Cohort analysis

Also known as: Customer cohorts

Cohort analysis is a method that groups customers by when they first bought, usually the month of their first order, and tracks each group's repeat purchases, revenue and profit over time.

Formula

Repeat rate (month n) = customers in cohort who reordered by month n ÷ cohort size
Cumulative revenue per customer (month n) = cohort revenue to month n ÷ cohort size

Worked example

200 customers placed their first order in January, spending $12,000. By the end of June, the same 200 customers have spent $18,000 in total, and 60 of them have ordered again.

  1. Six-month repeat rate = 60 ÷ 200 = 30%.
  2. First-order revenue per customer = $12,000 ÷ 200 = $60.
  3. Six-month cumulative revenue per customer = $18,000 ÷ 200 = $90.
  4. Repeat revenue = $18,000 − $12,000 = $6,000, or 50% on top of the first order.

Comparing January with later cohorts at the same age shows whether retention is improving.

Why it matters for a Shopify store running ads

Store-wide averages mix old loyal customers with new ones, which hides trends. Cohorts compare like with like: customers of the same age. That is the only reliable way to see whether a new acquisition channel, product or offer brings customers who come back.

For stores buying customers on Meta and TikTok, cohorts are also where honest LTV and payback numbers come from. A cohort acquired during a heavy-discount sale often behaves very differently from one acquired at full price.

Common mistakes

  • Reading young cohorts as if they were complete; a two-month-old cohort hasn't had time to reorder.
  • Using calendar months for repeat rates instead of months since first order.
  • Mixing countries or channels with very different behaviour into one cohort.
Work it out with your own numbers: LTV calculator

FAQ

What is a cohort in ecommerce?

A cohort is a group of customers who share a starting point, most commonly the month of their first order. Tracking cohorts shows how customers acquired at different times behave over their lifetime.

How often should I review cohort analysis?

Monthly is enough for most stores. Look at each cohort at the same ages, such as 30, 90 and 180 days, and compare cohorts from different channels, offers or seasons.

Related terms

Want this number for your own store without a spreadsheet? ProfitFly connects Shopify, Meta and TikTok and shows your real profit every day, free up to 50 orders a month.

Start free