What the users metric counts
Analytics does not see people. It sees an identifier: a value stored in a browser, a logged-in User-ID if the site supplies one, or a signal from a platform account where that is enabled. The users metric counts how many distinct identifiers appeared in the period you are looking at.
GA4 reports more than one version of it. Total users counts every identifier seen; active users, which fronts most standard reports, counts those that were actually engaged. New and returning describe whether an identifier had been seen before. These are different questions, and a report that says only users without saying which is inviting an argument later.
The count is also period-bound and cannot be added up. A month’s users is not the sum of that month’s daily users, because someone who visited on four days is one user in the month and four in the daily rows. Any spreadsheet that totals a column of user counts produces a number that describes nothing.
Why users matter
Users is the closest available answer to how many different people the site reached, and it is the right denominator for questions about audience size — how many people saw the site at all, how many came back, how many needed more than one visit before enquiring.
Read next to sessions it also describes buying behaviour. A high number of sessions per user means people are returning to think it over, which is normal for expensive or considered purchases, and tells you the site needs to be persuasive on a second visit rather than only on the first.
Common mistakes with users
The biggest is presenting the figure as a headcount of real people. One person on a phone and a laptop is two users. One person who cleared their browser data, used private browsing, or came back after their browser expired the stored identifier becomes a new user again.
This is particularly pronounced where traffic arrives through social apps. Links opened inside the Facebook, Instagram or Messenger browser sit in separate storage from the phone’s main browser, so the same person arriving from a post and later from a search can be counted twice. In markets where most visits come from a phone and most of those from a social app, that inflation is not a minor rounding issue.
The third mistake is comparing users with a platform’s reach figure. Reach is counted against logged-in accounts and covers a different population entirely.
How to act on it
Name the metric precisely in every report — total users or active users, for a stated period — and never sum it across periods. If the site has genuine accounts or logins, implementing a User-ID gives a materially better count, because it stitches the same person’s devices together where you have a legitimate basis for doing so.
Expect the number to be affected by consent choices and browser restrictions, and treat the trend as more trustworthy than the absolute value. Agreeing these definitions once, before they reach a dashboard, is a large part of useful reporting and dashboards.