Why this matters more than any campaign setting
Every optimisation decision you make is downstream of measurement. Smart Bidding optimises toward the conversions you report. If those conversions are wrong, the algorithm is being trained to buy the wrong traffic — efficiently, and at scale.
Broken tracking does not usually produce obviously absurd numbers. It produces plausible ones that are too good, which is why it survives for years.
Mistake one: counting thank-you page views as leads
A thank-you page view is not a form submission. People refresh. People bookmark the page. People arrive there from a browser history suggestion. Every one of those counts as a lead.
Worse, if the page is reachable by URL it can be indexed and crawled. The fix is to fire the event on a confirmed submission response, not on a page load.
Mistake two: the same conversion counted twice
This happens when a conversion is imported from Analytics into Google Ads and also tracked natively by the Ads tag. Both fire, both report, and the account shows double the leads it produced.
Check the conversion actions list. If a lead appears under two names with similar totals, only one should be marked as a primary conversion action.
Mistake three: GA4 installed twice
A site that has the Google tag hardcoded and the same measurement ID deployed through Tag Manager sends every page view twice. Sessions halve in duration, bounce rate collapses, and everything looks better than it is.
Open the page source and search for your measurement ID. It should appear once.
Mistake four: consent handling that silently discards data
If a consent banner blocks the analytics script entirely rather than using Consent Mode, visitors who have not yet clicked are missing from the data completely — and they are disproportionately new visitors, which is exactly the segment acquisition work is judged on.
How to check yours in an afternoon
None of this requires a paid tool. Work through it in order and stop at the first thing that does not match.
- Submit your own contact form and confirm exactly one conversion appears — not zero, not two
- Search your page source for your GA4 measurement ID and count the occurrences
- In Google Ads, open the conversion actions list and check what is marked primary
- Compare Ads conversions with GA4 key events for the same date range and ask why they differ
- Check whether your form's success state depends on the server confirming delivery, or just on the click
The number that settles it
Count the enquiries that actually reached your inbox last month. Compare that with what the platforms reported. If those two numbers disagree, believe the inbox.
Common questions
Should Google Ads and GA4 numbers match exactly?
No — they use different attribution models and conversion windows, so a gap is normal. A large or inconsistent gap is not, and is usually a setup problem rather than a modelling one.
Do I need Google Tag Manager?
Not for a simple site. It earns its place once you have several tags to manage, because it lets you change tracking without a code deployment.