Marketing & growth
Analytics & tracking
Numbers you can act on: a measurement plan written before any tag is added, GA4 and Tag Manager configured against it, consent mode wired to your banner, and every event documented.
The short version
A great many GA4 properties were migrated in a hurry and never audited since. Events fire twice, internal traffic gets counted as customers, the thank-you page records a conversion on every refresh, and four reports produce four different numbers. That is worse than having no analytics, because wrong numbers make people confident. We start from the decisions you need to make, write the measurement down, and only then touch a tag.
What is included
- A measurement plan document: the questions the business needs answered, then the events, parameters and naming convention that answer them, signed off before implementation begins.
- GA4 configured deliberately — key events marked, internal and developer traffic filtered, cross-domain tracking set, unwanted referrals excluded, and data retention raised to 14 months rather than left on the 2-month default.
- Google Tag Manager rebuilt on a data layer instead of CSS-selector triggers that break the next time somebody edits a page, with named versions and annotated changes.
- Consent mode wired to your consent banner and tested in both the granted and denied states, with a written record of what is collected before consent and what is not.
- Server-side tagging in a container on your own subdomain where the volume justifies it, so first-party cookie lifetime survives Safari's restrictions and fewer events are lost to blockers.
- A short dashboard covering acquisition, funnel and key events, plus a tag inventory document so whoever inherits this receives documentation rather than a mystery.
How we work
Our approach
The stages this work runs through, in order. Every one of them ends in something you can see or sign off.
-
Step 01
Write the questions first
Before any tag, we write down the decisions this data has to support: which channels get next quarter's budget, where the funnel leaks, which pages earn their keep. Those questions produce a short list of events and the parameters each event needs, in a naming convention that stays readable once there are 60 of them. You sign the plan off. Arguing about an event name in a document costs far less than renaming it across three platforms six months later.
-
Step 02
Audit what fires today
We then look at what is actually happening, in debug view rather than in a summary report. Duplicate events, conversions counted on a page refresh, self-referrals, redirects stripping campaign parameters, consent blocking things nobody realised it blocked. We reconcile GA4 against your back office or CRM and against what the ad platforms claim, then write down the size of each gap. Knowing why two numbers differ is more useful than picking whichever one you prefer.
-
Step 03
Rebuild on a data layer
Tags should read structured events pushed by your site, not guess from CSS classes that change with the next design tweak. We specify the data layer for your developer, or work within whatever your platform allows, then build the container to read it. Everything is tested in preview mode against a checklist, event by event, in the states a real visitor actually arrives in. Versions get named and annotated so a change made next year stays traceable.
-
Step 04
Handle consent, then scale
Consent mode gets wired to your banner and tested in both directions, because a setup that only works when everybody accepts is not a setup. We keep personal data out of GA entirely, strip it from URLs where forms leak it into query strings, and document what is collected under each state. If your volume and your reliance on ad platforms justify it, we move measurement to a server container on your own subdomain and verify everything again from scratch.
Before you ask
Questions we get asked first
How long does it take?
A clean GA4 and Tag Manager rebuild for a standard site runs 2 to 3 weeks: roughly a week on the plan and the audit, a week implementing, then a QA pass across devices and consent states. Ecommerce with item-level data, or a server-side container, adds 2 to 3 weeks and needs developer time on your side. After go-live, allow 28 days of collection before comparing anything to anything, because partial data invites the wrong conclusion.
How is it priced?
A fixed price in pounds against the event list in the signed-off measurement plan. Server-side tagging is quoted separately, because it carries a running cloud hosting cost that scales with your traffic and that bill goes to you directly rather than marked up through us. If the audit uncovers work outside the plan, we price it and you decide whether it is worth doing now or later. There is no hourly meter running behind any of it.
What do you need from us?
Admin on GA4, Tag Manager, your ad accounts and your consent tool. A developer who can push events to a data layer, or a clear picture of what your platform allows without one. Access to the back office or CRM figures we will reconcile against. And a single owner who can say which numbers the business genuinely reports on, because a measurement plan built to capture everything ends up measuring nothing well.
How do we know the tracking is right?
By reconciliation rather than faith. Key events in GA4 get compared against orders in your back office or leads in your CRM, and we agree the tolerance in advance, typically within a few percent for a client-side setup and tighter once measurement runs server-side. Every event is verified in debug view in both consent states before sign-off. You keep the tag inventory and the QA checklist as documents. We then re-check the whole setup 30 days after launch.
Want this scoped for your situation?
Tell us the goal and the deadline. You get an honest read on whether this is the right service, a fixed price if it is, and a straight no if it isn't.
Or email [email protected] directly. We reply within one working day.