Cohort Retention Template

What This Template Is For

Aggregate retention metrics lie. A product that reports "78% monthly retention" might be improving rapidly or decaying slowly, and the single number cannot tell you which. Aggregate rates blend old cohorts (who have already churned the most) with new cohorts (who have not had time to churn yet). The result is a number that always looks better than reality for growing products and worse than reality for shrinking ones.

Cohort retention analysis fixes this by tracking each group of users (typically grouped by signup week or month) independently over time. You can see whether January signups retained better than December signups. You can see if retention flattens at month 3 or keeps declining. You can see whether a product change improved the retention curve or just shifted it.

This template provides the structure for building cohort retention tables, plotting retention curves, comparing segments, and identifying the critical retention inflection points. It pairs with the Product Analytics Handbook for measurement methodology and the retention rate glossary entry for calculation conventions.

Use the NPS Calculator to correlate satisfaction scores with retention cohorts. The activation funnel template helps you understand whether retention problems start at activation or emerge later.

When to Use This Template

• Monthly or quarterly. Run cohort retention analysis at least monthly to track trends before they become crises.
• After a major product change. Compare the retention curve of cohorts who experienced the change versus those who did not.
• When evaluating growth quality. Fast growth with declining cohort retention is a leaky bucket. Slow growth with improving retention is compounding value.
• When diagnosing churn. If aggregate churn is rising, cohort analysis tells you whether the problem is in new cohorts (activation issue) or old cohorts (value delivery issue).
• Before board meetings. Cohort retention is the single best metric for demonstrating product-market fit or lack thereof.

How to Use This Template

Step 1: Define the Retention Event

Decide what "retained" means. For most SaaS products, it is "logged in and performed a core action." A login alone is weak because users who log in to export data before cancelling count as retained. Define the specific action.

Step 2: Choose Cohort Granularity

Weekly cohorts for high-volume products (1,000+ signups/week). Monthly cohorts for lower-volume products. Use signup date as the cohort key unless you have a specific reason to use first purchase date or activation date.

Step 3: Build the Retention Table

Fill in the table with the percentage of each cohort that performed the retention event in each subsequent period. Period 0 is the signup period. Period 1 is the first full period after signup.

Step 4: Plot Retention Curves

Overlay the curves for each cohort on one chart. This visual reveals whether newer cohorts retain better, worse, or the same as older ones. Look for the "flattening point" where the curve levels off.

Step 5: Segment and Compare

Cut the same cohorts by segment (plan, acquisition channel, persona, geography) to find which populations retain best and worst.

The Template

Retention Definition

Cohort Retention Table

Period-over-Period Decay Rates

Flattening point: Retention curve flattens at period [N] with approximately [%] retained.

Segment Comparison

Best-retaining segment: [Segment] at [%] P6 retention

Worst-retaining segment: [Segment] at [%] P6 retention

Gap: [X pp]. This gap represents [implication for product/growth strategy].

Retention Benchmarks

Improvement Tracking

Analysis Checklist

• ☐ Defined retention event (specific core action, not just login)
• ☐ Chose cohort granularity and cohort key
• ☐ Built retention table for last 8-12 cohorts
• ☐ Calculated period-over-period decay rates
• ☐ Identified the flattening point
• ☐ Compared retention curves visually (newer vs. older cohorts)
• ☐ Segmented by plan, channel, persona, and geography
• ☐ Benchmarked against industry targets
• ☐ Identified the highest-leverage retention improvement opportunity
• ☐ Linked improvement initiative to specific cohort for measurement
• ☐ Scheduled monthly refresh of cohort data

Filled Example: B2B Project Management SaaS

Retention Definition

• Retention event: Created or updated a task (not just logged in)
• Cohort granularity: Monthly
• Period length: Calendar month
• Exclusions: Free trial users who never converted, internal QA accounts

Cohort Retention Table (Monthly, Oct 2025 - May 2026)

Key Findings

1. Retention is improving. P1 retention rose from 58% (Oct) to 69% (Mar), a gain of 11 percentage points. The January onboarding redesign appears to be the inflection point.
2. The flattening point is P4-P5. After month 4, cohorts lose less than 2 pp per period. Users who survive to month 5 are long-term retained.
3. The critical window is P0 to P1. An average of 36% of each cohort churns in the first month. This is where the biggest improvement opportunity remains.
4. Paid users retain 22 pp better at P3. Free plan P3 retention is 31%. Paid plan P3 retention is 53%. This gap justifies investing in free-to-paid conversion during the first 30 days.

Action Taken

Focused Q1 2026 on first-month retention. Redesigned onboarding to get users to create their first task within 10 minutes of signup (previously took an average of 2.3 days). Result: Jan-Mar cohorts show 64-69% P1 retention vs. 55-61% for Oct-Dec cohorts. This improvement, if sustained, adds approximately $48,000 in annual retained revenue per monthly cohort.

Key Takeaways

• Cohort retention is the single most honest metric for product-market fit. It cannot be gamed by growing faster
• The flattening point reveals your natural retention ceiling. Improving the P0-to-P1 drop raises the entire curve
• Segment the data. Aggregate cohort retention still hides differences between plans, channels, and personas
• Track improvement initiatives against specific cohorts. "Retention improved" is vague. "February cohort P1 retention is 67% vs. October cohort at 58%" is actionable
• Run this analysis monthly. Quarterly is too slow to catch problems before they compound

About This Template

Created by: Tim Adair

Last Updated: 3/5/2026

Version: 1.0.0

License: Free for personal and commercial use