What Are Unique Users? Total, Active, and New Users in GA4
Shusaku Yosa
Unique users are the users counted once within a specified period under a measurement system’s identity rules. They are not visits, page displays, or a verified count of people. In GA4, choose a named metric such as Total users, Active users, or New users instead of reporting an unexplained “UU” figure.
Unique users, page views, and sessions count different things
One identified user can return several times and view many pages. A different device or browser can create another observed identity for the same person. Conversely, several people sharing a device can appear under one identity. These limits matter when interpreting audience reach.
Metric | Counting unit | Useful question |
|---|---|---|
Web page views | Measured page displays | How often was content viewed? |
Sessions | A measured period of interaction | Which visits arrived and produced an outcome? |
Users | Deduplicated identities within the period | How broad was the observed audience? |
See the page-view counting guide for repeated displays and the session timeline examples for visit boundaries. Separating these units prevents a rise in viewing frequency from being reported as a larger audience.
Choose the GA4 user metric that answers your question
Some standard GA4 reports use “Users” to mean Active users. That does not make Active users a universal substitute for total reach. The official user-metric definitions distinguish the following groups.
Metric | Qualification | Interpretation |
|---|---|---|
Total users | Unique users who triggered an event during the period | The breadth of collected user activity |
Active users | An engaged session or a qualifying platform-specific condition | Users meeting GA4’s activity rules |
New users | Unique users with first_visit or first_open | Users identified as first-time visitors or app openers |
Returning users | Users with a previous session | Repeat usage; this group can overlap with New users within a period |
Active does not mean “high-intent prospects only.” In addition to engaged sessions, Web users qualify when first_visit or engagement_time_msec is collected; Android uses first_open or engagement_time_msec; iOS uses first_open or user_engagement. Do not assume every active user stayed for a minimum number of seconds.
A user can first visit and then return within the selected period, so New users plus Returning users is not a reliable formula for Total users. “New” also reflects available identifiers: deleting cookies can make a repeat visitor appear new. It is not proof of that person’s first-ever contact with your business.
Understand the identity used for deduplication
The current GA4 reporting identity documentation describes Blended, Observed, and Device-based identities. Blended uses User-ID, device ID, and modeling; Observed uses User-ID and device ID; Device-based uses device ID. Google signals should not be presented as a current reporting-identity method.
Identity input | What it helps connect | Important limit |
|---|---|---|
Device ID | Activity associated with a browser or app instance | Another device, browser, or reset identifier can split usage |
User-ID | Activity associated with a consistently implemented first-party ID | Requires appropriate implementation; does not reconstruct every anonymous or historical visit |
Modeling | Eligible unobserved behavior through modeled estimates | Availability depends on conditions; it is not a list of identified people |
In a fictional example, one customer uses a laptop and a phone with separate device IDs. Device-based reporting may count two users. Consistently sending an appropriate User-ID can associate eligible activity under identity methods that use it. This is conditional, not a promise of perfect cross-device coverage. Shared devices can create the opposite problem, and consent-related gaps can change what is observed. There is no universal rule that user counts must always overstate or understate people.
Do not add daily users to calculate monthly unique users
Consider a fictional two-day dataset. Day one contains A and B; day two contains A and C. Both days have two users, but the combined period contains three unique identities: A, B, and C. Adding the daily values gives four user-days, not four unique users.
Request a deduplicated value for the entire week, month, or quarter you intend to analyze. For activity frequency, use DAU, WAU, MAU, and stickiness ratios with a defined activity condition and matching windows. A period total and the sum of daily activity answer different questions.
Record the denominator in per-user metrics
Suppose a fictional Web dataset has 3,000 page views, 1,000 Total users, and 800 Active users. Page views divided by Total users is 3.0; divided by Active users, it is 3.75. Neither result is inherently incorrect, but they describe different populations.
GA4’s Views per user uses Active users as the denominator, and Views includes both page and app-screen views. It is therefore not automatically equivalent to Web page views divided by Total users. The GA4 Data API metric definitions make these units explicit.
More views per user can reflect useful reading, difficulty finding information, reloads, or duplicate tracking. Pair audience and viewing metrics with the outcome that matters, such as a successful inquiry or continued product use.
A checklist before interpreting a change
- Name the exact user metric and selected period.
- Align time zone, platform, stream, hostname, and language paths.
- Record changes to reporting identity, User-ID collection, consent, or tagging.
- Check thresholds, modeling, and whether data is still being processed.
- Break changes down by acquisition source and audience composition before attributing them to a campaign.
A useful audience report keeps the number together with its period, metric definition, and identity method. That makes changes interpretable without claiming more certainty about people than the measurement can support.