Which cohorts should I use?
Start with install cohorts: players grouped by the day or week of their first launch. Then split them by one more trait, such as the first build, the platform or the acquisition source. Keep one trait per comparison, so that you know what made the difference.
Sources: Unity Analytics terminology documentation; Manalens retention documentation.
What is a worked example?
Example numbers: 1,000 players start your game in the week of 1 September. In the next week, 400 of them come back, so week 1 retention is 40%. You ship a new tutorial, and 1,200 players start in the week of 8 September. 540 of them come back in their next week, so week 1 retention is 45%. The newer cohort retains better.
Sources: worked example; Manalens retention documentation.
Why do cohorts matter for indie PC and web games?
An average over all players mixes old fans with new players. After a launch, a sale or a creator video, the new players can pull the average down even when the game got better. Cohorts compare like with like: players who joined at the same time, on the same build. In October 2026 the median PC game kept 6.54% of new players on day 1, so a small change per cohort can matter.
Sources: GameAnalytics Benchmarks Hub.
How do I build cohorts in Manalens?
Add a Retention chart to a dashboard. The entry event defines the cohort: people are grouped by the week or day of their first entry event. The Install cohorts chart on the Game KPIs template groups players by their first game_open and counts a return with session_start. Read rows to compare cohorts and columns to see where players stop.
Sources: Manalens retention and metrics documentation.
Frequently asked questions
What is the difference between a cohort and a segment?
A cohort is usually fixed by when a player started, such as the week of their first launch. A segment is any subset of players split by a trait, such as platform or source. Unity defines segments as subsets of your player base, split apart by key differentiators. You can apply a segment to a cohort chart.
Why does my newest cohort show a dash?
Because its period is not complete yet. In Manalens, periods that are not complete at the end of the date range show a dash, not a low number. A cohort from last week cannot have week 3 retention yet. Compare cohorts only on periods that are complete for all of them.
Which trait should I use to split a cohort?
Start with install cohorts: players grouped by the day or week of their first launch. Then split them by one more trait, such as the first build, the platform or the acquisition source. Keep one trait per comparison, so that you know which trait made the difference.