Retention Analysis: N-Day Matrix & Curves | Keentics

Retention analysis answers the question that decides whether a product survives: do the people who try it come back? It groups users by the day they first showed up and tracks how many return on day 1, 7, 30 and beyond. For game and app teams, retention is the single most honest growth signal — acquisition can be bought, but retained users compound. Keentics computes retention on full event data so the curves reflect every user, not a sample.

Read the N-day retention matrix

The classic view is a cohort matrix: each row is a sign-up day, each column is a day offset, and each cell is the share of that cohort still active. Reading down a column tells you whether retention is improving release over release; reading across a row tells you the shape of a single cohort’s decay. Keentics renders both the matrix and a smoothed retention curve, so you can spot the day-1 cliff and the long-tail flattening at a glance. See the retention rate glossary entry for the precise definitions.

Define “retained” the way your product works

Not every product means the same thing by “active.” Keentics lets you choose the return event — a session, a level start, a purchase — so retention reflects the behavior that matters to you. You can measure classic N-day retention, unbounded (rolling) retention, or bracketed retention for slower loops. A hyper-casual game and a B2B app will draw very different curves, and both should be measured on their own terms.

Compare retention across segments

A single curve is a baseline; the insight is in the comparison. Split retention by cohort, acquisition channel, platform, country or first-session behavior and watch which groups stick. This is how you learn that users who complete onboarding retain twice as well, or that one paid channel looks cheap but churns by day 3. Pair it with DAU/MAU to connect cohort retention to the active-user numbers leadership tracks.

From retention to revenue

Retention and money are the same story told twice. Users who come back are the users who eventually pay, so retention curves are the leading indicator for LTV analysis. When a retention experiment moves the day-7 number, you can follow those same cohorts into the LTV matrix and watch the payback improve. And if a funnel is leaking, retention shows whether the survivors were worth keeping.

Why teams pick Keentics for retention

• Full-data cohorts, no sampling — small early cohorts stay trustworthy.
• Flexible return definitions for fast and slow products alike.
• Read-only SQL to verify any cohort against raw ClickHouse events.
• Fixed pricing, never per project, with a free tier to start and full export at any time.