Why your Google Ads and GA4 numbers disagree
The common reasons GA4 and Google Ads report different conversion numbers, how to diagnose the discrepancy, and how to build tracking you can actually trust.
If you have ever opened Google Ads and GA4 on the same day and seen two different conversion counts, you are not imagining it. The platforms are counting different things, at different moments, with different rules.
Sometimes the gap is harmless. Sometimes it means you are optimising on a number that does not exist. Here is how to tell which situation you are in.
1. They count different events
Google Ads conversions are usually configured on a specific action you imported or created — a form submission, a call, a purchase. GA4 counts every event you send it, including ones you forgot about. If an old Google Ads tag is still firing alongside the GA4 import, you get two conversions for one enquiry.
Check: in GA4, open Admin → Conversions and list every key event. Then compare against the conversion actions in Google Ads. Duplicates are common after a migration.
2. Attribution windows are different
Google Ads defaults to a 30-day click window for many conversion actions. GA4's advertising reports use their own attribution settings — often data-driven, with a different lookback. A conversion from a click 20 days ago may count in one and not the other.
Check: compare conversion action settings in Google Ads with the attribution settings in GA4 Admin → Attribution settings.
3. Consent mode is blocking one tag and not the other
If a consent banner blocks analytics until the user accepts, GA4 loses those sessions. Google Ads may still receive modelled conversions from its own tag, or vice versa. On a site with a 60% consent rate, a large gap is expected — and it is the single most common cause I see in the UK and EU.
Check: compare the same date range with consent interactions in GA4's consent report, and confirm both tags respect the same consent state.
4. Forms fire on the wrong trigger
A classic: the conversion fires when the form page loads rather than when the form is submitted successfully. Add a validation error or a bot submission, and the numbers inflate. This usually makes ads look better than reality.
Check: submit a test enquiry and watch Tag Assistant. The event must fire after the success state, not on page view or button click.
5. Bots and spam
GA4 filters known bots, but form spam still counts as a conversion if it fires the event. If your inbox has spam enquiries, your conversion count includes them.
Check: reconcile a week of conversions against actual enquiries received. The difference is your noise floor.
6. Server-side and browser signals are out of sync
If you have moved to server-side tagging or the Meta Conversions API, the server event may carry a different event name or deduplication ID than the browser event. Done correctly, the platform deduplicates. Done carelessly, you double-count or drop events entirely.
Check: confirm the same event ID is sent from both browser and server. In Meta Events Manager, look for deduplication warnings.
How to build numbers you can trust
- One event schema. Write down the events that matter and name them consistently across GA4, Ads and Meta.
- Reconcile against reality. Every month, compare reported conversions with enquiries actually received. That ratio is your calibration factor.
- Use offline conversions. Import qualified leads from your CRM so ads optimise on sales, not form fills.
- Test after every change. Tag changes break silently. A five-minute check after each deployment prevents months of bad data.
- Document it. A one-page measurement plan saves arguments later.
Which number should you trust?
Neither, entirely. Trust the reconciliation: enquiries in the inbox, orders in the shop, calls answered. Platform numbers are directional signals for optimisation, not accounting. If the gap between reported conversions and real enquiries is stable, you can optimise on the platform number with a known correction. If the gap moves every week, fix the tracking before spending another pound.
If you want a second pair of eyes on your setup, send the site and the platforms you use. A short audit usually finds the cause within an hour.
Written by Rasel Rony — web developer and e-commerce consultant, fourteen years of client work. If this raised a question about your own setup, send it over.