AnalyticsFunnels, cohorts and retentionIntermediateAnalyses your data

Cohort retention analysis and interpretation

Turns a cohort table into retention curves and a careful read of what changed between cohorts, with hypotheses clearly separated from findings.

For: Performance marketers, Marketing analysts, Growth marketers · Works with: Any capable chat model (ChatGPT, Claude, Gemini, others), Google Analytics 4, Google Sheets / Excel

Public beta. This prompt was drafted with AI assistance and checked by automated rules, but it has not been reviewed or tested by a person yet. Treat it as a starting point and check the output. It is hidden from search engines while in beta. Use the “Was this prompt useful?” box to tell us what works.

Your prompt

Act as a product-and-marketing analyst. Analyse the cohort retention table below. Definition: a cohort is users grouped by signup month; "retained" means completed at least one session in the period; the period is month. Business context: a note-taking app with a free plan. Data (cohort | size | retained counts by period): Cohort | Size | M1 | M2 | M3 Jun | 2000 | 900 | 700 | 620 Jul | 2300 | 940 | 690 | n/a Aug | 2600 | 1100 | n/a | n/a Tasks: 1. State assumptions and check data consistency (retained counts should not exceed cohort size; flag incomplete recent periods). 2. Compute retention percentages for every cohort and period from the counts, showing the formula, and recompute to verify. 3. Describe the shape of the curve (initial drop, flattening or continuing decline) and compare cohorts, noting only differences that are large relative to cohort size; give the reason for your caution about small cohorts. 4. List hypotheses for any cohort differences (acquisition source, onboarding changes, seasonality, product changes) as hypotheses with the data needed to test each, without claiming causes. 5. Suggest 3 cuts to run next (by source, plan, geography, device). 6. List missing data. Rules: do not invent retention numbers, benchmarks or causes. If data is missing or not provided, say so. Distinguish correlation from causation. Do not compare cohorts at different ages as if equal.

Customize

The fields start with example values so you can see how the prompt works. Everything stays in your browser.

How cohorts are formed.

Retention definition.

Week, month, etc.

Business context.

Paste your table. The default is small fictional sample data so the preview works; replace it with your own.

Example of what to expect

Illustrative only. It describes the kind of result this prompt aims for; real output varies by tool, model and run.

A consistency check, retention percentages per cohort and period with formulas, a description of curve shapes, cautious cohort comparisons, hypotheses with tests and next cuts. (Illustrative.)

Expected format: Assumptions, retention table, curve description, hypotheses, next cuts.

How to use it

  1. Make sure cohorts are compared at equal age.
  2. Check cohort sizes before comparing.
  3. Cut by acquisition source next.

Limitations

  • Small cohorts produce noisy retention rates.
  • Retention definitions change conclusions; be explicit.

Platform notes

Any capable chat model (ChatGPT, Claude, Gemini, others)

Official docs read · checked 2026-10-11

Works in any capable chat model. Paste a small anonymised sample and recompute the key numbers yourself: chat models can misread columns or miscalculate, and they cannot run your statistical tests reliably.

General notes on Any capable chat model (ChatGPT, Claude, Gemini, others)
  • These prompts are plain text and work in any chat assistant. For analytics prompts, paste a small, anonymised export; a chat model can misread columns or miscalculate, so recompute key numbers yourself.

Google Analytics 4

Official docs read · checked 2026-10-11

GA4 exports and the interface can differ (sampling, thresholds, attribution, data freshness); state which one your numbers came from.

General notes on Google Analytics 4
  • The BigQuery export has daily events_YYYYMMDD tables and intraday tables; intraday lacks some fields, and late-arriving data can update a daily table for up to 3 days.
  • Exported data and the GA4 interface can differ; the export documentation links to separate comparison articles that we did not read.

Google Sheets / Excel

Not verified: third-party guidance only · checked 2026-10-11

Test any formulas on a small range first; syntax depends on locale and version.

General notes on Google Sheets / Excel
  • Formulas are written for common functions in both products. Locale settings (decimal and argument separators) can change what you must type.

Source, license and attribution

Origin
Original by MarketerTools
Publisher
MarketerTools
License
Original work by MarketerTools, free to copy and use. Informed by the linked documentation; no third-party text is reproduced.
Platform assumptions checked
2026-10-11

Documentation and references behind this prompt

These informed the structure and the platform notes. A reference is not a license, and no third-party prompt text is copied here.

Spotted an attribution error, or are you a source owner with a request? Use the chat button at the bottom right and quote prompt ID ana-030. We will correct or remove it.

Related prompts

  • AnalyticsIntermediateAnalyses your data

    LTV model sanity check from cohort data

    Stress-tests an LTV figure against cohort retention data, exposing hidden assumptions about lifespan, margin and discounting.

    Any capable chat model (ChatGPT, Claude, Gemini, others) · Google Sheets / Excel

  • AnalyticsIntermediateAnalyses your data

    Conversion funnel drop-off analysis with hypotheses and tests

    Finds the biggest relative and absolute drop-offs in a funnel table and turns them into prioritised, testable hypotheses.

    Any capable chat model (ChatGPT, Claude, Gemini, others) · Google Analytics 4 · Google Sheets / Excel

  • AnalyticsIntermediateAnalyses your data

    App retention diagnostics: D1, D7 and D30 by acquisition source

    Compares day-N retention across acquisition sources with maturity and sample-size checks, so weak channels are flagged only when the data supports it.

    Any capable chat model (ChatGPT, Claude, Gemini, others) · Google Sheets / Excel