Why Your Ad Platform ROAS Numbers Are Wrong

If you have ever added up the revenue each ad platform claims it generated and found a number significantly higher than your actual store revenue, you are not alone. This is one of the most common frustrations in eCommerce marketing, and it has gotten worse since iOS 14.5 changed the tracking landscape.

The core issue is that every ad platform has a financial incentive to claim credit for as many conversions as possible. Facebook uses a default 7-day click, 1-day view attribution window. Google Ads uses a 30-day click window. TikTok uses 7-day click, 1-day view. When a customer interacts with ads on all three platforms before purchasing, each platform counts that single purchase as its own conversion. Your store processed one order for $80. Facebook says it drove $80 in revenue. Google says the same. TikTok agrees. Suddenly you have $240 in reported ad revenue from an $80 sale.

This double-counting and triple-counting is not a bug. It is a feature of how platform attribution works. Each platform operates in its own silo with no visibility into what the other platforms are doing. They cannot deduplicate because they do not share data with each other.

The problem has been compounded by iOS 14.5 and subsequent privacy changes. When Apple introduced App Tracking Transparency, a large percentage of users opted out of cross-app tracking. This means Facebook and other social platforms lost visibility into a significant portion of conversions. To compensate, these platforms now rely heavily on statistical modeling to estimate conversions they cannot directly observe. The reported numbers include modeled conversions alongside observed ones, and there is no way to tell which is which in the standard reporting interface.

The fix requires moving your source of truth from platform-reported data to an independent measurement system. This means collecting your own first-party conversion data and attributing it using consistent rules across all channels. When an order comes in, you look at the customer's actual journey across your site and attribute the conversion once using a single, consistent model.

Here is a practical framework for getting accurate ROAS. First, implement server-side tracking so you capture every conversion regardless of browser privacy settings. Second, build or use a cross-platform attribution system that assigns each order to a single source based on the customer's actual touchpoint history. Third, compare your independently attributed data to what each platform reports. The gap between the two tells you the platform's overreporting rate, which you can use as a correction factor for campaign-level optimization.

Some brands find that Facebook overreports revenue by 30 to 60 percent, while Google branded search is relatively accurate because it captures high-intent users who are about to convert anyway. The specific numbers vary by business, audience, and vertical. The point is that you cannot know the real performance of your campaigns without an independent source of truth.

This is not about discrediting ad platforms. Their data is still valuable for understanding relative performance between campaigns within the same platform. If Campaign A shows twice the ROAS of Campaign B in Facebook Ads Manager, Campaign A is almost certainly the better performer. The platform numbers are useful for intra-platform optimization. They are unreliable for cross-platform budget allocation and for calculating your actual return on ad spend.

The brands that figure this out gain a significant edge. Instead of making budget decisions based on inflated platform numbers, they allocate spend based on real, deduplicated attribution data. They know which channels are truly driving incremental revenue and which are merely taking credit for conversions that would have happened anyway. In a competitive market where margins matter, that clarity is worth its weight in gold.