Which Conversion Number Should You Trust: GA4, Meta or Google Ads?

Meta says the campaign drove 50 sales. Google Ads says 35. GA4 says 28. Your bank account says something else entirely. Same week, same business, four different numbers. So which conversion number should you trust?

The frustrating answer is that none of them is wrong, and none of them is the whole truth. Each platform counts conversions its own way, for its own purpose, and comparing them directly is like comparing a thermometer to a clock. The skill is knowing what each number is for and which one to act on.

This guide explains why the numbers never match, what each platform is really counting, and how to build one version of the truth you can actually run your business on.

The numbers differ because the platforms are answering different questions. Meta wants to show what Meta drove. Google Ads wants to show what Google drove. GA4 wants to map the whole journey. Each is built to credit conversions through its own lens.

None of this is a bug. It is the result of different attribution models, different time windows, and different amounts of modeled, estimated data. Once you understand that, the goal stops being to make the numbers match. They never will. The goal is to read each one correctly.

Each tool counts a "conversion" differently:

• Meta. Self-reported and generous by design. By default it credits a sale within 7 days of a click or 1 day of a view of your ad, and it counts conversions it can only estimate. Meta has tightened its click definitions over time, but it still claims credit broadly.
• Google Ads. Attributes the conversion to the ad interaction, using data-driven attribution by default, which distributes credit across interactions based on your data. A click on Monday and a purchase on Friday show as a Monday conversion.
• GA4. Uses data-driven attribution to spread credit across the whole path, and counts the conversion when it happens. It sees more channels but credits each ad platform less than that platform credits itself.

So Meta and Google both over-claim relative to GA4, because each ignores the other. That gap is structural, not a mistake.

Two ideas explain most of the difference. An attribution window is how long after someone clicks or views an ad a conversion still counts. A longer window credits more sales. Meta's default window is wider than what GA4 typically gives it, so Meta's number is higher.

A quick illustration. A customer sees your Meta ad on Monday, clicks a Google ad on Wednesday, and buys on Friday. Meta counts a view-through sale, Google Ads counts a click conversion, GA4 splits the credit across both touchpoints, and your store records one order. Four reports, one sale. Multiply that by every customer in a month and the gap between your dashboards stops being mysterious.

Modeling is the estimate platforms make for conversions they cannot directly see, because of privacy changes, blocked cookies, and iOS limits. Both Meta and Google fill the gaps with machine-learning estimates. Those estimates are useful for optimization but they are not a headcount of real sales. The deeper breakdown of these mismatches is in our guide on why GA4 data does not match Meta Ads Manager, and the broader problem of trusting your numbers at all is covered in why businesses can't trust their marketing data.

The trick is to use each number for the job it is good at:

• Optimizing inside Meta? Use Meta's number. It is the signal the Meta algorithm itself learns from, so it is the right one for in-platform decisions.
• Optimizing inside Google Ads? Use Google Ads' number, for the same reason.
• Comparing channels against each other? Use GA4. It is the closest thing to a neutral referee across all your traffic.
• Deciding what actually made money? Use your backend or CRM. Real orders and real revenue are the only true conversions.

Trouble starts when you use one number for the wrong job, like comparing channels using each platform's self-reported figure.

You stop the confusion by deciding, in advance, what counts as truth. For most businesses that is the backend: actual orders, actual revenue. Everything else is a directional signal.

The setup that makes this work is server-side tracking feeding clean, consistent conversion data to each platform, with GA4 as your cross-channel view and your CRM or store as the final arbiter. Without that foundation, you are reconciling noise. With it, every platform is measuring the same events, so the gaps shrink and make sense. None of this works if the underlying tracking is broken, which is why launching ads without a proper tracking setup wastes budget before it starts.

You do not need to match the numbers. You need to know their relationship and watch it. Once a month:

• Pull all four. Meta-reported, Google Ads-reported, GA4, and backend revenue for the same period.
• Track the ratios. How much does each platform over-report versus your backend? That ratio is usually stable.
• Watch for drift. When a ratio suddenly changes, something broke: a pixel, a tag, a privacy update.
• Decide on the backend. Make budget calls on real revenue, and use the platform numbers only to optimize within each platform.

A stable set of ratios is worth more than four numbers that agree once and never again.

This is the tracking infrastructure we build and run. Server-side tracking, clean conversion events, GA4 configured as a true cross-channel view, and reconciliation against the backend, so you stop guessing which number is real.

Most advertisers argue about which platform is lying. We build the setup that makes the question disappear. See where your tracking stands today on our tracking and analytics service.