Google has been getting more aggressive at promoting good Android apps in the Play Store lately, mainly by demoting bad apps in its search rankings. The latest way that’s happening, as Google puts it, is that the company is “incorporating app quality signals in our ranking algorithms.”
What that means is that if you look at the top app rankings or search for an app, it’s less likely that an app that has a history of crashing or a bunch of low user ratings will rise to the top. Google says that it’s looking at “performance data, user engagement, and user ratings” in creating these “quality signals,” and that it will affect an app’s ranking in both search and app listings.
It’s part of a more general effort Google seems to be making to promote high-quality Android apps. Earlier this month, for example, it announced that it is using machine learning to identify apps that are behaving badly — e.g., by asking for more permissions than they ought to be. It has also created a whole new program called “Android Excellence” to curate apps the company thinks are good.
In all, it’s a mix of a lot of machine learning and a little human curation judging the apps that developers are submitting to the store. Android generally gets a somewhat-deserved rap for having lower-quality apps than iOS, and Google seems to be challenging that perception with its very favorite tool: AI.
It’s also a small stick (or carrot, depending on how you look at it) Google can use to encourage developers to simply make apps that crash less. The question, of course, is whether developers will know if their apps are getting downranked by Google’s algorithm or for any number of other reasons.
The thing about algorithms is they’re mysterious. It’s easy to know what to do when your app gets flat-out rejected by an app store: you appeal it. But as news organizations who’ve tried to parse Facebook’s inscrutable sharing algorithms can tell you, there’s no easy way to appeal to a machine.