[youtube http://www.youtube.com/watch?v=oUiYcwTlXIw&w=560&h=349]

The first statistic we see for any application usually falls into the category of “vanity metrics,” explained Anton Commissaris, chief revenue officer for Appsalar. Vanity metrics are those “top line” metrics such as number of new users, number of sessions, and daily active users. These stats often only act as “feel good” metrics, not helping you understand anything about your users’ behavior. They can hide significant problems or opportunities. A better way to look at user statistics, so as to improve user engagement, retention, and monetization, is to conduct cohort analysis, explained Commissaris. A cohort analysis is the process of grouping users by something they’ve done collectively, such as:

• All users that registered on a specific date.
• All users that downloaded a specific song this month.
• All users that reached level five in a game.

Once you’ve created such a cohort, you can then analyze their behavior with respect to engagement, retention, and monetization. One example Commissaris gave was the cohort of users that reached level five in your game. Say you track that cohort and notice that if users get to level five, they continue using your game for six weeks more, as opposed to the cohort of users who don’t reach level five (they might only last two more weeks). If that’s the case, then you want to build an experience in your app to try to drive more people to reach level five, which will hopefully lead to greater retention and monetization.