TL;DR
Meta typically reports 20 to 50% more conversions than your store or analytics show. The gap comes from three structural differences: Meta counts view-through conversions (saw the ad, bought later, never clicked), it uses longer attribution windows than last-click analytics, and every marketing channel claims the same buyer, so the totals across channels always exceed reality. None of this means Meta is fabricating sales. It means Meta's number answers "how many purchases were we involved in" while your store answers "how many purchases happened." Judge Meta on incrementality and blended numbers, not on its self-reported scoreboard.
The three reasons the numbers diverge
1. View-through attribution. This is the big one. Meta takes credit for purchases where someone saw your ad, didn't click, and bought later within the attribution window. Your analytics, running last-click, credits whoever got the final click (often direct, organic, or email). Meta's view: the ad influenced the sale. Your analytics' view: the ad did nothing. Both are defensible. Neither is the whole truth.
View-through is not fraud. Ads genuinely work this way; people see things and buy later. But the credit is generous by design, and the generosity scales with your window. A 7-day click, 1-day view window is standard and reasonable. Anything longer, and Meta starts claiming ancestors.
2. Window length. Meta's default attribution window (often 7-day click, 1-day view) is longer than the last-click model most stores use implicitly. Longer windows mean more purchases fall inside Meta's claimed territory. Shorten the window and the gap shrinks. It also shrinks your reported ROAS, which is why most advertisers never touch the setting and then complain about the gap it creates.
3. Everybody claims the same buyer. Your Meta ad, your Google ad, your email flow, and your influencer's discount code can all claim the same $120 order. Add up every channel's reported revenue and you'll get $300 to $400 of "attributed" revenue for $120 of actual money. This isn't a Meta problem; it's a multi-touch reality. But Meta is usually the loudest claimant because its windows are the longest and its view-through net is the widest.
Which number do you trust?
Your store. Always. Revenue that your payment processor confirms is the only number that pays salaries.
Meta's number is useful for one thing: relative comparison. If Ad A reports 3.2x and Ad B reports 2.1x under identical settings, Ad A is probably better, even if both numbers are inflated. The inflation is roughly constant across your own ads, so the ranking holds. What doesn't hold is comparing Meta's number to your store's number and concluding either one is lying. They're answering different questions.
The practical setup:
- Store revenue is the target. All budget decisions ladder up to blended ROAS: total store revenue divided by total ad spend, across every channel. If blended ROAS is healthy, the machine is working, regardless of what any single platform claims.
- Platform ROAS is the steering wheel. Use Meta's numbers to compare ads and ad sets against each other, not against reality. Kill losers, feed winners, and don't confuse the dashboard with the bank account.
- Watch the ratio, not the absolute. If Meta usually reports 30% above your store and suddenly reports 80% above, something changed: tracking broke, a window shifted, or view-through is doing something weird. The ratio is your early warning system.
The incrementality question nobody wants to ask
Here's the uncomfortable follow-up: of the sales Meta claims, how many would have happened anyway? A returning customer who was always going to buy, sees your ad in passing, and gets counted as a view-through conversion. That's real attribution by the rules and zero incremental value to your business.
This is why incrementality testing exists, and why it's the only honest answer to "is Meta working." Holdout tests, geo experiments, the occasional brave pause: these tell you what Meta adds, not what Meta claims. You don't need to run them constantly. One clean test a year recalibrates everything, including your faith in the dashboard.
FAQ
Why does Meta report more conversions than Shopify? Meta counts view-through conversions and uses longer attribution windows than last-click analytics, and it claims influence on purchases other channels also claim. Expect Meta to report 20 to 50% higher.
Is Meta inflating its numbers? Not fraudulently. Meta's attribution rules are published and consistent. The numbers answer "purchases we were involved in," which is legitimately larger than "purchases we caused."
What is view-through attribution? Credit given when someone sees your ad without clicking and converts later within the attribution window. Standard on Meta (1-day view in the default window).
Should I change my attribution window? The 7-day click, 1-day view default is reasonable for most. Shortening it reduces the gap but also reduces reported ROAS; lengthening it does the reverse. Pick one and keep it constant so comparisons stay valid.
What is blended ROAS? Total store revenue divided by total ad spend across all channels. It's the number that reflects reality, and the one budget decisions should ladder up to.
How do I know if Meta is actually incremental? Run holdout or geo tests periodically. One clean incrementality test per year tells you what Meta adds versus what it claims, which recalibrates every other number you look at.
MetaMaxd is a Meta ads agency. We run the accounts; you run the business. Want a second pair of eyes before Q4? That's what we're for.
