Cohort Retention Analysis

A clear retention question, an activity rule and cohort anchor, a first triangle and baseline, monthly refresh and review routines, segment and revenue cuts, and specialist methods from survival curves to layer cake forecasts.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Then open the Marketplace in the app, search for Cohort Retention Analysis, and click Install. It goes into your own files, and works offline from then on.

What it installs.

1

Pillar

1

Area

50

Projects

200

Starter tasks

📊 Data & Analytics

Cohort Retention Analysis

Grouping users or customers by start date or behaviour and tracking how many stay over time to see whether retention is improving.

  1. Writing the retention question before the first cohort table
  2. Defining an active user for retention purposes
  3. Choosing the cohort anchor event and date
  4. Matching the retention period to how often people use the product
  5. User activity table for cohort analysis
  6. First monthly acquisition cohort triangle
  7. Retention curves for the whole customer base
  8. Reconciling cohort sizes with signup and order totals
  9. Retention baseline and where the curve flattens
  10. Monthly cohort table refresh with commentary
  11. Weekly early retention watch on new cohorts
  12. Cohort definitions note and changelog
  13. Quarterly retention review with product and growth leads
  14. Retention annotations for launches, outages and campaigns
  15. Gross and net revenue retention by cohort
  16. Automated checks on the cohort tables
  17. Reactivated and resurrected user tracking
  18. Retention slide for the monthly founder update
  19. Reading a cohort triangle by row, column and diagonal
  20. Classic, rolling and bracket retention compared
  21. Mix shift and survivorship bias in blended retention
  22. Uncertainty in retention rates for small cohorts
  23. Survival curves for time to lapse
  24. Writing cohort retention SQL with window functions
  25. Presenting a cohort chart people can read
  26. Stickiness ratios alongside cohort retention
  27. First-week behaviour cohorts and the activation moment
  28. Retention by acquisition channel
  29. Retention by plan, tier and billing period
  30. Choosing a product analytics tool for cohort work
  31. Moving cohort logic out of spreadsheets
  32. Pruning retention segment cuts nobody uses
  33. Rules for pauses, merges and recreated accounts
  34. Excluding internal, test and fraudulent accounts
  35. Retention evidence for a fundraising data room
  36. Retention read on a major product launch
  37. Annual retention review and next year's target
  38. Cohort retention workshop for the leadership team
  39. Cohort read after a price change
  40. Retention for an early startup with tiny cohorts
  41. Renewal cohorts for annual subscriptions
  42. Two-sided retention for a marketplace
  43. Repeat purchase cohorts for an online shop
  44. Account-level retention for B2B products
  45. Taking over someone else's retention dashboard
  46. Growth accounting with new, retained, resurrected and churned users
  47. Seasonality in cohort comparisons
  48. Layer cake forecast of active users by cohort
  49. Projecting long-run retention beyond observed periods
  50. Cohort retention with consent gaps and lost tracking

Where it starts.

Each project arrives with its purpose, milestones and first tasks written out. Here is how the first three read.

Writing the retention question before the first cohort table

Retention work goes wrong when someone builds a cohort chart before anyone has said what decision it should inform. Writing down the one question, such as whether customers acquired since the onboarding redesign stay longer than those before it, fixes the cohort, the period and the activity definition before any SQL is written. It also tells you when the analysis is finished.

First tasks

  • Ask the requester what decision the retention analysis should change
  • Write the retention question and the decision in one sentence each
  • List the open choices: anchor event, period length and activity definition
  • Get the requester to confirm the question in writing

Defining an active user for retention purposes

Whether a user counts as retained in month three depends entirely on what counts as active, and logging in, opening an email or receiving a push notification are not the same as using the product. Choosing the one core action that represents real value, such as sending an invoice, booking a lesson or placing an order, makes retention reflect behaviour customers chose.

First tasks

  • List every event that could count as activity with its weekly volume
  • Strike out passive events such as email opens and push deliveries
  • Choose the core value action and write the rule in one sentence
  • Compare six cohorts under the old and new activity rules

Choosing the cohort anchor event and date

Cohorting by signup date mixes people who never set up their account with those who became paying customers, while cohorting by first purchase or activation only counts the committed. Neither is wrong, but the choice changes every number on the chart. Picking the anchor deliberately, and stating the time zone and date grain, means month-one retention means the same thing to everyone reading it.

First tasks

  • Pull last quarter's counts for signup, activation and first payment
  • Note how each anchor would change who enters a cohort
  • Choose the primary anchor and record the time zone used for dates
  • Count users whose anchor date is missing or falls before launch

Make it yours.

Install Cohort Retention Analysis, then tell Aurum the date you're aiming for. It checks whether the plan fits the weeks you actually have.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Blueprints are general information to help you organise your own work. They are not medical, financial, legal or other professional advice. See the Terms.