Retention is an important metric to look at when predicting the success of a game. The general definition of retention is how many players have stayed in a game on a given day relative to when they first started playing, but as you'll see below, there are several ways to calculate this.

When evaluating games, retention is one of the key pieces of information that we're looking at. Strong retention coupled with strong monetization is the real winning combo, but strong retention won't mean the same thing for every game. The baseline of good metrics for a CCG will be very different from the baseline of metrics for an idle game, but before comparing across different genres or games, you have to understand what's being measured.

It is important to understand that different analytics platforms calculate this differently. A true comparison can only happen when the calculation is the same; due to this it is important to know the source of your data. It is also worth noting that different platforms will label these differently.

Hard Retention (Kong Preferred)

• How it is calculated: what percentage of players came back to the game on Day N
• Where it is used: This is the most popular calculation and can be found in all of Kongregate’s Dashboards and analytics platforms such as GameAnalytics and DeltaDNA.
• Why it is used: It gives a true idea of the overall retention level of the game.
• Day 7 example: Only players who came back precisely on their D7 will be considered retained. It doesn’t matter if they came back on previous days or not.

Rolling Retention

• How it is calculated: what percentage of players came back to the game on or any day after Day N.
• Where it is used: This is not very popular, but some prominent platforms, such as Flurry, use it as their default.
• Why it is used: It provides the churn rate, as this is really the % of players who have not dropped out of the game.
• Day 7 example: Players who came back on or any day after D7 are considered retained for D7; they don’t necessarily need to show up on D7.

Full Retention

• How it is calculated: what percentage of players came back every single day up to Day N
• Where it is used: Not very popular and there are no prominent analytics platforms leveraging this.
• Why it is used: It is very restrictive, but it does give an idea of the level of engagement with the game.
• Day 7 example: Only players who came back every single day from D0 to D7 are considered retained.

Kongregate uses Hard Retention to benchmark all of our games, as well as the games that we evaluate, because it gives the truest picture of how many people are sticking around in a game. Obviously, the sample size needs to be big enough to get a significant value for the true retention, but that's a topic for a different post. You can find more information about how to avoid this, and other data pitfalls, in Emily Greer's "Data Gone Wrong" talk.