We noticed something interesting last week in the Android Market rankings: some of our apps increased in rank and others decreased – well beyond the normal ups and downs applications typically see.
After looking at it more closely, we believe that Android Market changed its app-ranking algorithm in advance of the launch of in-app billing, rewarding apps that engage users over those that merely recruit or purchase them.
In fact, the ranking change was dramatic. myYearbook’s core Android app moved up from #63 to #11 in the Top Free Social category as of 3/24, as indicated by Android App Tracker. Now we appear right below Foursquare for Android and just above AIM’s app, where previously we were not even on the first page of results.
While myYearbook was exploding in rank and new installs per day, we saw sudden decreases in smaller apps that we own. It appears to us that the only reasonable explanation is an Android Market algorithm change between 3/20 and 3/23.
So then the natural question is: why did “myYearbook for Android” increase in ranking while an app like “Tic Tac Toe” decreased in ranking?
We believe the new ranking algorithm is now more complex, taking into account a measure of activity in the application and not simply new installs per day. In particular, we believe the ranking now considers either Daily Active Users (DAU) or the ratio of Daily Active Users to Monthly Active Users, which is sometimes referred to as “the sticky factor.” As the chart below makes clear, apps that benefited from the change, like “myYearbook for Android,” are much more engaging for users than those that suffered, like “Tic Tac Toe.”
This conclusion appears to be born out by the fact that other apps known to engage users well, which would be expected to have relatively high DAU/MAU ratios, also increased in rank during the same time. For example, in the Social Free category, Seesmic went from #39 to #10, and Gowalla went from #78 to #18 – all virtually overnight.
Many have discussed Google’s recent change to its search algorithm, which targeted content farms in an attempt to improve search quality, but to date discussion of recent Android Market ranking changes has been minimal, despite the importance of these rankings to app developers. On Reddit, we found a single post pointing out a major improvement for an unofficial Reddit client, but it had received only 11 comments as of April 1. One reason for the lack of chatter is that fewer companies have insight into these rankings, a problem companies like Android App Tracker appear to be helping to solve.
The Android team has acknowledged that they’ve been focused on the ranking algorithm, but only with tangential references in press outreach about their in-app billing product, leaving it up to the development community to sort it out. With in-app billing, it is now in Google’s best interest to highlight more engaging apps, as those apps are more likely to generate paying users.
We think the change is clearly a good thing for users and for creators of engaging applications, but it’s not a good thing for everyone. In particular, less game-able rankings mean that simply buying 50,000 low-quality, incented installs doesn’t guarantee you a top ranking anymore.
We hope that Apple makes the same change to their App Store, as rankings that depend on DAU and retention reward good apps with more users, encouraging the entire app development community to focus on improving engagement – not just spamming users or buying low-quality installs.