Event taxonomy document
Every tracked event named, defined, and mapped to a real business action, with the parameters attached to each one.
Foundation
GA4 and Google Tag Manager, set up so the data is trustworthy
Most GA4 installs are a container ID pasted into the header and left alone. We build the event taxonomy, the consent logic, and the cross-domain handling that make the numbers worth trusting.
Almost every GA4 property we inherit was set up the same way: someone pasted a measurement ID into the site, accepted the default event collection, and never looked at it again. It reports page views. It reports the automatically detected events GA4 guesses at. It does not report anything close to what the business actually needs to know, and nobody notices until a decision gets made on top of bad numbers.
GA4 setup done properly is not a five-minute task. It is a deliberate design exercise: deciding what a conversion actually means for this business, naming events consistently, deciding what parameters travel with each event, and making sure the data collection respects consent without silently breaking measurement the moment someone declines a cookie banner.
We do this work in Google Tag Manager rather than hardcoding tags into the site, because GTM gives your team the ability to add or adjust tracking without a developer for every change, and it gives us a single place to audit what is firing, in what order, and under what conditions. A messy GTM container is almost as bad as no GTM container, so we also clean up dead tags, duplicate triggers, and firing conflicts we find along the way.
The result is a property where the events map to things the business actually does: a qualified lead, a completed checkout, a scheduled call, a signed-up trial. Not just 'click' and 'scroll' with no business meaning attached to them.
What it is
What the setup actually covers
GA4 and GTM implementation touches taxonomy, consent, cross-domain tracking, and the defaults that are usually wrong out of the box.
Event taxonomy comes first. We map out what actions matter to the business, name them consistently (not a mix of past tense and present tense, not three different names for the same action across pages), and define which parameters ride along with each event so the data can be segmented later without re-instrumenting anything.
Consent handling is not optional anymore in most jurisdictions, and it is also frequently implemented in a way that quietly breaks all measurement for a large share of visitors. We wire Google's consent mode correctly so that when someone declines analytics cookies, the property still receives modeled, privacy-respecting signal instead of nothing at all, and so that accepted consent actually enables the tags it should.
Cross-domain and cross-subdomain tracking matters whenever checkout, booking, or a portal lives on a different domain or subdomain than the marketing site. Left unconfigured, GA4 will report every handoff between those domains as a new session with no source, which destroys attribution and inflates bounce rate on exactly the pages you care about most.
We also fix the common broken defaults: unfiltered internal traffic sitting in the main data view, referral exclusions missing for payment processors, session timeout settings that do not match the actual buying cycle, and Google Signals or user-provided data settings left in whatever state the property defaulted to.
Fit
Who this is for, and who it is not for
We would rather say no early than sell a program that cannot work.
Right fit
- Your GA4 data does not resemble how the business actually operates, and you have stopped trusting the dashboard.
- You are migrating a site or adding a new domain and need tracking to survive the transition intact.
- You need consent-compliant tracking that still produces usable data, not an all-or-nothing measurement gap.
Not the right fit
- Your GA4 property is already well instrumented and you just need report building. See reporting dashboards instead.
- You want a tool that tells you exact individual user identities. That is not what GA4 does or should do.
- You need this done with no involvement from anyone on your side. Consent and legal requirements vary by business and we need input on both.
Deliverables
What you get
A GA4 and GTM implementation documented well enough that someone else could maintain it.
Clean GTM container
Tags, triggers, and variables audited, deduplicated, and organized so future changes do not require reverse-engineering the setup.
Consent mode configuration
Consent mode wired to your existing banner so declined consent still yields modeled signal instead of a measurement blackout.
Cross-domain tracking
Linked domains and subdomains configured so sessions and attribution survive the handoff to checkout, booking, or a portal.
Default corrections
Internal traffic filters, referral exclusions, and session settings fixed so the numbers are not quietly wrong from day one.
Validation report
A documented test pass confirming every event fires correctly, once, and with the right parameters before we call it done.
How we run it
How the implementation runs
A structured build, not a container ID and a hope.
Step 1: Define the taxonomy
We work with your team to decide what a meaningful event and conversion actually are for this business before touching any code.
Step 2: Audit the existing setup
We document what is currently firing, what is duplicated, and what is missing in the current GA4 property and GTM container.
Step 3: Build and wire consent
Tags and triggers are built to the taxonomy, and consent mode is wired to the existing banner or platform.
Step 4: Cross-domain configuration
Linked domains, referral exclusions, and session settings are configured for the real structure of the site and checkout flow.
Step 5: Validate and document
Every event is tested in the GTM preview environment and in DebugView before handoff, with documentation for future maintenance.
Why is default GA4 tracking usually wrong?
Default GA4 collects generic automatic events with no business meaning, does not filter internal traffic, and often breaks under consent banners. It was never designed to answer specific business questions without deliberate configuration.
Does consent mode reduce the amount of data collected?
Yes, but properly configured consent mode still produces modeled data for users who decline cookies, rather than a total gap. Poorly configured consent handling often loses far more signal than legally required.
Where this connects
Where GA4 setup connects
Clean measurement feeds directly into the rest of the analytics and paid media work.
Once events are trustworthy, the next question is usually which of them count as a real conversion worth optimizing bidding around.
The taxonomy we build here becomes the source for reporting dashboards that answer specific business questions instead of restating raw event counts.
Ongoing account monitoring and tuning is covered separately under strategy and optimization once the initial build is stable.
Clean audience and event data is a prerequisite for effective retargeting since retargeting audiences are only as good as the events used to build them, and broken GA4 signal means retargeting the wrong people or nobody at all.
For a broader view of what the data is showing across channels, see insights which builds on top of a correctly instrumented property.
Questions
GA4 setup questions we get asked
Get analytics data you can actually make decisions on
A GA4 and GTM setup built around your real business events, not the defaults.