Why Meta and Google report more sales than you actually made

In short

Each platform counts a sale if it had any part in it, within a time window it sets, so the same sale is counted on both Meta and Google. The way to see the truth is to measure against one source, the store or the CRM, track the ratio of total sales to total spend, and test from time to time what happens when you lower the budget.

The report doesn’t lie. It counts differently

When the Meta report shows 120 sales, the Google report shows 90, and the store has 150, it’s not necessarily a bug. Each platform answers a different question: “Which sales did I have a part in?” and not “Which sales wouldn’t have happened without me?”.

Three reasons for the gap

1. Attribution windows

On Meta, the default setting for websites counts a sale that happened up to 7 days after a link click, and up to a day after viewing or engaging with the ad. Someone who scrolled past the ad, didn’t click, and came in tomorrow from Google to buy can be counted on Meta.

2. Double counting across platforms

A customer saw an ad on Meta, searched for your name on Google and clicked an ad, then bought. Both Meta and Google attributed the sale to themselves. The store has one sale, and the reports have two.

3. Sales that would have happened anyway

Someone who searched for your brand name on Google, or a regular customer who saw a remarketing ad, would probably have bought without the advertising. The platform still counts them.

What changed in Meta’s measurement in 2026

In March 2026 Meta changed the definition of a click: only a link click counts as a “click”. Conversions that came after a like, comment, share or video view moved to a separate engagement category. Meta also offers an “incremental” attribution model, which tries to count only sales the model estimates the ad caused. These changes improve the picture, but don’t replace measuring against actual sales.

How to measure the truth

One source of truth. Sales are counted in the store or the CRM, and leads are counted as real leads, not by the report.
MER or total cost per customer. Total sales divided by total spend on all advertising. If every report improves and MER doesn’t move, advertising isn’t adding anything.
A separate brand campaign. Searches for the business name are measured separately, so they don’t inflate the campaigns that bring new customers.
UTMs on every link. That way you see in Google Analytics and in the store which channel each customer came from.
A holdout test. Lower the budget for a set period, or leave a region without advertising, and check how many sales actually disappeared. It’s the most direct way to know what advertising adds.

What it means for your decisions

Decide on budget by your actual sales numbers, and use the platform reports to compare campaigns and ads within the same platform. That way no ad looks successful just because it’s good at attributing sales to itself. For calculating the targets themselves, see the articles on target cost per lead and break-even ROAS.

Want to know where you stand? We’ll check the numbers against your account, free of charge.

Get a fit check

Frequently asked questions

What is Meta’s default attribution window?

For campaigns aimed at website conversions: 7 days after a link click, and 1 day after viewing or engaging with the ad.

Which report is right, Meta’s or Google’s?

Neither on its own. Each platform attributes to itself sales that others also attribute to themselves. The figure to decide by is the number of sales or customers in your own system, relative to total spend.

What is an incrementality test?

A test that compares a group exposed to advertising with a group that wasn’t, to see how many sales advertising really added. You can run it with the platforms’ tools or with a controlled budget reduction.

Sources

How Meta Ads Attribution Works in 2026 – Jon Loomer

Pini Smolanski

Founder of SMOLANSKI Media. 14 years in advertising management, 10 of them at Israel’s leading media and advertising agencies.