GA4 User Metrics Explained: New vs. Returning Users, Sessions, and Active Users

GA4 User Metrics Explained: New vs. Returning Users, Sessions, and Active Users

Google Analytics 4 replaced Universal Analytics with a fundamentally different measurement model. The shift from session-based to event-based tracking changed how GA4 counts users, defines sessions, and distinguishes new from returning visitors. Most teams that migrated from UA to GA4 copied their old KPI dashboards without checking whether the metrics mean the same thing. They do not.

Key takeaways

1. GA4's primary engagement metric is "active users" -- users with at least one engaged session -- not total users. A user who bounced without engagement is counted in total users but not active users.
2. A GA4 session expires after 30 minutes of inactivity or at midnight; a user can generate multiple sessions in one day, but each session counts only once toward the unique session count.
3. "New users" in GA4 are identified by the absence of a previous user_pseudo_id cookie; a user who clears cookies or switches browsers will be counted as new again.
4. "Engaged sessions" require at least one of: 10+ seconds on page, a conversion event, or 2+ pageviews in the session. Standard bounce-equivalent sessions below these thresholds are not engaged.
5. For acquisition reporting, use "new users"; for retention reporting, use "returning users" and "engaged sessions per user"; for total reach, use "total users" -- do not mix these in the same KPI.

What GA4 actually counts as a "user"

User in GA4: a device identified by a user_pseudo_id -- a device-scoped cookie set by the Google tag. Unlike Universal Analytics, which used a client ID, GA4 uses user_pseudo_id as the default identity space for web reporting.

GA4 supports identity stitching via User-ID (for logged-in users) and Google Signals (for signed-in Google account holders). When these are available, GA4 blends them to report cross-device users more accurately. When they are not, each distinct device and browser counts as a separate user.

Two types of users appear in GA4 reports:

• Total users: every user who triggered any event, including bounce sessions with no engagement signal. This is the broadest count and most comparable to UA's user metric.
• Active users: users who triggered an engaged session during the date range. This is GA4's default user metric in most reports -- it excludes traffic with no engagement, which is why active users typically appears lower than total users.

The switch from total users to active users as the primary metric was intentional. GA4 treats bounced sessions differently from Universal Analytics: in UA, a single-page session with no interaction was still a session. In GA4, a session where the user left in under 10 seconds without triggering any conversion or second pageview is not an engaged session.

New users vs. returning users: what each measures

New users: users for whom GA4 does not recognize a previous user_pseudo_id on the device. Technically, this is users who have not previously triggered a first_visit event in GA4 on that device-browser combination.

Returning users: users who have a previous user_pseudo_id recognized by GA4 on that device.

The distinction matters for acquisition vs retention analysis, but with important caveats:

• Cookie clearing: a user who clears browser data appears as new to GA4 even if they have visited 50 times previously. On some browsers (Safari with ITP), GA4 cookies have a shortened lifespan, inflating new user counts.
• Cross-device: a user who visits on mobile is counted as new when they visit on desktop for the first time, unless User-ID is configured. Without User-ID, GA4 new/returning is a device-level metric, not a person-level metric.
• Incognito: any visit in private/incognito mode is counted as a new user.

For acquisition campaigns, new users is the right metric: it measures whether the campaign is reaching people who have not previously engaged with the site. For retention and re-engagement, returning users and returning user rate (returning users divided by total users) tracks loyalty.

Sessions, engaged sessions, and what bounce rate became

Session in GA4: begins with a session_start event, expires after 30 minutes of inactivity, at midnight, or when campaign parameters change. A single user can start multiple sessions in one day.

Engaged session in GA4: a session where at least one of these conditions is met:

• Duration of at least 10 seconds
• At least one conversion event fired
• At least 2 pageviews or screenviews

This replaces bounce rate from Universal Analytics. In GA4, engagement rate is the percentage of sessions that are engaged (the inverse of "not engaged" which is analogous to a bounce). The default GA4 engagement rate benchmark differs by industry: e-commerce typically sees 55-65% engagement rate; informational blogs 40-55%.

Sessions per user: how many sessions the average user generates in the date range. Higher sessions per user means users are returning and re-engaging within the period -- useful for measuring content or app stickiness.

Prooflytics

Turn scattered analytics into one clear picture

Every source in one brief. The whole picture. Your decision.

Start free trialSee pricing

14 days free · no credit card

The operational problem: wrong metrics driving wrong decisions

The ICP problem this creates for marketing teams: a dashboard shows "Users" rising month-over-month, which is reported as acquisition success. But the "Users" metric is GA4 active users -- which means engagement is required. An ad campaign that drives 10,000 visits where 8,000 users leave in 3 seconds without engaging will show only 2,000 in the "Users" report. The 8,000 bounced visits are invisible. The campaign is perceived as performing better than it is because the metric is filtered to engaged traffic only.

Conversely, a retargeting campaign should drive returning users, but if the dashboard reports new users, the retargeting performance looks weaker than it is.

By the Google Analytics Unique Visitors framework documented in the Prooflytics knowledge base (sourcing MeasureSchool's GA4 metrics reference), the six metrics most commonly confused -- total users, active users, new users, returning users, sessions, and engaged sessions -- each answer a distinct question. Using active users where total users is appropriate will systematically undercount reach by the bounce percentage. Using sessions where users is appropriate will overcount reach for users who visit multiple times.

Prooflytics surfaces session-level and user-level metrics from GA4 in the daily briefing side by side, with engagement rate as the discriminating signal: when engagement rate drops while sessions hold steady, the briefing flags it as a traffic quality shift, not an acquisition success.

How to read each GA4 user metric correctly

For paid acquisition reporting: Use new users and new user rate (new users / total users). This tells you whether paid campaigns are reaching net-new audiences or cycling through the same returning visitors. Benchmark new user rate: most B2B sites see 60-75% new users on paid channels; if below 50%, the paid audience may be overlapping heavily with existing customers.

For SEO and organic content reporting: Use sessions and engaged sessions. SEO content drives multiple visits over time; sessions captures the volume. Engaged sessions identifies which content actually holds attention. Content with high sessions but low engaged sessions suggests high-ranking but low-relevance matches -- the keyword and intent are misaligned.

For email and retention reporting: Use returning users and sessions per user. Email campaigns should increase the returning user percentage among the target cohort. If email drives sessions but returning user rate stays flat, the email is reaching lapsed users who do not re-engage.

For audience sizing and total reach: Use total users. This is the GA4 equivalent of UA users -- the full count including non-engaged visits. Required when measuring top-of-funnel reach for awareness campaigns.

Setting up a user metrics comparison in GA4

Method 1: Standard report with segments

1. Open GA4 and navigate to Reports -- Acquisition -- User acquisition.
2. The default report shows new users. Add a secondary dimension or use the "Add comparison" button to overlay returning users from the User segment.
3. The default date range is 28 days. Switch to a custom range for period-over-period comparison.

Method 2: Exploration report

1. Navigate to Explore -- Blank.
2. Add dimensions: Session default channel group, New / returning (found under User attributes).
3. Add metrics: Sessions, Engaged sessions, Active users, New users.
4. Use a table visualization to compare new vs returning behaviour by channel.

Method 3: Custom events for identity stitching

For accurate new vs returning across devices: implement User-ID by passing a hashed user identifier to GA4 when users log in. This allows GA4 to recognize the same person across devices, reducing inflated new user counts from cookie clearing and cross-device visits.

gtag('config', 'G-XXXXXXXX', { 'user_id': 'HASHED_USER_ID' });

With User-ID active, GA4 reports switch to the User-ID identity space, blending device-level and user-level data for more accurate new vs returning classification.

Bottom line

• GA4 uses active users as the default, not total users -- active users excludes non-engaged visits, making it lower than the user count you saw in Universal Analytics.
• New users in GA4 is device-level without User-ID; cross-device journeys and cookie clearing inflate new user counts.
• Engaged sessions (10+ seconds, or conversion, or 2+ pageviews) replaced bounce rate; use engagement rate as the quality signal.
• Match the metric to the question: new users for acquisition, returning users for retention, sessions for content traffic volume, active users for engagement reach.
• Configure User-ID if you have a logged-in user base -- it is the only way to get accurate new vs returning classification across devices in GA4.
• You can read independent reviews of Prooflytics on G2 and compare it to alternatives in the marketing analytics category.