What’s a user?

One of the major changes in GA4 compared to the old Google Analytics is that GA4 was designed to be more user-centric. But what’s a user isn’t that simple. In theory a user is supposed to represent a unique person who interacts with your website(s) or app(s). But then we come to the main problem of any analytics tool: how do we tell that someone is the same person when they might be visiting on different devices and browsers?

• If your website doesn’t get a lot of traffic and has no login area (or other way for people to identify themselves), the answer is that generally you can’t. GA4 would be using its default way to differentiate between users, which is dropping a cookie with a randomly-generated ID, a bit like when you take a ticket when you enter a deli. Any events with the same ID are considered as being from the same person. But the problem is cookies are specific to the device and browser, so the same person will show up as multiple people in your reports. This generally means your real user count is lower than GA4 says. And conversely your real user conversion rate is higher than your GA4 says. Because user conversion rate is conversions / users, if the denominator is over-reporting, the fraction is under-reporting. Also note that whenever a user clears their cookies (or their browser does it for them), they will be counted as a new user, further inflating your user count.
• If your website requires the user to login to do anything then your user count will be very accurate, 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.
• A lot of websites have both a public and a login area. Then your user count will still be an overestimate for public pages but will be accurate for your login area.
• If your website lets the user identify themselves without logging in (eg. by filling out a lead form), you can still improve your user count by implementing user-provided data collection. But this wouldn’t affect the majority of sessions where the user doesn’t provide any identifiers.
• If you have enough traffic and activate Google Signals, Google can also model the real user count. It does this by trying to tie people’s browsing behaviour to signals like their Google account, since you might be logged into Google on multiple devices even if you aren’t logged into the website itself. This often doesn’t work and you need quite a bit of traffic for it to be useful but it’s an option.
• Finally it’s important to remember that all your GA4 data including users is also undercounting because of people who use ad blockers (which generally block GA4 completely) or have not consented to cookies on your website (if you have that set up).

What’s the difference between “active users” and “total users”?

This one’s a bit easier. Total users is everyone that GA4 is aware of. Active users is anyone who also engaged with your website or app in your reporting period. So it would exclude people who left the website very quickly. In our experience, the two metrics are very close for a lot of industries, to the point where you might not want to bother differentiating. But it’s best to check this out for yourself.

What’s a session?

A session is a single instance of a user interacting with your website. So if the user does multiple things in a short period of time, this counts as a single session. This metric becomes important when you want to see how often people are coming back, since the same user can have multiple sessions. Two things to note:

• A new session starts after the user stops interacting with your website for at least 30 minutes, but you can adjust this. Whether you want to do this depends on what you consider as a session for your own brand/website.
• A user clicking through to your website from a different traffic source (eg. Google search results vs an email newsletter link) does NOT start a new session unless they’ve had the 30 minute timeout.