Short answer
GA4 and ad platforms count sales by different rules: different attribution models and windows, ad date versus conversion date, view-through conversions and cookie consent. That gap is normal. A sudden or unexplained gap usually points to a setup mistake, such as a duplicate tag or the wrong trigger.
Say that at the Monday meeting Google Ads shows 42 conversions, GA4 shows 31 and the CRM has 27 new customers. This is the point where someone asks “why don’t GA4 and Google Ads match?”, and the answer is rarely a single mistake. Part of the gap in GA4 conversions comes from the two tools counting the same event by different rules, and that is normal; part of it comes from something missing in the setup, and that needs fixing. In this article we explain how to tell the two apart.
In measure before you raise the budget we explained why measurement needs auditing before budget decisions are made. Here we take a step back and look at why the numbers drift apart.
Why GA4 and Google Ads count differently
Attribution model and attribution window
Each tool decides by its own rule whom to credit with a sale. GA4 shares a conversion across all channels: if the same customer first clicked an ad and then came back from a newsletter and bought, it may give the credit to the newsletter or split it between the two channels. Google Ads, on the other hand, sees only its own ads and doesn’t take account of an email or organic visit in between.
The attribution window is different too. The window sets the longest time that can pass between a click and a conversion. According to Google Analytics Help, GA4’s lookback window for key events is 90 days by default; in Google Ads, each conversion action has its own conversion window. If the window differs between the two tools, a sale that comes long after the click is counted in one and not in the other. Check which window is selected in your settings on both sides, and make a note of it.
Ad date versus conversion date
Google Ads records a conversion on the day of the ad that led to it, and not on the day of the sale; Google Ads Help calls this the “ad impression date”. GA4 records the event on the day it happened. If a user clicked the ad on 28 September and bought on 3 October, Google Ads shows the sale in the September report and GA4 in the October report. A large part of the gap in month-end comparisons comes from this alone. Google Ads also updates past days’ figures retrospectively as late conversions come in; the number you looked at yesterday may have changed today.
Platforms crediting themselves
Meta, Google Ads and other ad platforms credit sales to themselves in their own reports. If the same customer saw ads on two platforms, the sale appears in both reports. On top of that come view-through conversions: if a user only saw an ad without clicking it, then came by another route and bought, the platform may credit the sale to itself. (Google Ads leaves view-through conversions out of its “Conversions” column by default; platforms such as Meta may show them in their reports.) GA4 sees only the visit to the site; it knows nothing about the view. That is why the total of the platform reports is almost always higher than GA4 and than the till.
Cookie consent, ad blockers and browsers
If a visitor has rejected the cookie notice, the tags in the browser either don’t fire at all or fire in a limited way. If consent mode is in use, Google’s tools fill this gap to some extent with modelled data; but GA4 and Google Ads don’t do this modelling in the same way. Ad blockers and some browsers’ tracking restrictions also cut off some tags entirely. Because each tool is affected by this loss to a different degree, the gap grows.
Cross-device
A customer may see the ad on their phone and buy on a computer. A platform that recognises the user through an account they are signed in to can join these two visits up; GA4, unless your site has a login system and user IDs, sees them as two different people.
Processing delay and modelled data
Data in GA4 takes a while to settle. Today’s and yesterday’s figures may change later. Always make comparisons for a period that closed at least a few days ago. If the reports contain modelled data, also bear in mind that it is an estimate, not a real count.
Time zone and currency
If the GA4 property and the Google Ads account are set to different time zones, sales close to midnight fall on different days. Currency causes the same problem: if one side reports sales in Turkish lira (TL) and the other in a different currency, or one side sends amounts including VAT, comparing revenue becomes meaningless. In a Google Ads account, the time zone and currency are chosen when the account is opened and can’t be changed later, so the GA4 side has to be aligned with it.
Is GA4 conversion tracking not working, or is it counting twice?
First, a note on names: GA4 now calls the important actions on your site “key events” and keeps the word “conversion” for what Google Ads measures. In this article we call both conversions, because that is the familiar term.
The gaps above are in the nature of measurement. The ones below are setup mistakes, and they distort the numbers.
Duplicate tags. The GA4 tag may be added directly to the site’s code and also loaded through Google Tag Manager. In that case every page view and every event is counted twice. The same thing happens with the Google Ads conversion tag.
Conversions counted twice. If the same purchase is both imported from GA4 into Google Ads and counted by a separate Google Ads tag, and both are marked as primary conversions, Google Ads sees the same sale twice. Recording a new sale every time the thank-you page is refreshed is another common mistake.
Form submissions measured with the wrong trigger. If a form is measured by clicks on the submit button, every attempt that is sent with a missing field and returns an error also counts as a conversion. The other way round, if the page doesn’t change after the form is submitted, a trigger tied to page views counts nothing. The right trigger is the event that shows the form was submitted successfully: the success message appearing, a real thank-you page or an event sent by the form itself.
A gap between platforms is not a problem in itself. The problem is when nobody can explain where the gap comes from.
Where the gap comes from: audit steps
Below is the order we follow when auditing a GA4 setup and a Google Tag Manager configuration. If you are setting up the property from scratch, work through the basics in our GA4 setup checklist before these steps.
- Write a measurement plan. List in a single document what each event means, which ones count as conversions and which one feeds the bidding strategy.
- Give each conversion event a single source. Every conversion should fire from one place on the site, preferably from GTM. Clean out old tags added directly to the site’s code.
- Test with GTM preview mode. Turn on preview, fill in the form, complete the purchase steps and see which tag fires with which trigger, and how many times.
- Confirm with GA4 DebugView. In the same test, check that events reach GA4 under the right name, with the right parameters, and only once. For purchases, check that the value and currency parameters are filled in.
- Decide: import into Google Ads, or a Google Ads tag. Importing from GA4 is managed through a single setup and brings the two tools closer together. The Google Ads tag makes more direct use of Google Ads’ own measurement and some of its features. Choose one of them as primary; if you are going to use the other, mark it as secondary so the bidding strategy doesn’t see the same sale twice.
- Set up consent mode. In preview mode, test that the choice made in your cookie notice really reaches the Google tags, and that the tags behave as expected when consent is refused.
- Set up the cookie notice in line with KVKK. KVKK is Türkiye’s Personal Data Protection Law, Law No. 6698. Get explicit consent for non-essential cookies, make accepting and refusing equally easy, and explain in plain language in the notice which data is processed for what purpose. Prepare the legal text with your legal adviser.
- Align time zone and currency. Bring the GA4 property into line with the Google Ads account settings, and make sure revenue is always sent on the same basis (for example, excluding VAT).
- Keep a manual test log. Trigger each conversion once yourself, and write down how many times it appears in GA4, in the ad platforms and in the CRM.
When you need server-side tracking and the Meta Conversions API
Because a browser tag runs in the visitor’s browser, it is exposed to blockers, browser restrictions and slow connections. Server-side tracking sends events through a server under your control; the Meta Conversions API also passes the purchase or form event straight from your server to Meta. On the Google side, enhanced conversions narrow a similar gap.
These setups make sense in the following cases:
- Your ad budget is large enough that the loss in measurement would noticeably mislead the bidding strategy.
- Some sales are completed on the payment provider’s page, in an app or outside the site.
- Meta ads carry a significant share of sales and the browser pixel misses some events.
- Sales close on the phone, in a quote meeting or in a shop, and you want to send that information from the CRM back to the platforms.
Server-side tracking is not a substitute for cookie consent. If the visitor hasn’t given consent, the same rules apply to an event sent from the server. Also, if the same event is sent from both the browser and the server, it must be deduplicated with the same event name and a shared event ID (event_id), as Meta’s Conversions API documentation describes. Otherwise, instead of fixing the measurement, you will have added double counting. On a small site, moving to server-side tracking while the basic setup is still not clean doesn’t solve the problem; it only moves it.
One shared source of truth, not one right number
Trying to make GA4 and the ad platform match exactly is usually wasted effort; the two tools answer different questions. The more useful route is to choose one shared source of truth within the company. In e-commerce this is usually net sales in the order panel or the ERP; in B2B, it is the deals won in the CRM.
Once you have chosen this source, each tool’s role becomes clear. The CRM or sales data answers the question “how much did we sell?” GA4 shows behaviour across channels and the funnel on the site. The ad platform is used for comparisons within its own campaigns. When you keep the three numbers side by side as a reconciliation table in your weekly report, the usual ratio between them shows itself within a few weeks. If the ratio suddenly changes, look at the measurement setup first and only then at the campaign.
Setting up this arrangement, writing the measurement plan and, where needed, bringing the data together in a single dashboard are part of our data and CRO service. On our advertising service page we explain how this data is used on the campaign side.
Common questions
Which number should I trust: GA4, Google Ads or my CRM?
For sales, trust the net figure in the CRM or the order panel. Use GA4 to compare channels and behaviour on the site, and the ad platform to decide between campaigns. What matters is knowing which question each number answers.
How much difference between GA4 and Google Ads is normal?
There is no ratio that holds for everyone; it varies with the site, the sector, the cookie consent rate and the channel mix. What matters is that the gap stays steady from week to week and can be explained. A sudden jump usually points to a change in the setup.
Does every website need server-side tracking?
No. First, the GTM configuration, single-source conversions and consent mode must be clean. If the ad budget is large, if some sales are completed outside the site or if Meta ads carry a significant share, server-side tracking and the Conversions API are worth considering.
Does consent mode bring back lost data?
Not entirely. On visits where consent wasn’t given, consent mode estimates some of the conversions by modelling them. That estimate is useful for bidding strategies, but it is not a real count and should not take the place of the till figure.
The gap between the numbers usually doesn’t close, and it doesn’t need to. The real question is whether it stays the same size from week to week. If a gap that was small last month has clearly grown this month, something has changed: a tag, a consent setting or a redirect in the payment step. Finding that change before you touch the campaigns costs far less than switching off the wrong campaign.
Sources
- Google Analytics Help: Select attribution settings (GA4)
- Google Analytics Help: Conversions vs. key events in Google Analytics
- Google Ads Help: Data discrepancies, factors and troubleshooting
- Google Analytics Help: Monitor events in DebugView
- Meta for Developers: Handling duplicate Pixel and Conversions API events
This article is general information, not legal, financial or technical advice. Platforms, their rules and the law change over time, so some details may have changed since 25 October 2026 or contain small errors. Before you make a decision for your business, check the current official sources or write to us.