Key takeaways
GA4 is a powerful tool, but if misconfigured, it produces inaccurate data that skews all your marketing decisions
Inflated "direct" traffic, "(not set)" values, and "unassigned" traffic are the most common symptoms of faulty tracking
The absence of consistent UTM parameters is one of the primary causes of poor attribution
Since the April 2026 update, discrepancies between GA4 and Google Ads have increased for many advertisers
Correcting your tracking is not just for experts: a few good practices are enough to ensure the reliability of your essential data
You have installed Google Analytics 4 on your site, your campaigns are running, and each week you carefully review your reports. Yet, something is wrong. An inexplicably high portion of your traffic is coming in as "direct". Some lines show "(not set)" without you understanding why. And the figures from GA4 do not match what you are reporting Google Ads or Meta. You are not a victim of an illusion — you are a victim of poorly configured tracking.
GA4 does not lie in the strict sense. It faithfully records what it is given to record. The problem is that what it is given is often incomplete, inconsistent, or simply incorrect. And when your data is wrong, your decisions are misguided: you cut budgets that were performing, you invest in channels that do not convert, you measure a ROI that does not reflect reality.
Inflated direct traffic : the first warning sign
In GA4, the "direct" source groups visits for which no origin could be identified. Theoretically, this corresponds to users who typed your URL directly into their browser, or who clicked on a bookmark. In practice, this category is often much broader — and for the wrong reasons.
Every time a visitor arrives at your site via an untagged link — an email without UTM parameters, a link in a PDF, a text message, a WhatsApp message, a Word document — GA4 cannot identify the source. It therefore categorises this visit as "direct". The result: you believe you have a lot of direct recognition, when in fact you have mostly a lot of untracked traffic.
If the Google Analytics tags are not properly configured, traffic sources may not be identified or categorised, leading to unassigned rows in channel groups, "(not set)" values, and an abnormally large share of traffic classified as direct. This is what most small and medium-sized enterprises experience when they have installed GA4 without prior audit.The UTM: the most underutilised tool in tracking
UTMs : The Most Underutilized Tracking Tool
UTM parameters are tags that you add to the end of your URLs to indicate to GA4 where the visitor is coming from. They consist of five key parameters: source (utm_source), medium (utm_medium), campaign (utm_campaign), content (utm_content), and term (utm_term). Without them, GA4 navigates blindly whenever traffic comes from a channel it does not automatically recognise.
The problem is not limited to their absence: their inconsistency is equally destructive. Common mistakes include using spaces instead of underscores, mixing naming conventions between platforms, or forgetting parameters on certain campaigns — each inconsistency fragments the data and makes attribution analysis nearly impossible.
Specifically, if your email campaign is tagged "Email" in one send and "email" in another, GA4 will create two distinct rows in your reports. Multiply this by the number of campaigns, channels, and collaborators creating links — and you end up with an unreadable dashboard. The solution is simple to implement: a shared UTM convention document, with clear rules on capitalisation, punctuation, and the naming of each parameter.
Attribution and models : what GA4 calculates (and what it forgets)
Attribution is the central question of any serious tracking: which channel deserves credit for a conversion? GA4 uses a data-driven attribution model by default, which relies on machine learning to distribute credit among the various touchpoints in the customer journey. It is more sophisticated than the old "last click" model of Universal Analytics — but it is not infallible.
GA4 systematically undervalues mobile advertising while over-attributing direct traffic and organic search. The cause: limitations of tracking cross-device, restrictions imposed by browsers and operating systems regarding cookies, and shopping journeys that span multiple sessions and devices. A customer who discovers your offer via an Instagram ad on their phone, who returns two days later from their computer via a Google search, and then converts by clicking on an email — this journey is extremely difficult to accurately reconstruct in GA4.
Source of the problem | Symptom in GA4 | Solution |
Untracked links (email, PDF, SMS) | Inflated direct traffic | Systematise UTM on all links |
Inconsistent UTM | Duplicated sources, fragmented data | Shared naming convention |
Cookie consent denied | Missing data, undercounting | Consent mode + cookie-less measurement |
Limited cross-device tracking | Misattributed conversion | GA4 / Google Ads connection + User ID |
GA4 update April 2026 | GA4 / Google Ads divergence | Post-update audit, double tracking |
The April 2026 update : an additional source of confusion
If you have noticed anomalies in your GA4 reports since spring 2026, you are not alone. The restructuring of the GA4 attribution model in April 2026 introduced a more interconnected system encompassing key events, conversions, report views, and linked advertising accounts — causing many businesses to see a change in the credit attributed to traffic sources in conversion reports, and increasingly visible discrepancies between GA4 and Google Ads.
This update is the most disruptive since the disappearance of Universal Analytics, and it is already affecting conversion tracking, audience lists, and attribution reports for thousands of Google Ads advertisers. If you have not audited your GA4 setup since this update, it is likely that some of your data is inaccurate — and that your Google Ads campaigns are running on degraded signals.
The short-term response is to not rely exclusively on GA4 for the management of your paid campaigns. Cross-referencing GA4 with Google Ads' native conversion tracking, and with Meta Ads' internal reports, gives you a more robust view than a single imperfect source of truth.
Where to start to ensure reliable tracking ?
No need to start everything from scratch. Three priority actions are enough to correct the main tracking issues for a micro or small business. The first is to check that Google Tag Manager is properly installed and that the GA4 tag is firing correctly on all pages, including purchase confirmation or form pages. The second is to standardise UTM parameters on all your outgoing links: emails, LinkedIn posts, email signatures, newsletters, ad campaigns. The third is to correctly configure the consent mode (Consent Mode v2) to avoid losing data from users who refuse cookies, through the modelled measurement proposed by Google.
FAQ
Is GA4 reliable if nothing is changed after installation?
No. A default installation without audit or UTM parameters produces partial and often misleading data. GA4 requires serious initial configuration and regular monitoring to be usable.
What is the difference between "(not set)" and "unassigned" in GA4?
"(not set)" means that a required dimension has no value — often due to a poorly configured event. "Unassigned" refers to traffic for which GA4 could not apply a classification rule in its default channel groups, often due to missing UTM parameters or redirects that have removed attribution data.
Should I use Google Tag Manager or install GA4 directly in the code?
Google Tag Manager is highly recommended: it centralises all your tags, facilitates updates without developer intervention, and allows for proper management of the Consent Mode.
How can I tell if my GA4 data is reliable?
Compare your GA4 data with your other sources: number of orders in your CMS, conversions in Google Ads, leads in your CRM. Significant discrepancies are a sign of tracking that needs auditing.