Cohort Analysis

Cohort analysis groups your users by a shared starting point, like their install week or the source that brought them in, and tracks how each group's retention and revenue hold up over time.

You are staring at a dashboard that says one campaign got you installs for cheap and another cost you a small fortune, and the cheap one looks like the obvious winner. Cohort analysis is what stops you from acting on that half of the story. Instead of judging everyone as one undifferentiated blob, you line up the users who arrived in the same week, or through the same channel, and you watch that specific group age. How many are still opening the app on day seven? Day thirty? How much has that group spent by day sixty?

The reason this matters is that it lets you compare like with like. A cohort from your December sale might churn fast and spend little, while a cohort from a boring evergreen campaign sticks around and quietly turns into your best revenue. On the flat dashboard those two look similar. On a cohort curve they look nothing alike, and you finally see which sources produce users worth having. A channel that costs you more rupees per install can still be the one you pour budget into, because its users pay you back many times over.

The catch is patience. Cohorts take real time to mature, and you cannot judge last week's cohort this week. A group that is only seven days old has barely started spending, so reading a fresh cohort like a finished one will fool you into killing winners early and scaling losers. Give each cohort the weeks it needs before you draw conclusions.

Stop asking what a campaign cost you, and start asking how its people aged.