How to Build a Web Analytics Report: GA4 Metrics and Action Plans
August 4, 2026
Author: Shusaku Yosa
A useful web analytics report helps its reader decide what to do next. In GA4, that requires clear distinctions between users, sessions and events, plus an explicit denominator for each rate. Use a short decision summary, matched comparisons and a worked example to keep observations separate from explanations.
Start with the reader and the decision
Write down who will read the report and what decision they need to make. An executive may need an investment decision, an editor may need a page to investigate, and an implementer may need reproducible measurement conditions. That choice determines the level of detail and which metrics earn a place.
For example, ask: “Can we verify the journey from an organic landing session to a successful application?” Traffic volume alone cannot answer this. If success has not been validated, report it as unmeasured rather than converting missing evidence into zero outcomes.
Summarize acquisition, behavior and verified outcomes
Summary field | Content | Essential context |
|---|---|---|
Question and scope | URLs, dates, landing-page and channel conditions | Production host, language and comparable days |
Acquisition | Sessions plus search impressions and clicks | Keep GA4 and Search Console series separate |
Behavior | Time, views and subsequent actions | Name the denominator and measurement conditions |
Outcomes | Verified success and observed event counts | Distinguish people, occurrences and successful sessions |
Next action | Hypothesis, owner and deadline | Evidence needed for the decision |
Use a line chart for a trend and a table for page comparisons. Segment by relevant languages, devices or entry pages, but retain actual counts when splitting creates small samples. The summary should lead to a decision without requiring the reader to interpret every exported column.
Define GA4 metrics, denominators and scopes
Consult the GA4 metric definitions before using shorthand such as “users” or “CVR.” Also distinguish traffic-source scopes: first-user acquisition, session acquisition and event attribution answer different questions.
Metric | Meaning | Common misreading |
|---|---|---|
Active users / Total users | Different measured-user definitions | Actual people, or additive page-row totals |
Sessions | Sessions on the site or app | Sessions minus users equals returning people |
Views | Page and screen view occurrences | Unique viewers |
Key events | Recorded key-event occurrences | Completed registrations or distinct people |
Session key event rate | Sessions with a key event ÷ all sessions | Key-event count ÷ sessions |
Average time | Total time ÷ the relevant denominator | Proof of reading or satisfaction |
GA4 bounce rate exists and is 100% minus engagement rate for the same selection. The official engagement criteria use a session lasting longer than the default ten seconds, an eligible key event, or at least two page or screen views. Check the configured timer. Do not import Universal Analytics’ single-page-visit interpretation.
Match periods, hosts, landing pages and channels
Match day counts, URLs, language, production hostname, landing-page definition and session channel conditions. Sessions that viewed a page are not the same set as sessions that began on it. A first-user source filter cannot substitute for a session acquisition filter.
Search Console performance describes search impressions and clicks, while GA4 describes site-side activity such as sessions. Their difference is not automatically a measurement defect or a count of people lost. Check dates, time zones, consent, tag receipt and filters. Document query anonymization and aggregation, row limits and capture times.
Calculate counts and session rates with a worked example
This fictional example has 100 sessions, 20 occurrences of a successful-action event designated as a key event and 12 sessions containing that event. Report the event count as 20 and the session key event rate as 12 ÷ 100 = 12%. Dividing 20 events by 100 sessions produces a different measure, not the session key event rate.
Selection | Sessions | Sessions with the event | Session rate |
|---|---|---|---|
A | 100 | 12 | 12% |
B | 300 | 18 | 6% |
Non-overlapping A + B | 400 | 30 | 7.5% |
For non-overlapping groups, the combined rate is 30 ÷ 400 = 7.5%, not the simple average of 12% and 6%, which is 9%. Overlapping selections require deduplication first. Whether the event truly represents success, whether it is duplicated, and how a person is identified are separate validation questions.
Separate observations, hypotheses and next actions
A key-event report reflects the important actions your implementation records. Before calling them business outcomes, verify the event name, success condition, production host, duplicate handling, authenticated or cross-domain transitions, and consent. “20 events; completed registrations unmeasured” is more precise than an unsupported people count.
Part | Fictional entry |
|---|---|
Observation | The selected landing sessions have a lower successful-session rate |
Hypotheses | Intent changed, the form failed, or measurement missed activity |
Next task | An owner reproduces the production success flow and checks measurement |
Decision condition | Choose a page change after success and comparison conditions are validated |
A low rate does not establish that landing-page copy caused the problem. Focused engagement time is also a clue rather than proof of reading. A before-and-after improvement alone cannot establish the edit’s causal effect.
Check quality and document recurring updates
- Match target URLs, language, host and comparison days.
- Write the numerator and denominator for every rate.
- Separate event occurrences, successful sessions, measured users and verified people.
- Avoid adding page or daily users, or averaging percentages without their denominators.
- Disclose anonymization, limits, rounding, missing rows and unmeasured outcomes.
- Assign an owner, deadline and the evidence to inspect next.
For recurring reporting, save a metric dictionary and reproducible extraction conditions. Append measurement and website changes with dates, and explain revisions to historical comparisons. Consistent definitions let readers follow the same decision over time.
Continue with the guide that matches your next task: GA4 Average Engagement Time: Denominators, Examples and Interpretation; GA4 APIs Explained: Data, Admin, Realtime and BigQuery; Landing Page SEO: Match Search Intent, Improve Content, and Measure Results; What Is a Search Query? Keywords, Intent and Search Console Analysis.
