Analytics You Can Trust: Getting GA4 Right

Here’s an uncomfortable truth: most analytics setups are quietly lying to you. The dashboards look busy, the numbers go up and to the right, and somewhere a decision gets made on data that was never trustworthy to begin with. You can’t optimise what you can’t measure — and you definitely can’t measure with a broken GA4.

If you’re going to trust a single number to steer your marketing budget, that number has to be real. Let’s make it real.

Why Most Setups Quietly Lie

A default GA4 install is not the same as a correct one. Out of the box, GA4 counts your own team browsing the site, bots crawling pages, the developer testing the checkout fifty times, and traffic from staging environments. None of that is a customer. Yet it all lands in the same reports you use to judge whether a campaign worked.

The result is subtle but corrosive. Your conversion rate looks lower than it is. Your best channel looks worse because internal traffic inflated a weaker one. You “learn” the wrong lesson, double down, and waste money — all while feeling perfectly data-driven.

Bad data doesn’t feel like no data. It feels like confidence. That’s exactly what makes it dangerous.

What a Trustworthy Setup Actually Needs

A GA4 you can trust isn’t complicated — it’s just deliberate. The essentials:

• A clean install. One tag, firing once, on every page. No duplicate tags double-counting sessions, no half-migrated Universal Analytics leftovers.
• Correct key events. Mark the things that matter as key events — a form submit, a booking, a purchase, a qualified lead. Not every button click. A “conversion” that fires on page scroll tells you nothing.
• Filtered internal traffic and bots. Exclude your office IPs and your team’s traffic, and lean on GA4’s bot filtering so robots don’t pad your numbers.
• Search Console linked. Connect it so you can see the actual search queries bringing people in, not just “(not provided)”.
• Sensible data retention. Set event data retention to the longer option so you can compare year over year instead of losing your history.

Get those five right and you’ve already beaten the majority of setups out there.

Common Mistakes That Cost You

The usual suspects show up again and again. Tracking everything as a conversion, so nothing is meaningful. Never filtering internal traffic, so your own team is your “best audience”. Letting two analytics tags run at once after a redesign. Obsessing over vanity metrics — pageviews, raw sessions, time on page — while ignoring whether anyone actually did the thing your business depends on.

The fix is focus. Pick the three or four numbers that map to revenue, report on those, and let the rest sit quietly in the background. A clean weekly view of qualified leads beats a forty-widget dashboard nobody reads.

Measure First, Optimise Second

This is the part people skip. Conversion rate optimisation, A/B tests, landing-page tweaks — all of it assumes your measurement is sound. Run a test on top of lying analytics and you’ll confidently ship the losing variant. Strong SEO in 2026 has the same dependency: you can’t tell which content earns customers if you can’t see who converts.

Fix the foundation first. Once the numbers are honest, every optimisation decision after that compounds in your favour. That’s the whole point of our analytics and CRO work — get the measurement right, then make it pay.

Not sure whether your GA4 is telling you the truth? Request a free audit and we’ll show you exactly where the numbers are leaking.