Why Ad Platforms Report Conversions Differently

If you manage paid media, you’ve likely seen this frustrating scenario play out. Google Ads reports 400 conversions last month. Meta claims 250. Microsoft Ads adds 60 more. Tally them up, and you’ve apparently sold to 710 people. Yet your finance team’s report shows only 480 actual sales hitting the bank.

So who’s fabricating these numbers? No one.

Most marketers assume the data is broken or deceptive. Ad platforms almost always report more conversions than your internal records , and they do so because they count conversions using fundamentally different rules. Once you understand those counting methods, the apparent contradiction becomes far easier to interpret and act on.

Start with the incentive

Here’s the uncomfortable truth that explains nearly everything: it’s in a platform’s commercial interest to report more conversions. The more conversions a platform shows you, the better its performance appears. The better it looks, the more you believe it works. The more you believe, the more you spend. That’s the business model. Always remember what drives platform revenue.

This isn’t malice; it’s rational economics. Given a choice between conservative and generous counting, every platform has a structural reason to choose generously. And they all do.

It’s counting, not lying

Here’s the reframe that prevents this from becoming another “platforms are lying to you” rant. The number of real conversions is fixed. There are only so many actual sales in a given period, regardless of how many get reported across Google, Meta, and Microsoft combined. Three platforms can each claim the same sale , and frequently do , but the customer still bought only once.

Focus less on reconciling every number and more on understanding how each platform counts conversions. Don’t chase a perfect, unified figure across every platform. Understand the differences, work with what you’ve got, and accept that “good enough to move the dial” is usually the right standard.

The structural reasons numbers don’t line up

If you need to explain gaps to a CFO , or to yourself , these are the concrete reasons platforms diverge from each other and from your own systems.

Attribution windows are one of the biggest culprits. Meta defaults to a seven-day click window (plus a one-day view). Google Ads with data-driven attribution looks back up to 90 days. Before you change anything else, those two platforms are already counting different conversions simply because they’re looking at different time frames.

What counts as an engagement also varies. On Meta, a carousel swipe, a video view, or a post share can earn attribution credit. On Google Ads and Microsoft Ads, you generally have to click the ad. The customer journey is the same; the attribution rules aren’t.

View-through conversions, especially on YouTube, are a major source of inflation. Display, programmatic, affiliate, and YouTube channels often count conversions from people who saw an ad rather than clicked it. YouTube view-throughs, in particular, can overinflate results because a view is invisible to your analytics, ecommerce platform, and CRM. Those systems can only track clicks or arrivals. There’s nothing wrong with optimizing toward YouTube conversions , just don’t mark your retargeting homework with view-throughs. They should be modeled and, ideally, validated against incrementality.

The in-platform attribution model also matters. Google’s default data-driven attribution (DDA) spreads fractional credit across every interaction in the Google Ads environment over 90 days, based on its machine learning model. Meta typically uses a last-touch, one-touch model. Different distribution logic produces different reported numbers for the exact same underlying journey.

Platform silos versus analytics platforms add another layer. Each platform can only see what happens inside its own walls. Google Ads tracks Google Ads. Meta tracks Meta. An analytics platform, CRM, or ecommerce system sees the whole journey across email, paid social, organic, affiliate, and direct channels, then assigns credit using its own logic , often on a last-touch basis. That’s why those systems report different numbers from the platform dashboards. It’s also why every platform can lay claim to the same conversion.

Modeled conversions have emerged as privacy changes broke the old way of tracking. Every platform built systems to fill the gap, each with its own methodology based on the data it has. Google uses enhanced conversions and Consent Mode. Meta uses data matching, looking for personally identifiable information (PII) indicators to determine whether someone who interacted with an ad later converted. The need is legitimate, but it’s also a real source of discrepancy , and, in places, a black box.

Cross-device tracking further complicates things. Google and Meta both model journeys that span multiple devices belonging to the same person. That modeling is another genuine reason the numbers diverge from each other and from your own systems.

The cost of misreading platform data

Conversion data leads to insight. Insight leads to decisions. If the insight is built on misread data, the decisions will be worse every time. The people who get burned aren’t the ones who understand the differences. They’re the ones who can’t explain why one platform says one thing and another says something else. That’s where you trip up and hand a stakeholder insight that’s simply wrong.

There’s also an accounting trap worth naming. The moment you treat platform numbers as the gold standard for accounting, the whole thing falls over. Conversion actions are tracked using different methodologies. They aren’t counting the money in your bank. Use them for optimization within performance marketing rather than for accounting. Those are two different jobs.

This is also where CMO and CFO skepticism comes in , and it’s usually misjudged. Leaders are often suspicious of platform-reported data. Assuming a clean setup , one data layer, consistent GTM triggers, and a robust tracking framework , there’s no reason to believe the numbers are wrong. They’re counted differently and counted generously. Understanding that distinction is what matters.

The pragmatic principle to land on

Here’s the rule of thumb that makes this usable rather than paralyzing: If all the numbers are moving in the right direction , overreported Google Ads, overreported Meta, overreported Microsoft , there’s a very good chance the business is moving in the right direction, too. You don’t need a perfect, reconciled figure to know whether your marketing is working. You need the platforms to be broadly trending in the right direction and the business data to confirm it.

What does good look like?

There’s no problem using platform metrics to optimize campaigns, feed the algorithms, and report up to your CMO, CEO, or CFO. The non-negotiable is understanding the counting methodologies behind each platform. That’s the line between useful insight and confidently wrong insight.

Mature advertisers go a step further. They move beyond raw platform counts through incrementality testing, marketing mix modeling, and first-party data that attributes performance to real customers and real purchases. The single most valuable move is feeding the right business data back into the platforms. Instead of obsessing over which platform claims more conversions, focus on lifetime value, CAC, product margin, returns, and lead quality. Those are the business signals that drive business results.

Comparing which platform reports more conversions is often a fruitless exercise. The edge comes from feeding genuine business signals to the algorithms, so they generate genuine business outcomes.

The one thing to do tomorrow

Ask your paid media team a single question: Do you understand the different accounting methodologies between the platforms? If they can’t explain why Google says one thing and Meta says another, that’s the gap to close first. Everything else follows from it.

Platform numbers aren’t wrong. They’re counted differently and generously. Use them to optimize, not to account. Feed your true business data back to the algorithms, and let those signals drive optimization.