Home/Conversion Tracking Setup
Conversion Tracking Setup
Conversion tracking setup that stops you losing sales
GA4, Tag Manager, Meta Conversions API and offline conversions configured properly, so the platforms bid on real revenue instead of a half-broken signal. Bad tracking quietly raises your cost per sale every single day it runs. Fixing it is usually the cheapest profit increase available to you.
The problem
Why your platforms never agree
This confuses more business owners than any other part of paid advertising, and it is genuinely not your fault.
They count different things
Meta will credit a sale to an ad someone merely saw. Google credits the click. Shopify credits whatever source came last. Three honest systems, three different answers.
They use different windows
A seven-day click and one-day view window is not the same as a thirty-day window. Change the setting and your ROAS changes without a single thing happening in the real world.
A third of the data is missing
Consent refusals, ad blockers and browser restrictions quietly remove a large slice of events. The platforms model the gap, and modelling is a polite word for estimating.
The goal is not to make them match. They never will. The goal is to pick one source of truth for money decisions, understand exactly what each platform over-reports, and read them without being misled.
What is included
What a proper setup actually involves
GA4 that means something
Events named and configured around your business, key events marked properly, and the default noise switched off so reports are readable.
Tag Manager, tidied
Containers cleaned up and documented well enough that someone else could pick them up. Duplicate and orphaned tags removed, not layered over.
Meta Conversions API
Server-side events with proper deduplication against the browser pixel, so you recover data that consent and blockers would otherwise erase.
Enhanced conversions
Google's hashed first-party matching configured so Smart Bidding is optimising against a fuller picture of who actually converted.
Offline conversions
For lead-gen: closed deals pushed back from your CRM into both platforms, so they learn which leads turned into money rather than which forms got filled.
Two weeks of verification
Test purchases, real-time debugging, and a fortnight comparing what the platforms report against what actually landed in your account.
Why it pays
Bad tracking is a performance problem, not an admin one
People treat tracking as housekeeping. It is not. Smart Bidding, Advantage+ and Performance Max all decide who to show your ads to based on the conversions you send them. Send them the wrong ones and they will optimise, efficiently and relentlessly, toward the wrong people.
I have seen an account where the purchase event fired on the cart page. Meta dutifully found the world's most enthusiastic cart abandoners and reported a wonderful ROAS. The client had been scaling that campaign for four months.
Fixing the measurement first also means every change afterwards can actually be judged. Without it you are changing things and hoping, which is expensive and slightly demoralising.
Signs your tracking is lying to you
- Platform-reported revenue is wildly above what your bank shows
- Conversions appear with no matching orders
- Mobile conversion rate looks impossibly low or high
- A campaign has a brilliant ROAS but you cannot find the customers
- Numbers changed the week your cookie banner was installed
Any two of these together and it is worth looking before you spend another month optimising toward a fiction.
Not sure whether your numbers are real?
Bring your platform reports and your actual sales figures. Twenty minutes is usually enough to tell where the gap is coming from.
Book a free strategy call →FAQ
Tracking questions worth asking
Because they count differently. Meta credits a view, Google credits a click, Shopify credits whatever came last. None of them is lying and none of them is complete. The job is not to make them match, it is to pick one source of truth for money decisions and understand what each platform is over-reporting so you can read them without being misled.
GA4 configured with events that mean something rather than the defaults. Tag Manager tidy enough that someone else could read it. Meta Conversions API sending server-side so consent and browser blocking do not erase a third of your data. Offline conversions pushed back from your CRM if you sell to leads. Then a fortnight of checking it against reality.
If a meaningful share of your audience blocks trackers or refuses consent, yes, because the platforms are optimising against a partial picture and it shows up as a rising cost per sale. If you are small and running a simple store, standard pixel plus a clean GA4 setup may be plenty. I will tell you which situation you are in before quoting anything.
Yes, and quite often that is the whole engagement. Plenty of accounts do not need new campaigns, they need the existing ones to be optimising on data that reflects reality. Fixing the measurement first also means we can tell later whether any campaign change actually did anything.
Next
Where to go from here
Ad account audit
A written review of accounts, tracking and landing pages, with every fix ranked by likely cost to you.
Read more →Meta Ads management
Once the pixel tells the truth, the account can finally be judged fairly.
Read more →Google Ads management
Smart Bidding gets dramatically better when the conversions behind it are real.
Read more →Ready to stop guessing what your ads are doing?
One call. Your numbers. A straight answer about what to fix first.