It’s very common for clients to be interested in what GA4 says about repeat visitation. This makes sense as getting people to come back over and over again is a major part of most websites’ strategies. And GA4 has these as metrics, as well as the first_visit event which fires whenever it thinks someone is visiting for the first time. However we usually tell people to not put too much stock in any of these.

The reasons are very similar to the reasons listed in the previous post about why most websites’ user counts should not be considered exact. Here’s the list again, adapted for first-time and repeat user metrics.

If your website doesn’t get a lot of traffic and has no login area

In this case, GA4 differentiates between first-time and repeat users by dropping a cookie with a randomly-generated ID — a bit like when you take a ticket when you enter a deli. A user who has this cookie is considered a repeat visitor. Someone that doesn’t is considered as a new visitor. There are 2 problems with this:

• Cookies are specific to the device and browser, so the same person will be considered a new user on a new browser/device.
• If a user deletes their cookies they will also be considered a new user. Note that some browsers do this automatically (eg. Safari would delete the cookie after just 7 days.)

This has these implications:

• Some of the new users in your GA4 reports are actually returning users
• Your returning user count is higher than GA4 says
• Your new user count is lower than GA4 says

If your website requires users to log in

In this case, your user count will be very accurate (including for new and returning users), as long as you’ve implemented the User ID feature. In that case, no matter what device someone is on, you are sending the same ID to GA4 for each event so it will know the true user count.

If your website gets a decent amount of traffic

In this case, you can potentially use Google Signals to turn on Google’s modelling. This will help you recover some of the lost data. For example, if someone uses multiple devices but are logged into Google with each device, GA4 may count them as a repeat user when they’re coming back through another device. But this still isn’t fooproof.

Bottom line is that these metrics will provide a basic trendline, and you’d want to watch for changes. However, like with most other analytics reports you’re better off not thinking of them as truths about your website but rather as the results of a messy experiment – meaningful but not definitive.