Content updated and sources checked on 19 October 2026. The original publication date has been retained.
More visitors does not automatically mean a better website. If you want to improve with GA4, start with a question: where do suitable visitors get stuck, and which action shows that the website helps? A measurement plan makes a dashboard more useful than a long list of isolated figures.
1. Choose a meaningful outcome
For a service provider, a successfully submitted enquiry may be important. For an online shop, the focus might be purchases and returns. Distinguish between clicking a button and actually completing the task. Establish which system confirms the final result so that a failed submission is not counted as an enquiry.
2. Understand bounce rate in GA4
In GA4, an engaged session is a session that lasts longer than ten seconds, has a key event or includes at least two page or screen views. Bounce rate is the proportion of sessions that are not engaged.[11] It is therefore not simply the percentage of visitors who leave after one page. Someone can read one article and still have an engaged session.
3. Check events before drawing conclusions
Go through a form yourself and check that the intended event appears. Also test a validation error, reloading and double-clicking. Only mark events as important when they genuinely say something about your goal. Record changes to measurement: a sudden increase may be a measurement error rather than a better result.
4. Compare groups with the same context
For example, examine the contact journey of mobile visitors separately from desktop visitors. Compare campaigns in the context of their own offer and landing page. Account for seasons, budget changes and small numbers. Use the figures to formulate a question and combine them with actually going through the page.
5. Measure online shop steps explicitly
Google documents separate ecommerce events for actions including product views, shopping basket actions, starting checkout, purchases and refunds.[12] General page tracking alone does not automatically report these steps correctly. Have the events and associated product and transaction data tested before assessing a sales funnel.
6. Account for consent in your interpretation
Consent Mode adjusts the behaviour of supported Google tags to consent choices; basic and advanced mode work differently.[10] Changes to that setup can affect your measurements. Do not use analytics as though every visitor is fully visible. Record when the banner, tags or settings change.
7. Going further: work with one improvement hypothesis
For example, formulate this hypothesis: visitors hesitate at the form because it is unclear what happens after submission. Add a short explanation and record what you want to assess afterwards. Do not change every field, the campaign and the offer at the same time. Choose an evaluation point based on sufficient data and business context; there is no universal measurement period for every website.
Frequently asked question: does GA4 prove that a campaign is profitable?
No. Compare recorded actions with suitable enquiries, orders, costs and any returns in your own records. Revenue is not the same as profit. Want your measurement plan to align with your campaigns and website? Discuss your measurement questions with Digital Pandas.
