How to Use GA4: Traffic Sources, Pages and Business Outcomes

April 8, 2026

Author: Shusaku Yosa

GA4の使い方を初心者向けに完全ガイド｜Googleアナリティクスの基本操作から活用法まで

Start using GA4 with three questions: where did visitors come from, what did they read, and did the visit produce the intended outcome? You do not need to master every menu before making a useful decision. This guide assumes collection is already working.

Confirm the property, reporting period and web/app scope. Avoid drawing conclusions from incomplete recent processing, and consider weekday composition and seasonality when choosing comparison periods.

1. Investigate traffic sources

Use Traffic acquisition for the source of sessions. User acquisition describes the source that first acquired the user. Navigation can vary by the business-objective collection configured for the property; use report search or the available library if the expected item is missing.

Question	Dimension scope	Supporting measures
Where did this visit originate?	Session source/medium	Sessions, selected outcomes and rates
Where was the user first acquired?	First user source/medium	New users and subsequent behavior
Which campaign generated visits?	Appropriate campaign dimension	Consistent UTM naming and dates

Direct traffic does not prove that everyone typed the URL. Investigate missing referral context and untagged campaigns. Use consistent UTMs for external email or campaign links, but do not add them to ordinary internal related-article links and confuse acquisition reporting.

2. Separate viewed pages from entry pages

Pages and screens describes content viewed during visits. Landing page describes where a session began. An article read later in the journey is not necessarily the acquisition entry point.

Question	Report	Interpretation
Which content was viewed?	Pages and screens	Views are not people
Which page started visits?	Landing page	Compare session entry points
Where did visitors go next?	Path exploration or an appropriate exploration	Keep population and path conditions consistent

For multilingual sites, separate language paths. Check for different URLs sharing the same title and for query parameters fragmenting the comparison. GA4 bounce rate exists, but it is not a completion-of-reading metric. A short satisfied visit and an unsuitable landing can look similar without further context.

3. Select the business outcome

An all-key-events total can combine purchases, registrations and intermediate actions. Select the outcome relevant to the decision, then show both counts and rates. GA4 cannot establish a completed business process that has not been implemented as an appropriate event.

Hypothetical entry page	Sessions	Sessions with registration	Registration session rate
Article A	1,000	20	2%
Article B	100	5	5%

B has the higher rate while A contributes more registering sessions. Scaling B will not necessarily preserve its rate. Compare search intent, acquisition source and customer fit before choosing between additional traffic and conversion-path work.

A weekly 30-minute review

Check collection health and measurement changes first. Review major traffic sources, then high-volume entry pages. Break meaningful differences down by device or language. Choose one hypothesis and record the owner, action, outcome metric and next review date.

If views rise while outcomes stay flat, investigate whether new visitors are less suitable or the application path is failing. If outcomes suddenly disappear, inspect collection and the destination journey before changing acquisition strategy.

When reports disagree

Advertising platforms and GA4 can use different windows, attribution rules, consent observations, timezones and scopes. The objective is to understand which measure serves which decision, not force every total to match. Read data-quality and threshold notices when present.

Bring four items to the review meeting: observed change, plausible explanation, proposed action and decision criterion. This keeps reporting connected to a choice instead of turning it into a tour of every available metric.