Most GA4 setups I audit were installed, not designed. Someone added the Google tag, turned on enhanced measurement, marked a few events as key events and moved on. Then an agency added its own tags, a developer hardcoded a purchase event, and the consent banner went live without anyone checking what it actually blocks. The result is a property that produces numbers nobody fully believes, feeding ad platforms that bid on whatever those numbers say. I do not advise on tracking from the sidelines. I build it, from the tracking plan to the dashboard.
How I work on GA4 & GTM
Start with a tracking plan
Before touching a tag, I write down what the business needs to know: which actions matter, which count as key events, which parameters travel with them and who owns each one. That plan drives the data layer, the GTM container and the GA4 configuration, and it stays useful long after launch because anyone can see why an event exists. If you want to see the structure, my tracking plan template is a good starting point.
Clean key events and ecommerce data
I remove duplicate installs, turn off enhanced measurement features that double count, and keep key events to the few actions that reflect real value. For ecommerce, I check that item data, transaction IDs and purchase values match your backend. For lead generation, I separate a form submission from a qualified lead, because they are not the same event and should not be treated as one. I also set data retention, internal traffic filters, cross-domain measurement and unwanted referrals, the settings that quietly decide whether your reports can be trusted.
Consent Mode that actually works
Consent is part of the build, not a banner added at the end. I connect your consent platform to Consent Mode, choose basic or advanced mode based on your risk tolerance and markets, and test every tag in every consent state. When consent is denied, the setup should behave the way you have told visitors it will.
Server-side tagging where it pays off
A GTM server container on a first-party subdomain gives you more control over what data leaves the site and can recover conversion signal lost to browser restrictions. It also adds hosting cost and maintenance, so I recommend it when the paid media budget justifies it, not by default. It never bypasses consent.
Ad-platform conversions and the CRM
I decide whether Google Ads bids on its own conversion tag or on imported GA4 key events, so the same action is never counted as primary twice. Enhanced conversions add hashed first-party data to Google Ads conversions, while the Meta Conversions API and other server events share an event ID with their browser pixels so each conversion is counted once. Click IDs and UTMs flow into your CRM so qualified leads and closed deals can be sent back as offline conversions.
Reporting people actually use
GA4’s standard reports are fine for quick checks. For decisions, I build Looker Studio dashboards around spend, leads, revenue and blended efficiency, and set up the BigQuery export when you need raw event data or event-level history beyond GA4’s data retention limits.
When you need a GA4 & GTM specialist (and when you don’t)
You likely need one when:
- GA4, your ad platforms and your CRM report numbers nobody can reconcile
- Nobody on the team can explain what half the tags in your GTM container do
- Your consent banner went live and conversions dropped, or never changed at all
- You are about to scale paid media and bidding depends on conversion data you do not trust
- Leads arrive in the CRM with no reliable source attached
If the first point sounds familiar, why GA4 and ad platforms don’t match explains which gaps are normal and which are bugs.
You probably do not, yet, when:
- You have a simple site, little ad spend and a basic setup that already records the few actions you care about
- The real problem is traffic or offer, not measurement. Clean data on a weak funnel only confirms what you already suspect.
How it fits the rest of your growth system
I do not treat tracking as a standalone project that ends when the tags fire. It is the first thing I fix in a growth engagement, as part of my marketing attribution and tracking work, because every decision after it depends on the numbers. Clean key events make Google Ads and Meta bidding smarter. Source data in the CRM shows sales which leads to call first. Reliable ecommerce events make CRO tests readable. When one person owns the tracking, the ads and the funnel, a broken event gets noticed in days, not at the end of the quarter.
For the kind of engagement this work sits inside, see the case study Premium Fashion Brand: Rebuilding Shopify Around Conversion.
What the first 30 days look like
| Week | Focus |
|---|---|
| 1 | Access and audit: GA4, GTM, consent, ad platforms, CRM and backend compared |
| 2 | Tracking plan agreed, data layer specified, quick fixes to duplicates and settings |
| 3 | GTM rebuild, Consent Mode, enhanced conversions and server events in testing |
| 4 | Changes live, data validated against your backend, dashboards and next 60-day plan |
I run new and old setups side by side where I can, so nothing important goes dark during the switch. By day 30 you have a documented setup, key events you can defend and a clear list of what comes next, such as server-side tagging or offline conversions.
Get it built
If your team argues about which number is right before it can decide what to do, start with a Growth Audit: $1,500 fixed, credited toward ongoing work if you continue. Tell me about your setup and I will show you what I would fix first.