GA4 is the analytics platform most businesses are now running, whether they chose it deliberately or migrated reluctantly when Universal Analytics was switched off. Many are using it without fully understanding what it is measuring, what it is not, and where the gaps are significant enough to affect decisions.
This is not a critique of GA4 as a product. It is a recognition that analytics tools have inherent limitations, and that treating any platform as a complete picture of marketing performance produces misleading conclusions.
The Session Model Has Changed
Universal Analytics built its reporting around sessions. GA4 is built around events. This is a meaningful architectural difference, not just a UI change.
In UA, a session was a relatively intuitive unit: a visit with a beginning and an end. In GA4, everything is an event, and sessions are constructed from those events using specific rules. The result is that session counts between the two platforms are not directly comparable, and metrics that look similar — bounce rate, time on site, pages per session — are calculated differently and mean different things.
This catches out businesses that are comparing current GA4 data to historical UA data expecting like-for-like continuity. The numbers will not match even when nothing meaningful has changed in actual user behaviour, because the measurement methodology has changed.
Attribution Is a Model, Not a Measurement
GA4’s attribution reporting attempts to assign credit for conversions to the touchpoints that contributed to them. The default attribution model has changed from last-click to a data-driven approach, which is theoretically more sophisticated but also more opaque.
The data-driven model uses machine learning to distribute credit based on patterns in your conversion data. It sounds rigorous. The limitation is that it is working from the data GA4 can observe, which is increasingly incomplete.
GA4 cannot observe cross-device journeys without user ID implementation. It cannot observe conversions that happen without a trackable click. It cannot account for touchpoints outside its tracking scope — a podcast mention, a word-of-mouth referral, a physical event. The attribution model distributes credit intelligently among what it can see. What it cannot see is simply missing from the analysis.
For businesses with multi-channel, multi-device customer journeys, this means GA4’s channel attribution is a partial picture presented with the confidence of a complete one.
The Privacy Gap
GA4 is subject to the same privacy constraints as every other analytics platform. Cookie consent requirements mean a significant proportion of users are opting out of tracking. Browser privacy settings, ad blockers, and intelligent tracking prevention further reduce observable data.
The result is that GA4 is not measuring all your users. It is measuring the subset of users whose behaviour is trackable under current privacy conditions, and extrapolating from that subset. For some businesses in some markets, the gap between measured and actual traffic is substantial.
GA4 does use modelled data to partially compensate for consent gaps, but this modelling has limitations and is not universally applied. The headline numbers in your GA4 reports are not a census of your user behaviour. They are a sample, with all the caveats that implies.
What Conversion Paths Actually Look Like
GA4 shows you the conversion paths it can observe. These tend to systematically underrepresent the early touchpoints in a customer journey.
Upper-funnel interactions — a social media post, a display impression, a video view — are less likely to generate a trackable click and more likely to contribute to later behaviour that GA4 attributes to a different channel. Paid social in particular tends to be underrepresented in GA4’s channel analysis, because a significant proportion of its contribution occurs before the last trackable click.
If you are using GA4 to evaluate channel performance and making budget allocation decisions based on those evaluations, you may be systematically undervaluing channels that work at the top of the funnel and overvaluing channels that work at the bottom.
Where GA4 Is Genuinely Useful
None of this means GA4 is not worth using. It is worth using. It provides genuinely useful information about on-site behaviour, content engagement, conversion funnel performance, and audience characteristics that cannot be obtained from ad platform reporting alone.
GA4 is useful for understanding what happens after the click — where people go, where they drop off, which content drives engagement, which pages correlate with conversion. It is useful for identifying technical issues with user experience. It is useful as one data source in a broader measurement picture.
What it is not useful for is being your sole or primary basis for evaluating marketing channel performance or making attribution decisions. That requires a broader framework.
Building a More Complete Picture
The businesses with the most honest view of their marketing performance are combining GA4 data with ad platform reporting, MER tracking, and where possible, incrementality testing or media mix modelling.
They are also triangulating against business outcomes rather than relying solely on tracked metrics. If revenue is growing and GA4 shows organic as the top channel while paid social shows poor attributed performance, the question worth asking is whether the organic growth might be partly driven by paid social awareness activity that GA4 is not capturing.
The tool is the starting point for the analysis, not the end of it.
0 Comments