Conversion Tracking

Campaigns optimize toward the signal they are given.

Five steps sit between a visitor action and an optimization decision. Break one and Google and Meta keep optimizing — just against the wrong data.

Book a demo

What is conversion tracking?

Conversion tracking is the path from a meaningful action on your website to a signal advertising platforms can optimize against: the action is captured in the browser, passed to GA4 and Tag Manager as an event, optionally routed through a server-side layer as well, counted as a conversion, and finally used by Google Ads and Meta to make decisions. The goal is not to measure as much as possible, but to measure the right things reliably.

The chain is only as strong as its weakest link

Website action → browser-side tracking → GTM/GA4 event → optional server-side layer → advertising signal → campaign interpretation → optimization decision. Most cases of "Google Ads does not work" are really a broken link two steps earlier.

It usually breaks in three places: a shop plugin pushes the same event itself, so conversions get counted twice. A single-page app changes the view without firing a page view. Or a form submits with no thank-you page to anchor the event to.

And a badly chosen conversion dilutes everything downstream: if newsletter signups, PDF downloads, and purchases all count as conversions, the platforms optimize toward whatever happens most often — not toward what earns money.

Which actions are worth measuring

Not every measurable action deserves the status of "conversion". Separating a tracked event from a counted conversion is the real conceptual work.

Form submissions

Contact and enquiry forms — usually the most important B2B conversion, measurable even without a thank-you page.

Contact clicks

Phone numbers and email addresses: high intent, frequently unmeasured entirely.

E-commerce steps

Cart, checkout start, and purchase. Steps behind a login or payment step are prepared from the detected checkout intent rather than observed from outside.

SPA view changes

Route changes without a real page load — otherwise exactly the funnel steps that matter go missing.

Context interactions

Downloads, video starts, accordion clicks: useful for understanding behaviour, deliberately not conversions.

Consent state

Mateq detects common consent banners and your site's Consent Mode v2 status and accounts for them during verification.

How a measurement plan comes together

1
Scan the website
Mateq opens your site in a real browser — a single page or a crawl of up to 25 — and detects forms, clicks, e-commerce steps, and SPA route changes.
2
Approve the events
You see the detected interactions as a list and make two decisions per event: should it be tracked, and should it count as a conversion?
3
Have it implemented
Mateq creates tags, triggers, and variables through the Google APIs, marks your chosen events as GA4 key events, and publishes the container.
4
Optionally add the server-side layer
On request, Mateq also provisions a server container and a first-party subdomain — a second measurement layer alongside the browser.
5
Verify and keep watching
Mateq then checks in a real browser whether the events actually fire, and repeats that on a schedule.

A measurement plan for a B2B website

Scanning a services website surfaces four relevant interactions. How they are classified decides how campaigns steer:

Illustrative example

Contact form submitted: event + conversion
Phone-number click: event + conversion
Brochure download: event only
Accordion opened: event only

Two conversions and two context signals — rather than four "conversions" that pull bidding toward the most frequent but least valuable action.

You make that call per event; Mateq proposes and implements, but does not decide for you.

Browser and server side are not an either-or

Browser-side tracking captures what happens in the browser. A server-side layer complements it: it gives companies more control over how measurement requests are processed and forwarded. Together they are an architecture, not substitutes.

The honest boundary matters: Mateq provisions that infrastructure — server container, hosting, first-party subdomain — and can verify that your domain resolves and that requests leave the browser. What an advertising platform books internally afterwards is not observable from outside. That is exactly why we word verification results as "observed", never as "arrived".

Server-side tracking is infrastructure, not an accuracy switch: it still depends on correct event definitions, on consent, and on clean implementation.

Common questions about conversion tracking

What do GA4 and Google Tag Manager each do here?

Tag Manager is the control room: it decides which event fires when and with what data. GA4 is one of the destinations, and the place where an event is marked as a key event — a conversion. Mateq creates both: the tags, triggers, and variables in the container, and the key events in GA4.

Why does conversion tracking matter for Google Ads?

Because bidding strategies optimize toward conversion signals. Without them the platform optimizes toward clicks. Defined wrongly, it optimizes reliably toward the wrong thing. Signal quality caps optimization quality — which is why Mateq treats "clicks but zero conversions" as a tracking finding.

Do I need developers for this?

Not for the setup: Mateq detects the interactions during the scan and creates the configuration through Tag Manager — you do not paste code per event on the site. Tag Manager itself has to be installed on the website, which it usually already is.

How many conversions should you define?

As few as possible, as many as necessary. A rule of thumb: anything that directly creates business value is a conversion; everything else is a context signal. Too many conversions dilute optimization, too few starve it of data.

Start from what your website actually offers

A scan shows which interactions are measurable — and which of them deserve to be conversions.