What is cohort analysis in ecommerce?
How does cohort retention analysis work?
Pick a trait that groups customers. The usual one is the month of the first order. Then follow each group over the months after that: how many bought again, and how much margin they brought.
That "something" might be the channel, the offer or the product. Shopify's blog describes cohorts by the channel or campaign of the first purchase, and by the first product bought1. Splitting the same customers those ways is how you find out.
Why does it matter?
An average mixes old and new customers, cheap and expensive acquisitions, strong and weak offers. A cohort keeps them apart.
One analytics vendor's article compares ways to estimate LTV (customer lifetime value): a simple equation, cohort analysis and predictive modelling. It calls cohort analysis "by far the best framework"2. The same vendor sells a cohort and LTV app for Shopify, so treat that as a commercial view. The underlying point still holds: LTV by cohort keeps early and late customers from being blended.
Our view: show cohorts in margin, not only revenue. Our convention is contribution margin before marketing: what is left per order after variable costs such as cost of goods, delivery, payment fees and returns. A cohort can look strong in revenue and still leave little money behind.
Worked example
Easily confused with
Cohort analysis vs Retention
The share of customers who buy again within a given period. Cohort analysis is how you track it across groups.
Cohort analysis vs LTV
What a customer is worth over time. Cohorts are one way to calculate it.
Cohort analysis vs Payback period
The number of months it takes for the margin, not the revenue, a customer generates to repay the cost of acquiring them. You read it off a cohort.
Cohort analysis vs nCAC (new-customer CAC)
Acquisition spend divided by first-time customers only, leaving out repeat buyers. It pairs with cohorts when you ask what a new customer cost.
FAQ
What is a cohort?
A group of customers who share a trait, most often the month they first bought. You follow the group over time instead of mixing it into everyone else.
What is an acquisition cohort?
A group of customers defined by when or how they were acquired: the month, channel or campaign of their first purchase, as Shopify's blog puts it1.
Is cohort analysis the same as retention?
No. Retention is a measure of customers coming back. Cohort analysis is the method that lets you see it group by group.
Why not just use an average?
An average can hide a weak group inside a strong one. Cohorts show where the change started, which an average cannot.
Let’s find what’s actually working.
Next step. Cohorts show who stays. The calculator shows what a sale is worth first. Use the Is your ROAS real? calculator.