Practice & troubleshooting · Knowledge
How do I properly test whether tracking works?
Trust no dashboard you have not tested against yourself: the five-minute routine below verifies any tracking setup – and finds the classic mistakes on the way.
The order matters: source code first (is the snippet there?), then the live check (does data arrive?), then the edge cases (cache, exclusions, ad blockers).
The four-step routine – and what each step catches
Step one, page source: open the site in a private window, view source, search for your tracker hostname. The snippet must appear exactly once in the head section – zero means plugin inactive or Joomla cache serving old markup (clear it), twice means a forgotten template-level copy competing with the plugin. Step two, live check: keep the private window open, visit two or three pages, then watch the Visits Log in Matomo (or GA4's realtime view) – your visit should appear within seconds, with the right page titles. Step three, network tab: the browser's developer tools show the actual request to matomo.php or the GA endpoint; a red or blocked entry here while the snippet looks fine points to an ad blocker, a consent tool suppressing scripts, or a content-security-policy header. Step four, the exclusion cross-check: open a page you deliberately excluded (JStats menu exclusions) and confirm it does not appear – a test most people skip and the reason exclusion bugs live long. Run this routine after every relevant change: plugin updates, template switches, consent-tool changes. It costs five minutes and replaces weeks of quietly wrong numbers.
Key facts
- Step 1 – source: snippet exactly once in the head; zero = cache/inactive, twice = template leftover.
- Step 2 – live: private-window visit appears in the Visits Log within seconds.
- Step 3 – network tab: blocked tracker requests expose ad blockers, consent tools or CSP headers.
- Step 4 – cross-check exclusions: excluded pages must stay silent – the most-skipped test.
- Rhythm: rerun after plugin, template or consent changes – five minutes against weeks of wrong data.