User Acquisition vs. Traffic Acquisition

Valid as of April 2026, GA4 property interface.

Two GA4 reports live next to each other under Reports → Acquisition. Their names differ by one word, they use nearly identical column layouts, and they almost always return different numbers for the same date range. This is the most frequently asked disambiguation in GA4: which report should I be looking at?

The short answer: User Acquisition groups users by the source that first brought them to your site. Traffic Acquisition groups sessions by the source that started each session. They answer different questions and neither is “the right one” — the right one depends on what you want to know.

The one-sentence difference

Report	Groups by	Assigned to
User Acquisition	Users	Source of the user’s first session (first-touch)
Traffic Acquisition	Sessions	Source of each session (per-session, last non-direct click)

Both reports use last non-direct click attribution regardless of your property’s attribution settings. Changing the attribution model in Admin → Attribution settings does not change what these reports show. That setting only affects the Advertising section and the Data API.

Worked example: a single user, three visits

Consider one real user over the course of a week.

User Acquisition view

Monday:    Google / organic  (first visit)
Thursday:  (direct) / (none)
Saturday:  (direct) / (none)  — converts

User is attributed to:
→ google / organic

All events, revenue, and conversions
from this user roll up to
google / organic in User Acquisition
reports for weeks to come.

Traffic Acquisition view

Monday:    google / organic     → 1 session
Thursday:  (direct) / (none)    → 1 session
Saturday:  (direct) / (none)    → 1 session, 1 purchase

Traffic Acquisition splits the
three sessions across two source
rows:
→ google / organic: 1 session, 0 revenue
→ (direct) / (none): 2 sessions, $89 revenue

Same user. Same three visits. Same one purchase. In User Acquisition the purchase and all three sessions are credited to google / organic. In Traffic Acquisition the purchase is credited to (direct) / (none) because that was the source of the converting session.

This is not a bug and it is not a data quality problem. The reports are answering two different questions.

Which question are you asking?

Use User Acquisition when… You want to evaluate channels as user-acquisition investments. 'Where did my customers originally come from?' Marketing-mix and lifetime-value questions live here.

Use Traffic Acquisition when… You want to evaluate what is driving on-site activity right now. 'Which channel drove this week's sessions and conversions?' Campaign pacing and weekly reporting live here.

User Acquisition use cases

• Evaluating long-term channel performance — which channels bring users who convert eventually, even if they return directly later.
• LTV analysis — revenue per acquired user, segmented by the channel that acquired them.
• Content marketing and SEO ROI — users who first found you via organic search and later converted through any channel.
• Paid social justification — users who discovered the brand via a Meta campaign and purchased weeks later via direct traffic.

Traffic Acquisition use cases

• Weekly marketing dashboards — session volume by channel this week.
• Campaign pacing — is the new paid search campaign pulling its weight in sessions?
• Conversion-rate benchmarking per channel — conversion rate of sessions starting from email vs. paid vs. organic.
• Page-level and landing-page analysis — the Traffic Acquisition dimensions (session source, session medium, session campaign) are the ones joined into landing-page reports.

Tip

If your stakeholder asks “is our paid search working?” and you only have time for one report, open Traffic Acquisition. If they ask “are we growing our organic user base?”, open User Acquisition. The common mistake is using Traffic Acquisition for both and then wondering why the organic story looks flat even though SEO is doing well — organic’s downstream direct-return sessions are landing in the direct row.

The dimension names give it away

The two reports use different default primary dimensions, and once you learn the prefix convention you can tell them apart in any custom report or exploration.

Prefix	Meaning	Used in
First user …	First-touch attribution (first session ever)	User Acquisition
Session …	Session-scoped (the session that is being measured)	Traffic Acquisition, landing-page reports

Examples of the same concept across both scopes:

User Acquisition dimension	Traffic Acquisition dimension
First user source	Session source
First user medium	Session medium
First user campaign	Session campaign
First user source / medium	Session source / medium
First user default channel group	Session default channel group

In Explorations, you will see both families in the dimension picker. Choosing First user source against a session-scoped metric like Sessions produces a valid but often counterintuitive report: sessions from returning users are grouped under the channel that originally acquired them.

Metric scope interacts with dimension scope

The scope mismatch between dimension and metric is where many mis-reads originate.

Metric	Natural scope	Works with first-user dimensions?
Total users	User	Yes — native fit
New users	User	Yes — native fit
Sessions	Session	Works, but sessions are grouped by first-touch channel of the user
Engaged sessions	Session	Same caveat
Key events	Event	Groups events by the user’s first-touch channel
Total revenue	Event	Groups revenue by the user’s first-touch channel

Mixing scopes is not wrong — it is often what you want (e.g., “revenue per acquired user, by acquisition channel”) — but you need to read the result with the scope combination in mind.

Direct traffic inflation in Traffic Acquisition

Traffic Acquisition uses last non-direct click. This means a session with no campaign parameters and no referrer inherits the most recent non-direct source, up to the lookback window. So a user who arrived via google / organic on Monday and returns with no referrer on Thursday is credited to google / organic in Traffic Acquisition on Thursday — not to direct.

The direct row therefore only contains sessions where GA4 cannot find any prior non-direct source in the lookback window. Users are:

• Genuinely new, with no referrer (typed URL, bookmark, dark-social paste).
• Returning after their last non-direct source aged out of the 90-day lookback.
• Victims of lost campaign tagging — missing UTMs on a paid campaign, broken referrer policies, or unlinked Google Ads accounts.

If direct traffic spikes, check the date range for a tagging regression before assuming users suddenly memorized your URL.

How each report handles User ID and Google signals

User Acquisition depends on stable user identity. If a user clears cookies and returns, GA4 sees a new user_pseudo_id, which counts as a brand-new user and a brand-new first-touch channel. User Acquisition numbers therefore drift over time as cookie churn reshuffles which channel “acquired” each user.

User ID implementation mitigates this: logged-in users are stitched across devices and browsers, and their first-touch channel is preserved. Properties without User ID see more volatility in User Acquisition than in Traffic Acquisition, because Traffic Acquisition re-derives the source from each session’s own signals.

Differences from Universal Analytics

Universal Analytics had a single Acquisition → All Traffic → Source/Medium report that used last non-direct click at the session level — closest in behavior to GA4’s Traffic Acquisition. UA’s Users metric in that report was “users who had at least one session from this source,” not “users whose first session came from this source.”

GA4’s split into two reports is the genuine improvement here: you can now ask first-touch and session-touch questions without leaving the standard report set. The tradeoff is that stakeholders used to the UA number sometimes read one report and think it should equal the other.

Validating the difference in BigQuery

The split is easy to replicate in BigQuery. Each event row carries two families of traffic fields:

• traffic_source.source, traffic_source.medium, traffic_source.name — first-touch (User Acquisition analog).
• collected_traffic_source.* — the collected source at the time of the event, which drives session-scoped Traffic Acquisition once resolved through last non-direct click.

For session-level work, group by user_pseudo_id + ga_session_id and pick the source from the session’s starting event. For user-level first-touch, group by user_pseudo_id and pick the minimum-timestamp session’s source, or use the traffic_source struct directly. See Common BigQuery Queries for worked patterns.

Common mistakes

Comparing the two reports’ totals and calling it a discrepancy

The total users in User Acquisition and the total users in Traffic Acquisition are usually the same (both count distinct users who had any session in the range). But the per-channel breakdowns will never match, because the grouping dimensions are different. “Organic search in User Acquisition says 1,200 users, but in Traffic Acquisition it says 800” is not a problem — it is the expected shape of the data.

Using Traffic Acquisition for cohort analysis

Cohorts are a user-scoped concept. Building a retention or LTV cohort using Session source / medium as the cohort dimension produces results that re-shuffle users into channels each session. Use First user source / medium for cohort and retention analysis.

Expecting attribution-model changes to affect these reports

Switching your property attribution from last click to data-driven does not change either of these reports. Both always use last non-direct click at their respective scopes. To see data-driven attribution, open the Advertising → Attribution reports.

Forgetting the 90-day lookback

A user whose last non-direct source was 100 days ago shows up as direct in Traffic Acquisition. This is invisible unless you think about the lookback window.

Related articles

Sessions and Attribution How GA4 defines sessions and applies attribution — the foundation for both reports.

Data Model Why everything is an event and how first-touch and session-touch are derived.

Data Discrepancies Other cases where GA4 numbers do not line up with expectations.

Common BigQuery Queries Rebuild both acquisition views in SQL.