Skip to content

Answers

How do you check cohort retention before buying a small SaaS?

Short answer: Ask for a subscription-level export from Stripe or the billing system, group customers by the month they first paid, and track what share, and what MRR, is still paying each month after. If recent cohorts fall faster than older ones, or curves never flatten, the headline churn rate hides a weaker business.

Updated 28 October 2026 · Figures checked against their sources on 28 October 2026 · By Joan Murphy (pen name) · How we check · Corrections

What is a cohort retention check?

A cohort is a group of customers who started paying in the same month. A retention check follows each group over time: how many are still paying one, three, six and twelve months later, and how much monthly recurring revenue (MRR) they bring. A single average churn rate can look fine while recent customers leave much faster than old ones. Cohorts show that difference, and they show whether growth comes from customers who stay or from a stream of signups that replaces those who leave.

What data should you ask for?

An export of every subscription from Stripe or the billing system: customer ID, plan, price, start date, cancel date and any refunds. Customer names can be removed.
At least 12 months of history, more if the business is older.
Notes on price changes, discounts, free trials and annual plans, because each changes how a cohort looks.
A screen share of the billing dashboard, so you can check the export totals against what you see; see how to verify Stripe revenue.

If the seller cannot or will not give a subscription-level export, the listing's churn figure cannot be checked. Our checker raises the no churn figure flag when a listing gives none at all.

How do you build the table?

Group customers by the month they first paid.
For each group, count how many were still paying at the end of each later month, and divide by the starting count. That is logo retention.
Do the same with MRR instead of customer counts. That is revenue retention, and it includes upgrades and downgrades.
Put one row per signup month and one column per month since signup, then read down each column: month 3 retention for the January group, the February group, and so on.

A spreadsheet is enough. Treat annual plans separately, because they cannot churn until renewal and will make early months look better than they are.

What should you look for?

Recent cohorts that fall faster than older ones. This is the most common thing an average hides, and it often follows a new ad channel, a lifetime-deal launch or a price cut.
Curves that keep falling. Healthy cohorts usually drop early, then flatten. Our rule of thumb: if no cohort has flattened by month six, value the business on the revenue it would have after the drop, not today's MRR.
Growth that depends on signups. If MRR is flat but each month needs many new customers to replace those who left, the business depends on its acquisition channel; see one channel over 60% of traffic.
Revenue retention far above logo retention. A few large customers upgrading can hide many small ones leaving; check customer concentration.
Stated churn that does not match the table. If the listing says 3% a month but the cohorts lose far more, ask how the seller worked it out.

What does a composite check look like?

This is a composite example, not a real deal. A small SaaS lists $3,000 MRR and "about 4% monthly churn". The cohort table shows customers from a year ago still at 70% after six months, but the last three cohorts, which came from a lifetime-deal promotion followed by a switch to monthly billing, are at 45% after three months and still falling. The buyer runs the SaaS churn stress test with the recent cohorts' rate, values the business on about $2,400 MRR, and offers on that figure, with part of the gap as an earnout on retained MRR.

Run the listing through the free red-flag checker first; it flags missing or high churn before you ask for the export. See also the small SaaS due diligence checklist.

Frequently asked questions

How many months of data do I need for a cohort check?

Our rule of thumb is at least 12 months of subscriptions, so you can see six months of retention for at least six cohorts. With less history, use the table to spot trends rather than to set the price on its own.

Should I use logo retention or revenue retention?

Look at both. Logo retention tells you whether customers stay; revenue retention tells you whether the money stays. A big gap between them usually means a few large accounts are carrying the business.

What if the seller only shares a churn percentage?

Ask for a subscription export with customer names removed. A single percentage cannot show whether recent customers behave differently, and a seller with healthy cohorts usually has no reason to refuse.

Sources

FirstPass Memo, red-flag checker rules, our own rules, published as /rules.json
FirstPass Memo, seller question bank, October 2026 (general checklist), our own questions
FirstPass Memo, free buyer calculators, our own calculators; the formulas are on each page

Related

Get the 47 seller questions for SaaS and software listings

Enter your email and the next page gives you the SaaS and software seller question checklist (web page and PDF) straight away, no waiting, along with the other free files: 11 seller question checklists (151 questions across eleven business types), the 72-hour vetting plan and a 14-column deal-tracker spreadsheet. We also email you the links so you have them later.

No newsletter: we send one email with your links,. Reply "stop" to hello@firstpassmemo.com to be removed. Looking at a live listing priced $10k–150k? Get a full memo: $79, pay after delivery instead; its confirmation page has the same downloads.

Composite examples are invented and labelled. Figures from other sites are quoted with a link and the date we read them; FirstPass figures are asking prices and seller-stated profit, not sale prices. Informational screening only, not investment, legal or tax advice.