GA4 vs Google Ads conversions (and Meta): why the numbers never quite match

When we took over Google Ads for Papírnictví VojTech in early March 2026, the account looked great. In April the ads claimed 167% of the store’s real revenue. More than the store had sold, full stop.

That’s an extreme case with a simple cause. Smaller gaps between GA4, Google Ads and Meta are everywhere. Each tool has its own attribution, its own window and its own calendar. Here’s which gaps are normal, which mean broken tracking, and what to decide by.

GA4 vs Google Ads conversions: two counts for one purchase

GA4 and Google Ads conversions don’t match because they measure different things. Google Ads counts conversions after an interaction with its own ads and books them on the day of the click. GA4 counts events on your site from every source and books them on the day of the purchase. A gap is normal. It’s an error when one tool reports more purchases than the store actually received.

Google Ads only sees Google’s world: clicks, YouTube views, impressions. GA4 sees site visits from every channel and splits the credit. One answers “how many purchases did my ads drive?”, the other “where did my buyers come from?”

Why GA4 and Google Ads conversions don’t match: six causes

Most gaps come down to one of these six. The first five are how the tools work. The sixth is a setup mistake.

1. A different day. Google Ads books a purchase on the day of the ad click, GA4 on the day of the purchase. Click on Monday, buy on Thursday: two different days, and at month end some purchases spill into the next month.
2. A different attribution model. Both default to data-driven attribution, each on its own data. GA4 also shares credit with channels Google Ads can’t see: email, organic, Meta. So Google gets a smaller slice in GA4.
3. A different window. Google Ads defaults to a 30-day click-through window, adjustable from 1 to 90 days. GA4 looks back 90 days for purchases by default.
4. Views without a click. Google Ads can count a purchase after a video view or an ad impression. GA4 works mainly with site visits.
5. Modeled conversions. When someone declines cookies, Google Ads fills in missing purchases with consent mode modeling. GA4 models by its own rules, only with enough data. Since 15 June 2026, the ad_storage consent signal alone decides which GA4 data your ads can use.
6. A different set of conversions. This is where most real errors live: add-to-cart set as a primary conversion in Google Ads, or the same purchase counted twice, once by the Google Ads tag and once by a GA4 import.

The Meta attribution window: why Meta claims more

For purchase campaigns, Meta’s default attribution setting is 7-day click, 1-day engage-through and 1-day view. Since March 2026 a click means a link click only: to your site, a lead form or an app. Likes, shares, comments and video views of at least 5 seconds now fall under engage-through, with a one-day window.

Two things push Meta above GA4. Views without a click: someone who saw your ad and came back through Google the next day is Meta’s purchase, and GA4 never knew about the view. And Meta books purchases on the day of the ad impression, not the purchase, and models part of its results. Comparing single days across tools doesn’t work.

GA4 vs Google Ads vs Meta: how each one counts

Google Ads and Meta each count their own ads’ purchases. Neither accounts for the other.

Why Google Ads plus Meta adds up to more than your orders

A customer clicks an Instagram ad on Sunday, a Google ad on Monday and buys on Thursday. That’s inside both click windows, so both platforms claim all of it. Your store back end shows one order; the platform reports show two.

What it does to your budget: add up ROAS from Google Ads and from Meta and you count the same revenue twice. The ads look more profitable than they are.

At MyDeko, where Meta creates demand and Google catches it, this is business as usual: Meta and Google each report their own purchases, and the sum is higher than the real number of orders. On top of that, both platforms claim people who would have bought anyway. That’s incrementality, which we break down in When 400% ROAS beats 500%.

Calculator: how many purchases are counted twice

Enter one month: back-end orders, purchases reported by Google Ads and Meta, and the share of orders GA4 puts outside paid ads (organic, direct, email). You get the minimum number of purchases both platforms claim.

This is a floor. Meta also claims part of the purchases GA4 sees as organic or direct, so the real overlap tends to be higher.

Data from our accounts

Three accounts where the gap between platform and back end shows most clearly.

At VojTech, counting carts as conversions did more damage than a bloated report. Smart campaigns like Performance Max learn from whatever you mark as success. Here they learned to find people who add to cart. Why people abandon the cart and how to measure it is in Cart abandonment: why shoppers leave right before they pay.

What we got wrong

In three accounts we didn’t fix tracking right away. Each mistake is now a checklist line.

Ten weeks on broken tracking

After taking over VojTech, we ran the account for another ten weeks on numbers that counted carts. April ended at a total ACoS of 13.7%, above the 12% target. Today tracking is the first thing we check on a takeover.

Contacts with zero value

At Letsport, campaigns kept learning from a “Contact” conversion with no value until August. That’s when we moved it to secondary. We should have done it in May.

A month without forms

At CarDetailer, Google Ads didn’t record submitted forms in July after the tracking switch. Leads came in; the report didn’t show them. Since then we verify a test conversion the same day after every tracking change.

Which number to trust

None of the ad platforms, as a final number. Each is good for a different job:

• Your store back end is the truth about revenue, orders and margin. That’s what budget decisions rest on.
• Google Ads and Meta are for steering campaigns inside each platform.
• GA4 shows the path across all channels and the share of orders outside paid ads.

The store back end (revenue, margin) decides, not the sum of what the platforms claim.

The rule we run client budgets by

In practice it’s one table a month: back-end revenue, Google cost, Meta cost and margin after ads. Next to it, total ACoS: ad spend divided by the whole store’s revenue. Grab the template on the left.

Key takeaways

Not sure whether your ads claim more than your store sold? Reconciling the platforms with the back end is the first thing we do when we take over an account as part of Google Ads management.

FAQ

Jiří Kopejska

Co-founder, Marketing ASAP

Co-founded Marketing ASAP. Starts every account takeover by putting the Google Ads and Meta reports next to the store back end and finding where they part ways.