Meta and Google double counting: what your combined ROAS hides

Meta and Google double count when both platforms claim credit for the same order, because each uses its own attribution model and sees only its own touchpoints. Adding their reported revenue together therefore overstates what paid media produced. You estimate the overlap by comparing combined platform revenue with actual paid-attributed orders in your store.

Why do both platforms claim the same sale?

A customer sees an Instagram ad on Tuesday, searches your brand on Thursday, clicks a Google ad and buys. Meta records a view-through or click-through conversion within its window. Google records a click conversion. Both are honest about what they observed. Neither knows about the other. Your finance team, meanwhile, sees one order.

Multiply that across thousands of orders and combined platform ROAS can sit far above reality. It is one of the most common hidden costs we list in our hidden costs checklist, and one of the least discussed, because nobody's dashboard shows it.

How do you estimate the overlap?

1. Take a period of at least four weeks.
2. Add Meta-reported purchase revenue and Google-reported conversion revenue.
3. From your store, take revenue from orders that had any paid touchpoint, using UTMs or click IDs.
4. The ratio of the two is your overlap factor. A combined platform total well above store paid revenue means heavy double counting.
5. Repeat by campaign type: branded search and retargeting usually overlap most.

The method is not perfect. Store tracking misses some view-through influence, and that influence is real. But it gives you a defensible ceiling on what paid media could have produced, which platform totals do not.

Where does overlap concentrate?

• Branded search: often captures people Meta already persuaded.
• Retargeting on both platforms: two networks chasing the same cart abandoners.
• Performance Max: broad inventory that can include branded and remarketing traffic.
• Long view-through windows: Meta credit for people who saw an ad and later bought via another route.

What should you do about it?

Do not try to settle which platform "really" deserves the sale. That argument has no winner. Instead, use three referees:

• Blended MER for the overall health of paid media.
• True-profit ROAS reconciled against real orders for each campaign.
• Incrementality tests, such as regional pauses, for the big questions like branded search.

Then make budget decisions on marginal profit, as described in our budget reallocation guide.

Definitions

• Attribution window: the time after a click or view within which a platform credits a conversion.
• View-through conversion: a purchase credited to an ad that was seen but not clicked.
• Overlap factor: combined platform-reported revenue divided by store-recorded paid revenue.

Why this matters for agencies and teams

When separate agencies run Meta and Google, each reports its own platform total and each looks excellent. The brand pays twice in credit and once in cash. A shared, reconciled profit view ends this quietly. It is also why we argue for evaluating agencies on contribution, as covered in our agency scorecard article.

Cross-channel measurement is a broader problem too: audiences move between platforms, and few teams can measure them consistently. SOMIN's paid-media intelligence approaches it from the audience side, reading what people respond to across channels rather than relying on one platform's view of the journey. For an agency example, see the Click2View Asia case study on somin.ai.

If you are rebuilding your measurement around reasoning models and agentic workflows, our sister studio GPT5 Marketing works with CMOs on exactly that redesign.

An illustrative overlap calculation

Suppose over four weeks Meta reports 60,000 in purchase revenue and Google reports 50,000, a combined 110,000. Your store shows 75,000 of revenue from orders with any paid touchpoint. The overlap factor is 110,000 divided by 75,000, roughly 1.47. In plain terms, for every unit of real paid revenue, the platforms together claim about one and a half. If the combined platform ROAS looked like 4.4 on 25,000 of spend, the reconciled figure is closer to 3.0, before any returns or fees. Now split the same calculation by campaign type. You may find prospecting campaigns overlap very little while branded search and retargeting overlap heavily. That split tells you where to test first.

How do you run a simple incrementality test?

Pick a region or a set of comparable regions. Pause or sharply reduce one campaign type, for example branded search, in the test region for two to four weeks while leaving the control unchanged. Compare total orders, not platform conversions, between test and control. If total orders barely move, the campaign was mostly claiming sales that would have happened anyway. If they drop, it was genuinely adding sales, and you now know by roughly how much. Keep the test simple and long enough to escape weekly noise; a clean small test beats an ambitious messy one.

Overlap is not a reason to distrust either platform. Both remain essential sources of signal about who saw what and when. The point is to stop adding their claims together and calling the total growth, and to make budget decisions on numbers that match the orders your business actually shipped and kept.

Two dashboards claiming one sale is not growth. It is bookkeeping.

How FMedia handles overlap

FMedia reads Meta and Google side by side and matches their claimed conversions to real orders from your store. It reports an overlap factor by campaign type and shows true-profit ROAS on reconciled revenue, alongside each platform's own figure. You see where double counting is heaviest, usually branded search and retargeting, and can test or cap those areas first. No more arguing about whose model is right; just one profit number both teams can work from.

Frequently asked questions

Why does Meta report more sales than my store?

Meta counts conversions within its attribution windows, including view-throughs, and cannot see that Google or another channel also touched the order. Combined platform totals therefore exceed real orders.

Should I turn off view-through attribution?

Not necessarily. View-through influence can be real. Shortening windows or reading results against reconciled store revenue and incrementality tests is usually more useful.

Which campaigns overlap most?

Branded search, retargeting on both platforms and broad Performance Max campaigns tend to show the heaviest double counting.