What Are Unique Users? GA4 Users, Sessions and Page Views Explained
Published:
Last Updated:
Category: Web Analytics
Authors: Shusaku Yosa
Unique users describe users counted once within a defined period and measurement identity. They are different from visits and page views. The number is not a verified census of every real person who used a site.
In GA4, report the actual metric name, such as Total users or Active users, rather than using “UU” as an interchangeable label. The period, identity, scope and collection conditions all affect what the number means.
Unique users are a measurement definition, not a census
Repeated activity associated with the same identity can be deduplicated within the selected period. However, a person using different browsers or devices may appear as more than one measured user if those identities cannot be connected. A shared device can also make measured identities differ from real people.
GA4’s reporting identity uses User-ID, device ID and modelling according to the selected method. It is not correct to say that everyone on one IP address always counts as one user, or that a hostname identifies a person. See Google’s reporting identity documentation.
Count users, sessions and page views in one example
The following fictional website example uses the same period and identity conditions throughout. A and B are measured identifiers, not confirmed individuals.
Identifier | Sessions | Page views |
|---|---|---|
A | 2 | 3 |
B | 1 | 1 |
Whole period | 3 | 4 |
Total users is 2, sessions is 3 and page views is 4. A’s second session does not create another unique identity. Web page views divided by Total users gives 4 ÷ 2 = 2 page views per measured user. Dividing by Active users instead would be a different calculation, so name the denominator.
A session groups activity. GA4’s default timeout is 30 minutes of inactivity, but check the setting used by the property. A session should not be explained simply as starting again every time someone leaves a page. Google describes the rules in About Analytics sessions.
Total, active, new and returning users in GA4
GA4 offers several user metrics. Check what a report’s “Users” label means rather than assuming it is Total users. Differences between the metrics can be expected; they do not automatically indicate an error.
Metric | What it counts | Common mistake |
|---|---|---|
Total users | Unique users who triggered any event in the period | Treating it as a verified list of people |
Active users | Unique users meeting engagement or specified event conditions | Assuming it has the same definition as Total users |
New users | New unique users with first_visit or first_open | Calling all of them first-time customers |
Returning users | Unique users with a previous session | Adding categories without checking overlap |
Sessions | Groups of activity | Calling sessions people |
Views | Page or screen views | Ignoring whether app screens are included |
Active-user conditions include engaged sessions and specified events or parameters, such as first_visit or engagement_time_msec on the web. In the A/B example, inspect those conditions before giving an Active users total; the view-count table alone does not establish every identifier’s status. The definitions are in Google’s user-metrics guide.
Why daily or channel user counts do not simply add up
Deduplication happens within the range being counted. If the same identity appears in two daily rows, adding those rows counts it twice.
Range | Identifiers | Unique users |
|---|---|---|
Monday | A and B | 2 |
Tuesday | B and C | 2 |
Both days | A, B and C | 3 |
The daily sum is 4, while the two-day unique count is 3. B appears on both days. Similar overlap can occur across channels or devices, so calculate the total using the required period and conditions instead of summing every row.
New, returning and channel categories are not automatically mutually exclusive. Before adding counts, ask whether the measure is additive across that particular breakdown.
Check the metric, date range, identity and data quality
Start by identifying the property and documenting the existing conditions. You do not need to change measurement settings just to read the report. If conditions changed, record when and assess whether the periods remain comparable.
- Write down the exact user metric.
- Record the date range and property time zone.
- Match the production host, web or app scope, language and channel filters.
- Check reporting identity and the data-quality information shown.
- Note limitations such as consent, blocked tags, collection gaps or identifier changes.
- Recalculate the required range instead of adding unique-user rows.
Thresholding or modelling may affect the available data. An integer on a dashboard does not mean every individual has been identified precisely. Small differences should be interpreted in light of the conditions and the quality information.
Use the numbers to form a testable question
More measured users may mean a page reached more identities. It does not establish that they were the intended audience, completed a useful action or became customers. Follow the journey using appropriate entry, engagement and verified-outcome measures.
For example, a fictional increase from 100 to 120 Total users does not prove more registrations if the registration success event has not been validated. Mark that outcome as unmeasured, then investigate the next observable stage. A useful report states the metric, range, conditions and limitations alongside the number.
