Why your Usermaven numbers differ from GA4

If you run Usermaven and Google Analytics 4 side by side, the numbers will not match — and that alone is not a sign that anything is broken.

The two tools answer subtly different questions. GA4 is built around Google's advertising and audience ecosystem; Usermaven is built around identifying people and connecting their behaviour to revenue. They count similar-sounding things using different rules, so a gap between them is expected. What matters is knowing which gap is normal and which gap means a setup problem.

This article explains how each tool counts, why the numbers diverge, and how to tell the difference.

The short version

Metric	Usermaven	GA4	Which is usually higher
Visitors / Users	Distinct anonymous IDs with activity in the date range	The "Users" metric in GA4 reports is Active users — users with an engaged session or a first session in the range	Usermaven
New visitors	Visitors with exactly one session inside the report window	Users whose first-ever session was in the range	Usermaven
Sessions	Grouped by anonymous ID, 30-minute inactivity break	Grouped by session ID, 30-minute inactivity break, timeout adjustable	Varies
Pageviews	Distinct pageview events received	`page_view` events received	Closest to equal
Bounce rate	Sessions with exactly one pageview ÷ sessions with at least one pageview	Inverse of engagement rate (10 s+, 2+ pages, or a key event)	Not comparable — different metrics

Start with pageviews. It is the most directly comparable metric in both tools, so it is the right place to begin a reconciliation.

How Usermaven counts

Sessions

A session begins with a visitor's first event and continues until 30 minutes pass with no activity from that same anonymous visitor. The next event after that gap starts a new session.

Two consequences worth knowing:

Sessions are not reset at midnight. A visitor active from 23:50 to 00:20 is one session, not two — so a single session can appear on two calendar dates.
Sessions are tied to the anonymous visitor ID, not to a campaign. If the same visitor arrives again from a different source within 30 minutes, Usermaven keeps one session and records the entry source for it.

Usermaven counts a session even if it contains no pageview — for example, a session built entirely from custom events or a server-side call. GA4 does not have an equivalent.

Visitors, new visitors, and returning visitors

Visitors = the distinct count of anonymous visitor IDs with activity in the selected range.
New visitors = visitors with exactly one session in the selected range.
Returning visitors = visitors with more than one session in the selected range.

This is a deliberately simple, window-relative definition. A person who first visited two years ago and comes back today is a new visitor in a today-only report, because they have one session in that window. GA4 would call the same person a returning user.

Bounce rate

Usermaven's bounce rate is based purely on pageview count:

Sessions with exactly one pageview, divided by sessions with at least one pageview.

So a visitor who reads a long article for four minutes and leaves is a bounce in Usermaven. GA4 would count that same visit as engaged, because it lasted more than 10 seconds.

This is one of the most predictable structural differences between the two tools, and it is purely definitional. It is not a data quality problem.

Traffic filtering

By default Usermaven excludes recognised bots. Your workspace can also exclude:

AI crawlers — by category
Proxy and anonymised traffic — VPN, Tor, datacenter, public proxy, web proxy, and session-based proxies

Each of these is a switch you control, and turning more of them on lowers your numbers relative to GA4, which is generally less aggressive here.

One caveat before you enable VPN exclusion: at the network level a legitimate corporate VPN is indistinguishable from a privacy VPN, so excluding VPN traffic will also remove real employees and remote workers. Those same visitors may still be counted by GA4, which widens the gap — often with no actual fault on either side.

Events

The Events metric counts autocaptured interactions plus your custom events. Internal system events and identity calls are excluded, so this number is not a raw event count.

How GA4 counts

GA4's definitions come from Google's own documentation on sessions and user metrics:

A session starts when someone views a page or screen with no active session. It ends after 30 minutes of inactivity by default — but this timeout is configurable, up to 7 hours 55 minutes. A session is not restarted at midnight, and not restarted when the campaign or traffic source changes mid-session.
"Users" in GA4 reports means Active users: people who had an engaged session or whose first session fell in the range. This is almost always the number you are comparing against. GA4 also exposes Total users, which counts everyone with a recorded event — including visitors who never had an engaged session. Total users therefore sits closer to Usermaven's Visitors than Active users does, so check which one you are looking at before comparing.
An engaged session is one lasting 10+ seconds, containing 2+ pageviews, or containing a conversion event.
Bounce rate is the inverse of engagement rate — the share of sessions that were not engaged.
GA4 identifies users using its own signals (client ID, User ID, Google Signals). With Consent Mode, GA4 may model missing data — but only when the property is eligible and configured for it.

Why the numbers diverge — the causes, ranked

1. The user definitions are genuinely different

Usermaven's Visitors counts anyone with a recorded anonymous ID in the range. GA4's headline Users metric counts only active users — someone who bounced after a single 3-second pageview may be present in GA4's event data but excluded from its active-user count. If you compare against GA4's Total users instead, the two come much closer.

Typical effect: Usermaven higher, because every recorded visitor counts — including those with no engagement at all. How large the gap is depends on how much shallow or bounced traffic your site attracts, and on whether you are comparing against Active users or Total users.

2. "New" means different things

Usermaven's new/returning split is relative to your selected date range. GA4's new users are first-ever visitors, regardless of range.

Typical effect: Usermaven new visitors noticeably higher, especially on short ranges. On a single-day report, almost every Usermaven visitor can be "new."

3. Bounce rate is not the same metric

As described above, Usermaven bounce = one-pageview sessions. GA4 bounce = not-engaged sessions. These will rarely be close, and the direction of the gap depends on your content: a four-minute single-pageview read is a bounce in Usermaven and an engaged session in GA4, while a fast two-page click-through is engaged in GA4 and not a bounce in Usermaven.

Do not treat a bounce-rate gap as an error. Compare like with like, or don't compare it at all.

4. Session boundaries mostly agree

This is the area where the two tools are closer than people expect. Both use a 30-minute inactivity gap, neither restarts a session at midnight, and neither starts a new session when the campaign or source changes mid-session. The main structural difference is that GA4's timeout is adjustable:

	Usermaven	GA4
Inactivity timeout	30 minutes, fixed	30 minutes by default, adjustable up to 7h55m
New session on campaign/source change	No	No
Reset at midnight	No	No
Session with no pageview	Counted	Not applicable

So if your session counts differ, the first thing to check is whether your GA4 session timeout has been customised away from 30 minutes. If you are comparing against historic reports from Universal Analytics, note that UA did start a new session on a campaign change — GA4 does not, so UA comparisons will inflate sessions on the GA4 side.

5. Bot and proxy filtering is unequal

Usermaven filters bots out of analytics by default, and lets you also remove VPN, Tor, datacenter, and AI-crawler traffic. GA4 filters known bots but is generally more permissive with proxy and AI traffic.

Typical effect: Usermaven lower, sometimes substantially — AI crawler traffic in particular can be a large share of raw requests on content-heavy sites.

Both tools lose events when consent is withheld or a script is blocked — but they don't lose the same events, and they don't respond the same way:

GA4 with Consent Mode may model missing sessions and conversions — but only when the property is eligible and correctly configured. Google's behavioral modeling requires Consent Mode properly implemented, plus a sustained volume of denied-consent events (on the order of 1,000 events per day for at least seven days), along with other eligibility criteria. Plenty of properties never qualify — in which case GA4 does not model anything either.
Usermaven does not model anything. It reports only what it actually received, so its numbers are literal.

Typical effect: GA4 higher on consent-restricted traffic, but only where modeling is genuinely active. If GA4 is ahead on every metric, check which tags your cookie-consent manager is blocking, and whether it is interfering with Usermaven's cookie specifically.

7. Your tracking configuration differs per tool

Common asymmetries:

Usermaven installed via Google Tag Manager in a container that fires on fewer pages than GA4's.
Duplicate Usermaven snippets on the same page, inflating pageviews.
Domains, subdomains, or staging environments tracked in one tool but not the other.
GA4 collecting from an app SDK that Usermaven isn't installed on.
SPA route changes captured by GA4's enhanced measurement but not firing a Usermaven pageview — or vice versa.

Pageviews that don't roughly agree almost always point to one of these.

8. Timezone and date-range boundaries

Usermaven reports in your workspace timezone; GA4 reports in the property timezone. If those differ, events land on different calendar dates and single-day or "today" comparisons shift by hours.

Also confirm both tools use the same inclusive/exclusive date convention, and that neither is applying a stale saved filter.

9. GA4 may threshold or hide rows

GA4 applies data thresholding to protect privacy, which can hide rows in reports and explorations — and it may blend in modelled data. Usermaven reports the stored value with no suppression.

Typical effect: GA4 shows less detail in some breakdowns, or totals that don't reconcile with its own rows.

10. Attribution taxonomies differ

Channel groupings, source/medium normalisation, and lookback windows are defined differently. Direct vs. referral classification in particular is a frequent source of "why is GA4 saying direct?" discrepancies in both directions.

Diagnostic checklist

Work through this in order — it resolves most cases quickly.

Match the timezone. Set both tools to the same timezone before comparing anything.
Use one complete past day. Today's partial data and event delays make live comparison meaningless.
Compare pageviews first. This isolates a tracking failure from a definitional difference.
Check for duplicate installation. Two copies of the Usermaven snippet is one of the most common causes of inflated pageviews.
Check your bot and proxy toggles. Note exactly what you have excluded in each tool.
Check consent setup. Confirm which tools are blocked when consent is refused.
Then compare sessions, and finally visitors. Each step narrows the cause.
Ignore bounce rate in the comparison. It is a different metric with the same name.

When a gap is actually a problem

Some gaps are expected. These are not — investigate them:

Symptom	Likely cause
One tool records nothing for a page or section	Tracking not installed there, or blocked
Pageviews differ by more than a few percent on a full past day	Configuration, duplicate tags, or SPA routing
Visitors far higher than GA4 users while pageviews agree	Visitor splitting — a consent manager deleting the cookie, or a cross-domain/subdomain cookie problem
Sessions much higher than visitors	Duplicate script, or events from multiple installs
Numbers changed suddenly on one date	A deploy, a consent-manager change, or a filter added

If visitors are inflated while pageviews look correct, that is the classic signature of the same visitor being counted repeatedly. The usual reason is a cookie consent manager deleting Usermaven's cookie — see Troubleshoot a cookie consent manager blocking Usermaven.

The practical rule

Trust each tool for what it is built to measure.

Use GA4 to reconcile with Google Ads and to align with Google's advertising ecosystem.
Use Usermaven to understand individual visitors, journeys, and revenue attribution — because it is built on identifiable people rather than aggregate sessions.

The thing to look for is not equality but consistency. A stable gap across periods is usually just two different definitions doing their job. A gap that suddenly changes size, or pageviews that stop tracking each other, is the signal worth investigating — see the table above.

We deliberately do not publish a "normal" percentage gap here. The right number depends on your traffic mix, how much of it is shallow or consent-restricted, and your filtering settings — so treat direction and stability as the guidance, not a fixed percentage.

Was this article helpful?