GA4 cohort exploration: retention and user lifetime reports
This sits in the wider ecommerce marketing technology stack. This page covers the retention and lifetime reports inside GA4. For the general idea, see cohort analysis.
What does it do?
Three reports are in play.
Cohort exploration. Google's glossary gives examples of cohorts such as all new app users today, or customers making their first transaction in the analysis period, and says cohort analysis can help show how well a site or app retains users1. You build it in Explore from the template gallery1. Cohort analysis in the plain sense means grouping customers by a shared trait, usually the month of their first purchase, and tracking them over time.
Retention overview. Google describes a pre-made report that summarises how well your site or app retains users, including the share of users who return each day in their first 42 days2. It groups users into cohorts by the day they were acquired and shows Day 1, Day 7 and Day 30 lines, plus the average revenue from new users over their first 120 days2. In the shop sense, retention is the share of customers who buy again within a period. GA4 measures returning users, which is not the same thing.
User lifetime exploration. Google says this technique shows how users behaved over their lifetime on your site or app, for example which source, medium and campaign brought users with the highest lifetime revenue3. Lifetime data covers only users active after 15 August 20203. Read the notes on sampling and user identity before you trust a number.
Why do e-commerce brands use it?
It is already there. The data is in your analytics account and the reports cost nothing to open.
It gives a first look at whether people come back, and which sources bring people who keep coming back. Our view: that is a useful question to ask early, before anyone builds a data warehouse.
How do we implement it?
Check purchase tracking
Purchase events and user identifiers must be recorded properly. If they are not, fix that before building any cohort.
Write the definitions
Decide whether a cohort starts at first visit, first purchase or first campaign. Compare only cohorts built on the same rules.
Build in Explore
Use the cohort and lifetime explorations for questions shop data cannot answer, such as how visitors from different sources return.
Reconcile with the shop
Compare GA4 results with shop data before anyone quotes them. For repeat purchase rate, use the shop's own figures.
When should you not use it?
Do not use GA4 as your retention number for the board. Use shop data.
Do not use it if purchase tracking is broken. Fix that first.
Do not compare cohorts built on different rules. Our view: a comparison like that only measures the rules.
What does it cost in effort?
No figures. The effort is:
- Set-up. Purchase events and user identifiers must be recorded properly, or the cohorts are built on gaps.
- Definitions. Decide what a cohort is: first visit, first purchase or first campaign. Write it down.
- Reading. Explorations are built by hand. Someone has to pick the settings and explain them.
- Reconciling. Compare GA4 results with shop data before anyone quotes them.
Pitfalls
Users are not customers
A new device can look like a new person. Retention in GA4 can differ from retention in your shop.
Signed-out activity
When reporting by User-ID, Google says the lifetime exploration uses only the signed-in part of a user's activity3. Guest checkouts can drop out of the picture.
Revenue as lifetime value
The lifetime exploration reports revenue-based value. It carries no cost of goods, returns or delivery.
Sampling
In the user lifetime exploration, Google samples once the date range includes more than 1M users on the free product (10M on the paid one), then scales the sample up3. Check whether your report is sampled before quoting it.
Mixing definitions
The repeat purchase rate formula you use changes the answer on the same data. Choose one and keep it.
Treating source as cause
A source whose users return often may simply attract buyers who would have returned anyway. That is a correlation until tested.
Our take: Selective
In our view, GA4 cohort exploration is Selective. That is our rating, not a market fact.
Our reasoning: GA4 counts users, not customers. Without a user ID, GA4 identifies a user by a device ID, which on websites is the Analytics cookie's client ID3. So one person on two devices can count as two users. Consent choices and blocked cookies can also leave visits unrecorded. Your shop knows who bought. GA4 knows who it saw.
Use it where the shop data cannot answer the question, such as comparing how visitors from different sources return. For repeat purchase rate and purchase cohorts, use your shop platform's own reports or export.
Our view: LTV (customer lifetime value) in GA4 is built from revenue. Revenue is not margin. For a profit view, someone has to subtract the variable costs of each order.
FAQ
What is a cohort in GA4?
Google's glossary defines it as a group of app or website users that share a common characteristic identified by an Analytics dimension, such as the day they were acquired1.
What does the GA4 retention report show?
Google describes it as a summary of how well your site or app retains users, with the share returning each day in their first 42 days and Day 1, 7 and 30 cohort lines2. It follows users, not confirmed customers.
Does GA4 show customer lifetime value?
GA4's user lifetime exploration shows metrics such as lifetime revenue per user, as a total, an average or a percentile3. That is revenue, so it is not profit per customer.
Is GA4 user lifetime data sampled?
It can be. Google says the user lifetime exploration samples above 1M users on the free product and 10M on the paid one, then scales results up3.
Is GA4 good enough for repeat purchase rate?
Not on its own. Our view: your shop platform knows who bought twice. GA4 only knows the users it recorded.
Why do my GA4 and shop figures disagree?
Different counting. GA4 counts users and events; your shop counts orders and customers. Consent and cookies also change what GA4 sees.
Sources
- Cohort - Analytics HelpPrimary source
- [GA4] Retention overview report - Analytics HelpPrimary source
- [GA4] User lifetime - Analytics HelpPrimary source
Let’s find what’s actually working.
Next step. Want to know which retention number to trust? Talk to us about your setup.