What About Removing App Store Ratings Entirely?

This post by John Gruber touched off a lively debate on whether and how developers should prod users for ratings. The nag screens sure are obnoxious, but good ratings are pretty important in bringing your app to the surface in a very crowded App Store. Rather than enter the debate, I wonder why users have to explicitly rate apps at all, when they are implicitly rating them all the time? What if, instead of reviews, Apple were to score apps by a variety of criteria, awarding some number of (undisclosed) points, on a per-user basis, for:

• Being currently installed
• Being in the dock
• Being on the first page
• Being on the second page
• Being on the same page (or in the same group) with a bunch of other apps that get a lot of use
• Being regularly updated
• Time spent in active use (maybe just for games, as it’s tricky given the wide range of usage patterns across app types)
• Having push notifications enabled and sending them with some regularity (it is implicit that the user finds this valuable rather than annoying, and this catches apps that serve a purpose but that the user doesn’t launch often, so aren’t in a prominent screen location)

And an app would lose points (or at least not score) for:

• Being on a page or in a group with a bunch of other apps that mostly never get launched and never send push notifications
• Crashing (maybe?)

This is a half-baked thought, and you’d want to weight these somehow, but it seems to me like this would be a decent measure of ongoing value to the user. Certainly, when I consider my own phone, these criteria give a very accurate picture of current value to me. Sure, there will be users that don’t organize their phones at all, but they would be pretty easy to discard (or deemphasize) from the dataset based on usage patterns, I’d think.

Apple could then incorporate this into App Store search weighting, along with category, search keyword, download trends, etc. Maybe they could also start spidering the web (if they don’t already) looking at incoming links for an additional weighting factor (in addition to direct links to apps in the App Store they could look at links to developer websites associated with those apps (which they have on file)).

This won’t solve the “rich get richer” problem (which does seems pretty baked into capitalism), but Apple could blunt the feedback loop a bit by:

• Introduce some randomness into the search result ordering
• Mix in some relative newcomers each time

So the rich will keep getting richer (as they should, in many cases) while still exposing some up-and-comers.

Finally, under a scheme like this developers aren’t penalized for updating their apps (currently an update wipes out ratings).