The basic model
In plain English: A user has sessions; sessions contain events; events carry parameters.
A page view, scroll or sign-up can be recorded as an event. Parameters add context such as the page location or sign-up method. GA4 then exposes descriptive dimensions and numerical metrics for reporting.
Web and app data streams can feed one property, but only data you deliberately collect and configure can answer your questions.
- event: sign_up · parameter: method = email
- dimension: Country · metric: Active users
What you see in GA4
In plain English: Reports answer common questions; Explorations support more flexible analysis.
Realtime is useful for recent activity and initial checks. Standard reports cover acquisition, engagement and content. Explorations let you combine dimensions, metrics, filters and segments for deeper investigation.
- Acquisition: where visits came from
- Engagement: what happened after arrival
- Pages and screens: which content was viewed
What GA4 is not
In plain English: Measurement is an estimate shaped by implementation, consent and identity choices.
GA4 is not your server log, not Google Search Console, and not guaranteed person-level identification. Browser restrictions, consent choices, blockers, devices and configuration can all change the totals.
Search Console describes visibility and clicks in Google Search. GA4 describes measured behaviour on or around your site after visits; their datasets should not be expected to match exactly.
Keep in mind: Do not treat a GA4 user count as an exact count of individual humans.
Start with a question
In plain English: Collect and mark data because it answers a real product question.
Begin with outcomes such as a completed enquiry, account creation or purchase. Confirm the relevant events and parameters, mark genuinely important events as key events, and verify the implementation before interpreting trends.