Cohort Retention

Cohort retention is the practice of grouping customers by their acquisition period (typically month or quarter) and tracking what percentage of each group remains active at subsequent time intervals. It reveals whether newer cohorts retain better or worse than older ones, making it the gold standard for measuring whether your product and onboarding are improving over time.

formula.sh

Cohort Retention at Month N = (Customers from Cohort Still Active at Month N) / (Original Cohort Size) × 100

> Original Cohort Size — number of customers who started in the same period (e.g., all January signups)
> Customers Still Active at Month N — cohort members who have not cancelled by the Nth month
> Each cohort produces a retention curve: 100% at month 0, declining to a floor over time
> Compare cohort curves across acquisition periods to detect improvement or deterioration

example

example.sh

200 customers acquired in Q1. By month 3: 160 remain (80%). By month 6: 130 remain (65%). By month 12: 110 remain (55%).

Month-3 retention: 160/200 = 80%. Month-12 retention: 110/200 = 55%.

→ Q1 cohort has a 55% 12-month retention rate. If Q2 cohort shows 65% at 12 months, onboarding improvements are working.

why it matters

Cohort analysis separates genuine product improvement from statistical noise. Your overall retention rate at any moment is a blend of all cohorts at different stages of their lifecycle. Only by separating cohorts can you see whether month-1 retention is improving, whether your new onboarding flow works, or whether a product change hurt a specific cohort.

For investors, cohort curves that stabilize at a high floor (called "flattening") are the ultimate proof of product-market fit. A cohort that drops from 100% to 60% in month 1 and then holds at 55–60% through month 24 shows that the customers who survive onboarding stay forever. That flat tail implies lifetime value far higher than the early drop suggests.

Cohort retention also reveals the impact of product decisions. If a pricing change in Q3 caused that cohort to churn faster than Q2, cohort analysis will show it clearly. Aggregate metrics would smooth it over.

common mistakes

✗Looking only at the aggregate retention curve instead of individual cohort curves — you miss trends in both directions.

✗Not waiting long enough to evaluate cohort performance — 12-month cohort data requires 12 months of patience; drawing conclusions at month 3 is premature.

✗Failing to annotate cohort charts with product changes, pricing changes, and acquisition channel shifts that explain cohort differences.

pro tips

✓Build a cohort retention heatmap: rows are acquisition months, columns are retention periods (M1–M24), cells are retention percentages. Color code by value. Patterns emerge immediately.

✓Identify your retention "floor" — the natural long-term retention percentage where cohort curves flatten. Every product improvement should raise this floor.

✓Segment cohort analysis by acquisition channel: SEO, paid, referral, and sales cohorts often have dramatically different retention profiles.

the mrrsucks take

Cohort retention is the truth serum of SaaS metrics. Your overall retention rate lets you lie to yourself. Cohort curves show exactly which customers you're failing and when you started failing them.

faq

How do I build a cohort retention table?+

Group customers by signup month. For each month N after signup, count how many are still active. Divide by the original cohort size. Present as a table with cohort months as rows and elapsed months as columns.

What does a "flattening" cohort curve mean?+

A flattening curve means churn stops accelerating after an initial drop. The customers who survived early months become long-term users. This is a strong signal of product-market fit — there is a core audience for whom your product is genuinely valuable.

The churn spiral›

related metrics

./install-the-daemon

$9. 365 roasts. one public endpoint of pure shame.