Why GA4 Is Installed Everywhere but Rarely Used

GA4 is present on almost every e-commerce site. In practice, most stores limit their use of it to checking user counts and total revenue — and nothing more. The rest of the data sits untouched.

This isn't about lack of motivation. It comes down to two things: GA4 is configured incorrectly (making the data unreliable), and nobody in the organisation knows where to start with analysis. The result is like a sofa wrapped in plastic film — the equipment is there, but not for use.

The cost of inaction is real. A poorly configured GA4 means weaker ad algorithms, because Google and Meta receive fewer conversion signals. If your competitor delivers 20% more purchase data to their ad platforms, their algorithms are 20% better trained. Every advertising euro works less efficiently as a result.

What "Correct GA4 Configuration" Actually Means

Correct configuration is not the same as installing the tracking script. It has several layers, each with its own failure points.

Consent banner for cookies. The most common mistake: GA4 fires before the user has a chance to interact with the consent banner. This is a GDPR violation — fines are calculated based on revenue, not profit, and can be severe. The correct setup means the GA4 script fires only after the user clicks "accept" or "decline." When a user declines, GA4 can still run in anonymous mode — it sends data but without user identifiers. Google then models this data to partially fill the gaps.

The typical effect of a misconfigured banner: the user lands on the page, clicks consent, nothing happens (because the script already fired without consent, or failed to fire at all), they navigate to the next page and only then does GA4 initialise. Result: corrupted data on traffic source and the user's entire journey.

UTM parameters — convention and consistency. Every link driving traffic to your site — from newsletters, paid campaigns, LinkedIn posts, offline QR codes — should be tagged with UTM parameters: medium, source, and campaign. This lets you see not just how many people clicked a link, but how many of them actually purchased.

One critical detail: UTM parameters are case-sensitive. "Newsletter" and "newsletter" are two different traffic sources in GA4. If several people on your team create links without an agreed standard, your analytics breaks down. The fix is simple: one shared spreadsheet or UTM generator with a defined naming convention for each channel.

Data layer and Google Tag Manager. GTM is not technically mandatory — there are plugins that work. But GTM gives you control. With a GTM implementation you can specify precisely when and which events are sent, what parameters accompany them, and distribute the same data simultaneously to GA4, Google Ads, and Meta. This matters especially for forms — automatic form tracking in GA4 without a data layer almost never works correctly.

Can GA4 Ever Show 100% of Your Data?

No. And that's not the goal. There are several reasons why there will always be a gap between GA4 and your store's backend data.

Not all users accept cookies. Ad blockers block analytics scripts. Browser scripts sometimes fail to fire correctly. Users who complete a purchase in a payment gateway and close the browser before returning to the confirmation page won't be tracked — unless server-side tracking is in place.

A practical rule: up to 10% discrepancy is an excellent result. 20% is normal and acceptable. 25% or more — start looking for the cause. What matters is that assuming the data loss is proportional across channels, comparisons between channels and campaigns remain valid. You're analysing trends, not absolute numbers.

On Shopify specifically, discrepancies tend to be larger than on other platforms — it's a closed environment with limited configuration flexibility. Apps like Little Data help better synchronise GA4 data with store data.

What to Measure: Events That Actually Matter

GA4 automatically collects several events as soon as the script is installed: sessions, page views, YouTube video interactions, and internal site search. That last one is rarely used — there's no ready-made report for it, you have to build one in Explorations. It's worth doing. Internal search data shows what users are looking for on your site: what's missing from your offer, which colours and sizes they want, which brands you don't carry.

Beyond automatic events, the recommended e-commerce events to configure include: product view, add to cart, checkout initiation, purchase. Each event should carry parameters — colour, size, payment method — everything the user selected. A clothing store can then see that the green version of a jacket is clicked five times more often than the black, and use that colour in ad creatives. The same data can inform stock ordering decisions — the right size and colour ratios based on actual demand, not gut feeling.

One important note: don't measure everything. We've seen cases where someone wanted to track every button click on the entire site. The result is a sea of data nobody ever analyses. Add an event when you know why you need it.

Two Reports Whose Difference Is Worth Understanding

GA4 has two reports with similar names that measure two entirely different things.

Traffic acquisition shows the source of each individual session. If the same user visited eight times — from SEO, from Google Ads, from Meta — each visit is attributed to its own source.

User acquisition shows where a given user came from for the very first time. If they first arrived via a Facebook ad but made their purchase on their eighth visit from Google, that purchase is attributed to Facebook.

User acquisition is the key report for understanding which channels are actually bringing new customers into your funnel. Traffic acquisition tells you what closes conversions at the bottom of the funnel. Both are necessary — but confusing them leads to wrong budget decisions.

Segmentation: Why Aggregate Data Misleads You

The biggest mistake in GA4 analysis: looking at aggregate numbers without segmentation. Total conversions and total revenue don't tell you whether you're acquiring the right customers.

Example: a store has an average order value of €250. Create two segments — customers buying above €250 and below. Or customers who've purchased three or more times versus those who've purchased once. With this segmentation, it may turn out that the campaign generating the most transactions is primarily acquiring low-value, one-time buyers. Another campaign, with a smaller budget and fewer conversions, brings in customers who return regularly and spend more. Without segmentation, it's easy to deprioritise or kill the second campaign entirely.

Common Questions About GA4 in E-commerce

Why is ROAS in Meta ten times higher than in GA4? Because they measure different things. Meta counts every ad interaction as a touchpoint — including a post like. GA4 attributes the conversion to the last-click source (or, in the data-driven model, distributes credit across all channels proportionally). Both are true. Neither should be treated as the single measure of performance.

Are branded campaigns worth running? Yes. If you don't run them, competitors can capture traffic from users who are already close to a purchase decision and are searching for your brand by name. Allocating 10–15% of budget to branded campaigns is a justified cost of protecting what you've already built.

How often should you check GA4? It depends on the role. A simple report comparing week-over-week traffic is worth glancing at daily — to quickly catch an SEO drop or paused campaigns. A marketing manager should block an hour each week, ideally Friday, to review the data. A store owner should check at minimum once a month; the more operationally involved, the more frequently.

What should you do when add-to-cart rate drops? GA4 will show you that a problem exists and where in the funnel it occurs. It won't show you why. For that you need a session recording and heatmap tool — for example Microsoft's free Clarity, which integrates with GA4. You can then watch session recordings for the segment of users who added to cart but didn't purchase, and see exactly where they get stuck.

AI in Analytics: Where It Helps, Where It Falls Short

AI tools that connect to GA4 and let you query your data through conversation genuinely speed up analysis. A tool like Go Marble lets you ask questions directly against your analytics and ad data without building reports from scratch. For people who don't know where to start, it can even suggest which questions to ask.

But AI draws conclusions only as well as you feed it context. Without information about your business, goals, budgets, hypotheses, and current activity, the answers will be generic: "Improve site conversion, optimise ad campaigns." That's advice you can find in any e-commerce article. The more context you provide, the more precise the recommendations.

AI won't replace an analyst when it comes to understanding the causes of anomalies, interpreting edge cases, or asking the questions you don't yet know you should be asking.

Summary: Three Questions GA4 Should Be Answering in Your Store

Where to invest your ad budget — GA4 shows which channels genuinely acquire valuable customers, not just generate traffic and vanity conversions. Where in the funnel you're losing users — e-commerce reports broken down by device, channel, and campaign show exactly where the largest drop-off occurs. Where your growth potential lies — internal search, customer segmentation, and cohort analysis reveal what your users are looking for and which customers you haven't reached yet.

Every percentage point improvement in conversion at each funnel stage, multiplied by store scale and twelve months, adds up to a real revenue number — gained or lost. GA4 is the tool that helps you find those percentage points and act on them. The only condition: it has to be correctly configured and consistently used.

‍