was reading is gone, and no amount of algorithm work recovers information that was never there.
The scale is not hypothetical
In the first half of 2026 Apple's App Store took roughly 560,000 new apps while downloads grew about two percent. Directories in this category index tens of thousands of servers with no quality signal attached to any of them.
The result everywhere is the same shape: a handful of listings capture almost everything, and the rest are functionally invisible. Sensor Tower put the top one percent of apps at over ninety two percent of in-app purchase revenue in 2025.
Curation inverts the economics
If listing is free, listing is worthless. If listing requires passing a test that costs real effort to pass, the act of listing carries information again.
That is the whole strategy here, and it is why the rubric is published: a barrier nobody can see is arbitrary, and a barrier everybody can read is a standard.
What a quality floor actually does
Every ranked surface on this site excludes listings below a completeness threshold rather than ranking them lower. The difference matters. A ranked-lower listing is still on page four accumulating impressions; an excluded one is not there at all.
The cost is that new listings have to be finished before they appear. That is the intended cost. A catalogue that shows unfinished work teaches people that most of what they see is not worth opening.
Interest and quality are different questions
Popularity is the easiest signal to collect and the one most easily gamed, which is why nothing on this site ranks by it. Wanting something is not evidence that it works.
So the two are kept apart. Want signals tell makers what to build. Verified reviews and measured runs tell buyers what works. Mixing them would give you the worst of both: a popularity contest wearing the clothes of a quality signal.