Retention Cohort Analysis Template

Cohort analysis is one of those things marketers say they do and actually do inconsistently. You pull a report when things go wrong, present it in a QBR, and then go back to watching month-over-month averages until the next crisis. Averages hide the cohort signal. By the time you notice a retention problem in aggregate numbers, the underlying cause is usually six months old.

This worksheet is not a dashboard. It is a forcing function. Fill it in quarterly and you will see the decay curve clearly enough to act on it - before it shows up in churn alerts.

Why cohort-level retention beats aggregate retention

Aggregate retention mixes customers who signed up in January with customers who signed up in October. If your January cohort is churning at 8 percent a month and your October cohort is churning at 3 percent - because you fixed onboarding in the spring - your blended rate looks fine. Nothing in the average tells you onboarding was the problem or that October cohorts are now behaving differently.

Cohort-level analysis keeps each signup month isolated. You watch Month 0 through Month N for a specific group of customers, and patterns that were invisible in the blend become obvious.

Logo retention vs revenue retention: pick the right lens

Logo retention counts the percentage of customers still active. Revenue retention - sometimes called net revenue retention - tracks the percentage of recurring revenue still active, including expansions and contractions.

A business can have 85 percent logo retention and 105 percent net revenue retention if surviving customers expand. A business can also have 90 percent logo retention and 85 percent net revenue retention if the customers who leave happen to be the biggest ones. Both numbers are telling different stories. This template asks you to track both so you know which story you are actually in.

If you want to understand what each customer relationship is worth over time, run your cohort retention numbers alongside your customer lifetime value calculator. The two analyses answer adjacent questions.

The worksheet

Reading the curve: what a healthy cohort looks like

A healthy retention curve drops in the first one to three months - some churn in any cohort is normal - and then flattens. The flat section represents the customers who are genuinely getting value. If you never see a flat section, you have not identified a retained core, which is a product signal, not a marketing one.

The useful comparison is cohort over cohort. If your January 2026 cohort has 5 percent better Month 3 retention than your January 2025 cohort, something you changed in product, onboarding, or activation is working. If the curves look identical across twelve months of cohorts, interventions are not moving the number and you need a different diagnosis.

Where this connects to your broader retention program

Cohort analysis is the measurement layer. It tells you where the problem is in the customer lifecycle. It does not tell you how to fix it. For the strategic layer - how to build programs that address each stage of the curve - the retention marketing strategy guide covers the playbook that sits behind these numbers.

One trap to avoid: pulling cohort data once, presenting it in a meeting, and moving on. The value compounds when you track the same cohort across multiple periods and watch whether your changes actually moved the curve. The template is only useful if you fill it in on a schedule.

Frequently asked questions

How many customers do I need in a cohort for the data to be useful?

There is no fixed threshold, but cohorts under 30 customers produce noisy curves where a few churn events swing the percentage significantly. If your volume is low, group by quarter instead of month, or combine plan tiers. The goal is a curve you can read with confidence - not statistical significance at a precise sample size.

Should I use logo retention or revenue retention as my primary metric?

For most B2B SaaS businesses, net revenue retention is the more decision-relevant number because expansion and contraction affect your actual revenue more than headcount. For consumer subscriptions with flat pricing, logo retention is usually the right primary. Track both regardless - the gap between them tells you something about your customer mix.

What counts as Month 0 vs Month 1?

Month 0 is the signup month itself. Month 1 is the first full calendar month after signup. Some teams define it by billing period instead of calendar month. Either works as long as you apply the same definition to every cohort. Mixing definitions is the most common reason cohort curves look inconsistent across time periods.

The worksheet takes about an hour to build the first time and fifteen minutes to update each month once it is set up. That is a reasonable investment for the only analysis that shows you whether the customers you are acquiring are actually staying.