--- title: "Retention analysis: cohort retention and what drives it" description: "Group people by when they started, then follow each group period by period. Count a return as any visit or as the one behavior that matters, and split it by…" url: https://countly.com/feature/retention last_modified: 2026-11-11 --- Retention # Who comes back, and what brings them back Group people by when they started, then follow each group period by period. Count a return as any visit or as the one behavior that matters, and split it by who people are to find the users who stay. In this demo car app, about seven in ten new drivers return in their first week. The curve shows where the drop stops and the regulars begin. Part of analytics ## One question, three honest answers “Did they come back?” depends on what you count. Pick the reading that fits the decision. The people in each cohort stay the same; only the way a return counts changes. - Did they come back that week? Classic active in that period, gaps allowed The everyday reading. Someone who skips a week and returns the next still counts for the week they came back. - Did it become a habit? Full active in every period, without a break The strict reading. Only people who never missed a period count, so what is left are your real regulars. - Who could we still win back? Unbounded active in that period or any later one The forgiving reading. It counts everyone who ever returned, which sizes the audience a win-back campaign can still reach. From curve to decision ## Four questions a retention review should answer Retention is one number in a deck and four questions in a planning meeting. Each has its own view, on one page. - What does retention look like? Milestones at weeks 1, 4, 8 and 12, a curve that shows the onboarding cliff and where it flattens, and every cohort row by row. - Do they repeat what matters? Swap visits for a behavior, a trip, a purchase or a lesson, and see whether it becomes a habit or stays a one-off. - Who stays longest? Split by platform, country, acquisition source or cohort, up to five values, each with its own curve over the whole. - Is it getting better? Follow one checkpoint across every cohort against its average, and check the cohort sizes before you celebrate. Numbers you can defend ## Ready for the questions people ask in the meeting A retention number is only useful if nobody can talk it down. The page answers the usual objections before anyone raises them. “The newest week always looks terrible.” Young cohorts A cohort that hasn’t reached week 8 has no week 8 value. The cell stays empty instead of counting as zero, so recent weeks never read as a collapse. “One small week is skewing the average.” Weighted curve Weighted lets larger cohorts count for more, and every row shows its size first, so a flattering rate from a handful of people is easy to spot. “Is this better than last quarter?” Comparison window Pick an earlier window and it appears behind the curve and the trend, and beside every cell in the grid. “Can we keep this view?” Shortcuts and dashboards Save the query as a shortcut that always opens on the latest cohorts, put it on a dashboard, or export the grid. Works with ## Retention, with the rest of Countly Every feature reads the same live data, so an answer in Retention doesn’t stop there. These are the handoffs built in. [All features](https://countly.com/platform/features) Cohorts Filter retention to a cohort, or split it by cohort to compare audiences side by side. Open [(Cohorts Filter retention to a cohort, or split it by cohort to compare audience](https://countly.com/feature/cohorts) Profiles Filter and split by profile properties and profile groups, your own custom properties included. Open [(Profiles Filter and split by profile properties and profile groups, your own cu](https://countly.com/feature/profiles) Dashboards Add the query on screen to a dashboard as a retention tile. Open [(Dashboards Add the query on screen to a dashboard as a retention tile. Open)(ht](https://countly.com/feature/dashboards) Data ownership ## Every return counted on data that stays yours Retention follows each person over weeks and months. That history is computed where it is stored, never copied to another service to be measured. [Data privacy](https://countly.com/platform/data-privacy) Aspect Flex Enterprise Environment Dedicated single-tenant instance Your cloud, your data center or air-gapped Data residency The Google Cloud region you choose Wherever you deploy, including air-gapped networks Vendor access Operated by Countly; access limited, granted by you and logged No default access ISO 27001 Certified ISO 27017 Certified ISO 27018 Certified SOC 2 Type II examined TISAX® Participant* *Countly Ltd. is a TISAX® Participant. TISAX® is a registered trademark of and governed by ENX Association. Learn more at [enx.com/tisax](https://enx.com/tisax)