Why it is your best buyers who go missing

If the undercount were random noise, you could shrug it off. A consistent 20% miss across every segment would shrink the absolute numbers but leave the shape intact, and the shape is what you make decisions on. The problem is that the loss is not random. It is biased, and it is biased in exactly the direction that hurts a B2B founder most.

The undercount has a demographic

Around 29.5% of internet users run an ad blocker, but that average hides the real story. Ad blocking skews young, technical and male. It peaks among 25 to 34 year-olds, and across the population men block at 49% against 33% for women. The people most likely to install an ad blocker are the people most comfortable installing software, reading about privacy, and caring about it: developers, technical founders, security-minded operators, the buyers a B2B SaaS most wants.

Sit with that for a moment. If you sell to technical buyers, the single segment most likely to be invisible in your analytics is your ideal customer profile. The visits you most want to understand are the ones the browser is best at hiding.

GA4 describes the wrong people

The damage runs deeper than volume. When the technical, ad-blocking slice of your audience disappears, GA4 does not just count fewer people, it describes a different audience. Your reported demographics skew older and less technical than your real customer base, because the technical cohort has been filtered out before it ever reached the report. The audience you see in the dashboard is the residue left after your best buyers have been removed.

Now follow that into the part that actually spends money. You feed those skewed demographics back into audience targeting and lookalike modelling. The platforms dutifully optimise toward the audience you described, which is the wrong one. Client-side data does not just under-report your best-fit segment; it quietly instructs your ad spend to go and find more of your worst-fit segment. You end up paying to acquire the people who least resemble your actual customers, guided by a measurement system that erased the right ones.

The compounding cost

This is why "client-side GA4 is good enough" is the most expensive sentence a paid-acquisition founder can say. Good enough for a vanity pageview chart, perhaps. But the moment that data touches a targeting decision or a bid, the bias compounds. Every campaign you optimise against it drifts a little further from your real market, and because the drift is invisible, you cannot diagnose why the leads keep coming in slightly off-profile. The fix is not better analysis of corrupt data. The fix is to stop corrupting the data at the source, which means moving the measurement somewhere the browser cannot reach in to bias it.