Google Tag Manager consultant for teams whose numbers do not add up
Analytics reports one figure, the ad platform reports another, and nobody can say which to trust. The causes are familiar: containers nobody has tidied in years, GA4 events that were never agreed, consent settings blocking tags they should not. I work with in-house marketing teams and as the implementation resource behind agencies, on tracking across the UK.
Google Tag Manager consulting
Container builds and clean-ups
Most containers I have picked up agency-side have grown by accretion: duplicate tags on the same event, triggers nobody left a note about. I rebuild around a naming convention and a documented data layer.
Server-side tagging
Server-side tagging moves the vendor connections off the browser and onto a server under first-party control, so tags fire from a domain the browser trusts. I build the GTM server container and either hand it to the client's own cloud account or host it myself, whichever is preferred.
GA4 implementation and event design
GA4 is only as useful as the events fed into it. I design the event and parameter schema against what the business reports on, implement it through the data layer, and set up conversions and audiences.
Consent Mode v2 and CMP wiring
Consent Mode tells Google tags what a visitor has agreed to, so in advanced mode they still run in restricted form before consent and in full after it. Version 2 is required for advertising features in the UK and EEA. I wire the consent platform in and check the signals reach Google rather than stopping at the data layer. There is a fuller explanation of what each consent signal does.
Meta CAPI and enhanced conversions
The Conversions API sends events to Meta from a server, so they survive ad blockers and browser restrictions. I have built one and run it in production. Google's equivalent is enhanced conversions, which sends hashed customer data with the conversion to improve matching, alongside the Google Ads tag running through the server container. I have set out how the hashed match data is sent in more detail.
BigQuery export and reporting
GA4's own reports run out quickly. The BigQuery export gives the raw event data, which answers questions the interface cannot — full-path attribution, joins against CRM and order data. I set up the export and write the queries.
Tracking audit
I would normally start with a written website tracking audit, setting out what is currently collected and what it would take to fix what is broken.
Container inventory
Every tag, trigger and variable, what fires each one, and what has become dead weight.
Data layer and events
What the site pushes, what GA4 receives, and where the two disagree.
Consent
How consent is captured, what Consent Mode sends at each state, and what is lost before a visitor decides.
Match quality
Deduplication between browser and server events, Meta match quality, and how platform conversion counts compare with the source of truth.
Findings come back ranked by impact, with a fix list another GTM consultant or an in-house developer could implement without me.
Why server-side tracking
Browser-side tracking has been degrading for years. Safari's Intelligent Tracking Prevention caps script-set cookies at seven days, so a visitor returning after a fortnight is counted as new. Ad blockers strip the analytics and advertising scripts outright for a meaningful share of traffic. The result is under-reported conversions and ad platforms optimising against incomplete data.
Server-side tracking does not recover all of it, and claims that it does are overselling. What it does is move the vendor connections into a first-party context: the browser sends one request to a subdomain of the site, and the server decides what goes on to Google, Meta and the rest. Cookies written by the server outlast the seven days Safari allows a script-set one, provided the endpoint is genuinely first-party. A CNAME pointed at a vendor's host gets capped the same way, which is one reason I host these myself. The setup behind a first-party endpoint is written up separately.
How I work
I run this setup myself. The GTM server containers I maintain sit on my own server behind nginx and Cloudflare, carrying production traffic, including the Meta Conversions API integration I built; the infrastructure they run on is documented. I am not recommending an architecture I have only read about.
Behind that, 10+ years agency-side in Manchester. At McCann I owned the GTM setups across the accounts I led, covering GA4, Consent Mode and CAPI; at Soap Media, GTM, GA4, CRM integrations and attribution modelling. As Head of Paid Media at Platform81 in Stockport I see the cost of bad tracking weekly, in bid strategies trained on the wrong conversions.
Clean event data is also what the AI agents and reporting automations I build run on.
Engagements
Find out what is actually broken
The written audit above, quoted on its own. No obligation to have me do the fixing.
Build or rebuild the tracking
Container rebuild, server-side stand-up, GA4 event design, consent wiring. Admin access, documentation and the server container are handed across at the end; the setup should not need me afterwards.
Keep it working
Tracking breaks when the site changes. A small monthly arrangement covers the container and new tracking requirements. Most setups do not need it.
Common questions
Do we actually need server-side tagging?
Often not. A site with modest traffic, one ad platform and no consent complexity gets most of the benefit from a clean web container. Server-side earns its cost where paid spend is significant or Meta match quality is poor.
Can you work alongside our agency?
Yes. I am happy to be the implementation resource behind an agency's own client relationship, working to their naming conventions and documenting every change. Whether the client ever hears my name is the agency's call.
How long does an audit take?
It depends on how much is in there. A tidy single-domain container is quick to read; one with hundreds of tags, several consent states and a server container behind it takes proper time. I scope it before quoting, so the timeline comes with the quote.
What drives the price?
The number of platforms involved and whether server-side is in scope. A GA4 rebuild on one container is a different job from standing up a server container with consent and CAPI behind it. The audit is quoted separately from any implementation.
Does this have to be local?
Tracking work is remote almost without exception: container access, a staging URL and the ad accounts cover it. Being in Manchester only matters if someone wants a person in the room for the kick-off, which I am glad to do around the North West.
Contact
Google Tag Manager consultant, Manchester
Send a URL and read access to the container, or just say what looks wrong. I will come back with what I think is broken and what fixing it involves. No form, and no qualifying call before there is something to talk about.