Google Analytics for Marketing Mentors: A Decision Guide

Which question should you bring to GA4?

Choose the decision before choosing the report.

A useful question names a choice. 'Should we revise the consultation page or investigate the traffic reaching it?' is actionable. 'How is the website doing?' leaves you with an open-ended collection of charts. Write down what evidence would change your advice before opening the numbers.

GA4 has different acquisition scopes. Google's traffic-source scope documentation, checked in September 2026, distinguishes first-user acquisition from session acquisition and event attribution. Use session-scoped dimensions for the source of visits in the period. First-user dimensions answer how users were originally acquired, which is a different question.

Keep the report name and dimensions in your notes. A chart titled 'organic performance' is not reproducible if one month uses first-user source and the next uses session source. Good mentoring analytics preserves the question and measurement context, not just the result.

Set access and a comparable baseline

Read-only advice should not require ownership of the client's analytics account.

Ask the client to retain administration of their property and grant access to your own account. Google's access and data-restriction guidance, checked in September 2026, describes a Viewer role for seeing data and working with reports, alongside more powerful roles. Start with the least access that supports the agreed work; request broader permissions only for a specific authorised task.

Confirm the property identifier, website and included data streams together. Record any restrictions on revenue or cost data so a missing value is not misread as zero. Never ask a client to send their password, and agree who removes your access when the engagement ends. Screenshots circulated to stakeholders should contain only the information those recipients need.

Your coaching baseline needs a measurement record: date range, property time zone, traffic filter, selected event and extraction date. Use equal-length periods with comparable weekdays when that suits the decision. Add a year-on-year comparison where seasonal demand matters and history is available; neither comparison eliminates other influences.

Google's data freshness guidance, reviewed in September 2026, says processing can take 24 to 48 hours and reports can change during that time. Record the extraction time and avoid presenting an incomplete current day as a finished comparison period. For attribution-sensitive decisions, also check whether recent credit is still being updated.

Use a compact measurement record

The denominator is part of the metric, not a footnote.

Choose a small set of measures whose interpretation you can explain. Google's engagement-rate documentation, checked in September 2026, defines engagement rate as engaged sessions divided by sessions. Under the default time threshold, an engaged session lasts longer than 10 seconds, has a qualifying key event or has at least 2 page or screen views; check the property's actual configuration.

Engagement rate is therefore not a satisfaction score. A session can meet the rule without producing a useful enquiry, while a short visit can still answer the visitor's question. Describe the metric as GA4 defines it and use other evidence to judge whether the business objective was met.

Google's key event guidance, checked in September 2026, explains how selected important events appear in reporting. For mentoring, pick the event that matches the decision rather than silently using all key events. A newsletter signup, form submission and purchase should not be treated as interchangeable commercial outcomes.

Inspect the report's quality indicators. Google's data-threshold documentation, reviewed in September 2026, describes situations where data is withheld to protect privacy. Missing or withheld detail is not evidence that no activity occurred. Record the limitation and choose a less granular view when appropriate.

What does a worked comparison tell you?

Traffic growth and a lower engagement rate can occur in the same useful reporting period.

The following numbers are entirely illustrative, not client results or a benchmark. Assume the same property, equal completed periods, unchanged tracking and one organic traffic segment. The selected event is a tested consultation-form submission; sessions with that event are counted separately from the number of event occurrences.

The lower engagement rate is not sufficient reason to rewrite the entire site. More sessions included a consultation submission, and their share also increased. The useful next check is whether the submissions are valid and commercially relevant, followed by a page or device breakdown that can identify a specific test.

The example does not establish statistical significance or a causal effect. Traffic composition, an offer change or external demand could explain the movement. Report both counts and rates; 2% to 2.5% is a 0.5 percentage-point increase, while its relative increase is 25%. Choose the wording least likely to exaggerate the practical scale.

Turn the reading into one accepted next step

Analytics becomes useful when someone accepts responsibility for the next check.

Finish the session with a small decision record. In the illustrative case, the client could agree to reconcile the 37 event occurrences with form records and report duplicate or invalid submissions before the next review. The mentor owns interpretation and the review question; the client or authorised analyst owns access to the underlying enquiry details.

Do not put personally identifiable enquiry data into an analytics report simply to make it more persuasive. Use an authorised, appropriately limited reconciliation process. If the outcome event is unreliable, repair measurement before recommending a larger spend or declaring the new page successful.

Place the decision, caveat and next evidence date in the coaching progress report. MentPass can display connected, read-only client analytics alongside the work being discussed; use the source account for checks beyond the available summary. For a financial value claim, follow the separate coaching ROI guide rather than renaming traffic growth as return on investment.