Stop 08 · Marketing Analytics & DataTopics · 7/11

Cohort Analysis for Marketing

In short

Total users going up can hide a nasty secret: new people arrive while old ones quietly leave. Cohorts show whether each batch of customers actually sticks.

Grouping users by a shared starting point (e.g. signup month or acquisition channel) and tracking their behavior over time — reveals whether newer customers are actually higher or lower quality than older ones.

The blob in a graduation cap poses with two small groups of classmates side by side.

Think of it like this

College batch reunions

The 2022 batch still has 40 people in its group chat. The 2023 batch has 12. Same college, very different stickiness. Comparing batches by their joining year is exactly how cohort analysis works.

Watch it happen

How many January sign-ups are still active

Week 0: 100%

Week 1: 45%

Week 2: 32%

Week 3: 27%

Week 4: 25%

1/5A cohort is a group that started together. Here, 1,000 people who signed up in January.

The big ideas

What you need to know

Group by start date

Cohorts usually group people by sign-up or first purchase week or month.

The cohort table

Rows are cohorts, columns are weeks since start, cells are the percent still active.

Compare cohorts

If newer cohorts retain better, your product or onboarding is improving.

Cohorts by channel

Group by acquisition source too. Some channels bring buyers who stay much longer.

Save yourself the pain

Common mistakes

Oops

Judging retention from total active users.

Do this instead

Totals mix old and new people. Cohorts separate them so you see real stickiness.

Oops

Comparing a cohort from last week with one from last year.

Do this instead

Compare cohorts at the same age, like 'week 4' for both.

Your mission

Read a cohort table

~25 min

Quick check

Monthly users are rising, but every new cohort's month-2 retention is falling. What does that suggest?

Go deeper

1Sketch a cohort table (acquisition month vs retention over time) by hand.
2Learn why comparing cohorts by acquisition channel can reveal which channels bring better long-term customers.
3Understand why raw growth numbers can hide a quietly worsening retention trend.