Your SaaS dashboard's cohort retention chart is probably wrong. It might say you lack product-market fit when you actually have it. Or it might overstate retention and give you false confidence.
Here are the three most common mistakes and how to fix them.
1. Wrong denominator: signups instead of paid customers
The mistake: Counting everyone who signed up in a month as part of that month's cohort.
The problem: Many signups never convert to paying customers. Including them skews your retention rate downward.
The fix: Base your cohorts on paid customers at the end of the cohort month. Someone who signs up in January but doesn't pay until February belongs in the February cohort.
Non-paying signups belong in no cohort.
This will make your retention curves tighter and higher, reflecting actual paying customer retention.
2. Misaligned cancellation buckets
The mistake: Counting churn in the month a customer cancels, not the last month they paid.
The problem: This overstates churn, especially for annual subscriptions. A customer who cancels in April but pays through December should be counted as retained through December.
The fix: For revenue retention, count customers as retained until their last paid period ends.
3. Mixing signup and activation cohorts
The mistake: Combining customers who signed up in a month with those who activated (performed a key action) in that month.
The problem: This mixes two different populations, making your retention curve an unreliable average.
The fix: Choose either signup month or activation month as your cohort definition. Be consistent and label your chart clearly.
- Product-led growth SaaS: Activation cohort is usually more meaningful.
- Sales-led SaaS: Signup cohort (or contract start cohort) often makes more sense.
Why it matters
A wrong retention chart can lead to bad decisions. You might abandon a viable product or double down on a failing one. Investors will scrutinize your data, so make sure your chart is accurate.
Correct calculation
For revenue retention, calculate each cell like this:
retention[cohort_month][period_n] = (MRR from cohort customers in cohort_month + period_n months) / (MRR from cohort customers in cohort_month) - Cohort customers: Customers whose first paid month was the cohort month.
- MRR in period: Recurring revenue from those customers in (cohort_month + period_n).
- Customers count as "MRR contributing" only in months they actually paid.
For customer retention, replace "MRR" with "count of paying customers."
Validate manually: Pick a cohort, pull the customer list, and manually check payments. Compare your manual calculation to your dashboard. If they don't match, your dashboard is wrong.
This is a shorter version of a guide I wrote at Frontkit Studio. The full post has the complete walkthrough.
