Aglet

Learn From Repeated Cohort Retention Confusion

A retention lesson keeps the member timeline that made a cell ambiguous. Record entry rule, eligible denominator, activity window, re-entry, labels, and decision context. Then improve the metric dictionary or table review so future cohort changes explain membership and return semantics before a curve is compared.

Keep the lesson for the next incident

  1. Retain cohort exemplars

    Keep early-return, late-return, no-return, re-entry, missing-date, and boundary cases with expected cells and member sets. Record period basis, filters, cumulative mode, and export labels. Preserve the original ambiguous table beside the agreed definition for future review.

  2. Classify recurring disputes

    Review prior questions for entry changes, denominator drift, period alignment, cumulative resets, missing activity, and labels that omitted return semantics. Group named cases by missing contract or review signal. Avoid assigning a universal cause to every retention disagreement.

  3. Assign measure stewardship

    Make the metric owner maintain cohort and return definitions while the report owner checks tables, charts, filters, and exports after changes. Add a trigger for new eligibility or window rules. State which timeline cases the fixture detects and which product questions need human agreement.

What to carry forward

The learning record is complete when it links retention confusion to a missing cohort definition or review cue, preserves member timelines, and names an owner and trigger. Keep intentionally different first-return and cumulative measures visibly separated so future readers do not merge them.

Technical background: Metabase documentation.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow