Twinkle Twinkle Little Star: Why App Markets Should Ditch the 5 Star Rating System

It’s 2012 and we’re still letting a five star rating system drive a multibillion dollar Android & iOS ecosystem despite all the problems with this sort of filter? People tend to rate only at the extremes (one or five stars). Lots of known and rumored spam issues with developers getting their apps uprated and voting down competitors. And does three stars mean everyone thought it was mediocre or is it the average of polarizing one and five star ratings?

While 3rd party curation and discovery app services such as Chomp are trying to solve the problem, there’s a great opportunity to make the native markets into better experiences.

What are some of the under-exploited recommendation/discovery/ranking paths today?

  • Co-Installation: What apps are also used by people who have the same apps installed as I do
  • Social graph: Which apps are my friends using – this is easy enough to do anonymized and in aggregate 
  • Consumer segmentation cohorts:
    • Apps used by people of same demographic, geographic
  • Actual Usage is a better signal than ratings
    • % of installs where still actively used 30, 180, 365 days after install
    • Intensity of usage/engagement
    • Is it on people’s home screen?
All of these could be used for improving search rankings within app markets, creating truly interesting app browse pages and generally helping connect consumers with new developers.
And of course this type of information isn’t just valuable to consumers, it’s also gold for developers. I hope to see an explosion in analytics for app developers driven by more information provided by OS owners.