How client-side tracking works
When someone opens your page, the browser downloads the page, then downloads the measurement code you have placed on it — a GA4 tag, a Meta Pixel, an advertising conversion tag, usually delivered through a tag manager. That code watches for things happening on the page: a page view, a click, a form submission, a purchase confirmation. When one occurs it packages up the details and sends a request from the visitor’s own device to the platform.
Everything about this happens on hardware you do not control. The visitor’s browser decides whether the script loads, whether the request is allowed out, and whether any identifier attached to it survives. That is the whole strength and the whole weakness of the approach.
Why client-side tracking matters
It is where almost every measurement setup starts, and for good reason. You can install it without a developer, change it in a tag manager while looking at the site, and see rich detail about behaviour on the page — scroll depth, which button was clicked, which product was viewed — because the code is sitting right there watching it happen. Nothing else gives you that much visibility for that little effort.
It also feeds the ad platforms the signals they optimise on. A campaign learning from browser-reported conversions is only as good as those reports, which is why the gaps below are a bidding problem as much as a reporting one.
Where client-side tracking goes wrong
Events go missing, and they go missing quietly. Ad blockers and privacy-focused browsers stop the request. A visitor closes the tab before the tag fires. A slow connection — common on mobile data in Nepal and anywhere else with patchy coverage — means the script never finishes loading. Cookie limits set by the browser shorten how long a returning visitor stays recognisable, so a conversion weeks later may not be joined to the click that caused it.
Two softer failures cause as much trouble. Tags accumulate until the container is slowing the page down, and nobody removes the ones for tools the business stopped using. And personal information ends up in tag parameters, where it is visible to anyone who opens the browser’s network panel.
What to do about it
Treat browser tags as your detail layer, not your source of truth for money. Build events on a clean data layer rather than fragile selectors that break at the next redesign, test every tag in preview before publishing, and audit the container once or twice a year so dead tags come out. A tag manager setup done properly makes all of this routine rather than a project.
For revenue events — purchases, qualified leads — add a server-side copy so the important numbers do not depend on a browser cooperating. Send both, join them with a shared identifier so the platform counts one action once, and keep the browser tag in place; server-side tracking is a companion to this, not a replacement.
