GA4, Google Tag Manager and Firebase measurement rebuild for a global event
GA4 implementation and attribution case study: GTM, GA4 and Firebase measurement rebuild for a global event — 564,959 events and $205,207 revenue tracked, 1 January–15 July 2026.
- Business type
- Global summit / event organiser
- Market
- Global
- Platforms
- GA4, Google Tag Manager, Firebase
- Time period
- 1 January–15 July 2026
Objective
Give the marketing team a measurement layer they could trust: every meaningful action on the site and app recorded once, named consistently, and attributable to the channel that brought the visitor — so paid media budget decisions used real registrations and revenue rather than clicks.
Starting situation
The event ran paid activity across search, cross-network and social, but reporting could not reliably answer which spend turned into registrations and revenue. Without a consistent event layer, the same action could be recorded under different names or missed entirely, and budget decisions rested on platform click numbers rather than business outcomes. Paid media can only optimise toward conversions it can see — so the measurement problem had to be solved before the media problem.
Strategy
- —Define one written event taxonomy first: a fixed list of events, their parameters, and what each one means, so nothing depended on whoever happened to add a tag.
- —Implement the taxonomy consistently through Google Tag Manager on the website, so events fired the same way on every page and form.
- —Set up event and acquisition reporting in GA4 so registrations, leads and revenue could be read by channel.
- —Instrument the app in Firebase so app behaviour fed the same measurement picture as the website.
Implementation
- —A GTM container built to the taxonomy, sending events into GA4 — including form_submit, call_click, whatsapp_lead and catalogue_download — each with consistent naming and parameters.
- —GA4 configured for event reporting and user acquisition reporting by first-user channel group, so the team could see where converting users actually came from.
- —Firebase instrumentation for the app, so app events sat alongside web events rather than in a separate silo.
- —Revenue tracking configured so registration income appeared in GA4 alongside the events that produced it.
Measurement setup
- —Every event was QA'd before it was used in reporting or bidding: fired on the real site and app, checked in GA4 realtime and debug views, and confirmed to fire once per action — not zero times and not twice.
- —Form tracking was validated end to end: 1,831 form starts against 1,543 completions, an 84% completion rate, confirming the form events recorded real user behaviour rather than misfires.
- —Acquisition reporting checked by first-user channel group, which is what surfaced the attribution picture: 54% of users arrived from paid search, cross-network and paid social.
Campaign changes
- —Paid channels were assessed on tracked registrations and revenue rather than clicks alone.
- —Budget conversations could reference the same numbers the site produced, instead of each platform's own self-reported view.
Results
Historical results for 1 January–15 July 2026.
- 564,959
- events from 37,921 users
- $205,207
- revenue tracked in GA4
- 84%
- form completion rate (1,543 of 1,831)
- 54%
- of users from paid search, cross-network and paid social
Source: GA4 events and user acquisition reports, 1 January–15 July 2026.
Limitations and context
- —Revenue is as tracked by GA4, not reconciled against financial accounts.
- —This is measurement work: the revenue and registrations are not claimed as caused by the tracking itself. The value of the work is that the numbers could be trusted and acted on.
- —Attribution shows which channel group brought users; it does not prove what would have happened without that spend.
- —Client name withheld on this site for confidentiality.
Past results do not guarantee future performance.
Lessons
- —A written event taxonomy keeps tracking maintainable after handover — without one, every new tag is a new inconsistency.
- —Validate events against real behaviour before trusting them: the 84% form completion figure is only meaningful because the underlying events were checked first.
- —Measurement is a prerequisite, not an afterthought: media optimisation is only as good as the conversion data feeding it.
Services used
Want a similar review of your account?
A free 30-minute call to look at your tracking and campaigns and agree whether there is work worth doing.
More case studies
Start with a conversation
Tell me what you need. I’ll tell you what I would do first.
Paid ads, SEO, tracking or an automation idea: we’ll talk it through together. No sales team, no hard sell, and no hand-off after the call.