How many months of data do you need for a useful retention curve?

At minimum, six months of data for each cohort gives a reasonable picture of retention patterns. Twelve months is better, especially for annual subscription products. You need enough time for the curve to reveal its shape: whether it flattens into a stable retention level or continues declining. Early-stage companies can start with shorter windows and extend as data accumulates.

Should retention curves use customers or revenue?

Both are valuable. Customer-based curves show whether people stay. Revenue-based curves show whether the money stays and grows. A customer retention curve might flatten while a revenue curve continues growing if retained customers expand their spending. Build both and compare them to understand the full picture of customer health.

What does a smiling retention curve mean?

A smiling curve initially declines but then curves upward, meaning some cohorts start spending more over time. This is rare but extremely positive: it indicates that retained customers are so engaged that their increasing spend offsets some of the initial churn. It is most common in revenue-based cohort curves for products with strong expansion revenue.

SaaS & Subscription MetricsIntermediate4 min read

What Is a Cohort Retention Curve?

A cohort retention curve tracks how customer retention changes over time for groups of customers who started in the same period. Learn how to read and use them.

Key Takeaways

A cohort retention curve groups customers by their start date and tracks what percentage remain active over time.
The shape of the curve reveals whether churn is front-loaded, steady, or accelerating.
Comparing cohort curves across time periods shows whether your product and onboarding are improving.

What a cohort retention curve shows

A cohort retention curve is a chart that tracks the retention rate of a specific group of customers who all started using your product in the same time period. The horizontal axis shows time elapsed since signup, and the vertical axis shows the percentage of the cohort still active. A steep initial drop followed by a flattening curve indicates early churn that stabilises. A continuously declining curve signals ongoing retention problems at every stage of the customer lifecycle.

Reading retention curves

The ideal retention curve drops initially and then flattens into an asymptote, meaning a stable group of long-term users emerges. If the curve never flattens, you have a leaky bucket where customers leave at every stage. The steepness of the initial drop reveals onboarding effectiveness. A curve that drops 40% in the first month but retains 55% after twelve months has a different story than one that drops 15% monthly forever. The shape matters more than any single number.

Comparing cohorts over time

The real power of cohort analysis is comparing curves from different time periods. If your January cohort retains better at month six than your October cohort did at month six, your product or onboarding improvements are working. This comparison isolates the effect of changes you have made from external factors. African SaaS startups iterating rapidly can use cohort comparisons to validate whether each product sprint genuinely improves the customer experience.

Building actionable retention curves

Create separate curves for different customer segments: pricing tiers, industries, acquisition channels, and company sizes. This segmentation reveals which customer types have the healthiest retention and which need attention. Overlay product usage data to identify the behaviours that predict long-term retention. If customers who complete onboarding within the first week retain at 90% versus 50% for those who do not, you have a clear lever to pull.

What Is Net Revenue Retention (NRR)?4 min · Intermediate What Is Gross Revenue Retention (GRR)?6 min · Intermediate What Is Logo Churn?5 min · Beginner What Is Net Revenue Retention?4 min · Intermediate What Is Gross Revenue Retention?4 min · Intermediate What Is Logo Churn?3 min · Beginner