Home / Blog / Google Analytics 4: a complete guide for 2026

Analytics

Google Analytics 4: a complete guide for 2026

Joonas Harimaa Joonas HarimaaFounder · updated · 10 min read

Google Analytics 4 is more capable than anything that preceded it and considerably less forgiving. The event-based model is genuinely better suited to how people use websites now, but the defaults answer almost none of the questions a business actually has.

This guide covers what to configure, what to ignore, and how to get from data to decisions.

What GA4 changed

The previous model counted sessions and pageviews and bolted goals on top. GA4 counts events — every interaction is an event, and sessions are derived from them.

The practical consequences:

  • Everything is an event. Page views, scrolls, clicks, form submissions. Flexible, and it means nothing is measured meaningfully until you decide what matters.
  • Cross-platform by default. Web and app in one property.
  • Modelled data. With consent mode, GA4 estimates behaviour it could not observe. Useful, and it means figures are estimates rather than counts.
  • Different reporting. Exploration reports replace the old custom reports, with a steeper learning curve and more power.

The setup that actually matters

1. Define conversions properly

The most consequential decision. Pick three to five events that represent genuine commercial intent: a contact form submission, a call click, a demo request, a meaningful tool completion.

Do not mark newsletter sign-ups, scroll depth and PDF downloads as conversions alongside them. Mixing intent levels makes the conversion number meaningless, and every downstream report inherits the problem.

2. Get the events right

Enhanced measurement captures scrolls, outbound clicks, site search and file downloads automatically. Beyond that, configure events for the interactions specific to your business — and name them consistently, because renaming later does not backfill history.

3. Filter internal traffic

Your own team's visits distort small-site data badly. Filter by IP or, better, by a marker your team's browsers carry. This is a five-minute job that most properties have not done.

4. Connect Search Console

Without this link, GA4 cannot tell you which queries brought organic traffic — which is most of what you want to know about organic. It is a two-minute connection and it is frequently missing.

5. Configure consent mode

Under UK GDPR and PECR, analytics cookies need consent. Consent mode lets GA4 behave lawfully before consent is given and model the gap afterwards. Get the legal position confirmed rather than assuming the default is compliant.

6. Set data retention deliberately

The default is short. If you want year-on-year comparisons, extend it — and note that the change is not retrospective, so doing it late loses history permanently.

The metrics worth watching for B2B

MetricWhy it mattersTrap
Conversions by channelWhich sources produce commercial intentLast-click attribution undervalues research-phase channels
Landing page conversionsWhich pages earn their existenceLow-volume pages look unreliable; judge over quarters
Engaged sessionsBetter proxy for interest than bounce rate ever wasStill a proxy; do not optimise it directly
Organic queries (via Search Console)What people actually searched to arriveHeavily sampled and truncated for long-tail terms
Returning usersB2B buyers research over weeksCookie loss inflates apparent new users

Mistakes that corrupt the data

  • Every event marked a conversion. The most common, and it makes the whole property uninformative.
  • No internal traffic filter. Especially damaging on low-volume B2B sites.
  • Inconsistent event naming. form_submit, formSubmit and contact_form counted separately, forever.
  • Search Console never linked. Half the organic picture missing.
  • Reading absolute numbers as truth. They are modelled estimates. Trends are trustworthy; the third significant figure is not.
  • Reporting that nobody acts on. A dashboard that produces no decisions is a cost, not an asset.

Building reports that drive decisions

Start from the question, not from the tool. For most businesses three recurring questions cover it:

  1. Which channels produce customers, not just sessions? Conversions by channel, compared against the previous period.
  2. Which pages earn their keep? Landing page report filtered to organic, sorted by conversions.
  3. What changed, and why? Movement in organic queries and positions from the Search Console integration.

Build one exploration for each, save them, and review monthly. Three reports somebody reads beats twenty nobody opens.

Connecting analytics to sales

The gap between "GA4 says 40 conversions" and "sales closed three deals" is where most measurement programmes stall.

Two practical steps close most of it. Pass a source parameter into your CRM when a form is submitted, so the record carries its origin. And reconcile monthly — compare GA4 conversions against CRM records created, and investigate the difference rather than ignoring it.

Once that link exists, the return calculation in our SEO pricing guide becomes something you can actually compute.

Frequently asked questions

Why do GA4 numbers not match our CRM?
They are counting different things. GA4 counts sessions and events with a modelling layer on top; your CRM counts records created by humans. Expect a gap and reconcile the trend rather than the absolute figure.
Do we still need cookie consent for GA4?
Yes. Under UK GDPR and PECR, analytics cookies require consent, and GA4 should be configured with consent mode so that behaviour before consent is handled lawfully. This is a legal question — take advice rather than relying on defaults.
What is the single most common setup mistake?
Not defining conversions, or defining every event as one. If everything is a conversion, nothing is measurable. Pick the three to five events that represent real commercial intent.
Should we use GA4 or something else?
GA4 is free, ubiquitous and integrates with Search Console and Google Ads, which is usually decisive. Privacy-focused alternatives are simpler and often more accurate on raw counts, but give up the integrations.

Measuring the right things yet?

The audit checks whether your organic traffic is being attributed correctly — and what it is actually worth.