Facebook Ads ROI Calculator for Amazon Sellers

ROI from Facebook ads is what the ads earn compared with what they cost. For a brand that sells on its own site, Meta's pixel sees the purchase and reports a return. For a brand that sells on Amazon, the purchase happens where Meta can't see it.

Amazon Attribution fills part of that gap. You put its tag on the links in your Meta ads, and it reports clicks, orders and sales from shoppers who clicked a tagged link and then bought within its attribution window. The Facebook ads for Amazon guide covers how to set it up.

So the return you can see is the sales Attribution credits, divided by Meta spend. That's a reported number. It isn't the same as the sales the ads caused, which is why this calculator keeps the two apart.

reported ROAS          = attributed sales ÷ Meta spend
cost per attributed order = Meta spend ÷ attributed orders
profit after spend     = attributed sales × margin − Meta spend
break-even ROAS        = 1 ÷ margin
caused sales needed    = Meta spend ÷ margin

1. Take spend and link clicks for the campaigns from Meta Ads Manager.
2. Take click-throughs, orders and sales for the same tags and dates from Amazon Attribution.
3. Work out your margin before ads with the break-even ROAS calculator.
4. Compare reported ROAS with break-even ROAS, then look at how much more the ads would need to cause to break even.

A made-up month of Meta ads pointing at one Amazon listing. These are the calculator's example values.

1. Reported ROAS: $2,880 ÷ $2,000 = 1.44. Cost per attributed order: $2,000 ÷ 96 = $20.83.
2. Clicks: Meta reports 4,000 link clicks and Attribution records 3,200 click-throughs, 80% of them. 96 orders from those is a 3% conversion rate.
3. Profit after spend: $2,880 × 40% − $2,000 = -$848.
4. Break-even: at a 40% margin, the ads must cause $5,000 of sales, 1.74× what Attribution credits.

The scenario table shows the same month if the ads had caused a different multiple of the credited sales. At 2×, they'd make $304; at 0.5×, they'd lose $1,424. Both are possible, and the report alone can't say which is closer.

Any ROAS above your break-even ROAS, counted in sales the ads caused. The catch is the second half. With Amazon Attribution alone, you only know the credited sales.

• Reported ROAS above break-even: the credited sales pay for the ads. Still worth testing, since some credited buyers might have bought anyway.
• Reported ROAS below break-even: the ads pay off only if they cause sales Attribution doesn't credit, like shoppers who saw an ad and searched for you later. The break-even multiple tells you how much they'd need to cause.
• Click-throughs far below Meta's clicks: check that every ad uses the tag and the link lands where you expect.

If you're enrolled in the Brand Referral Bonus, tagged traffic can earn a credit on referral fees. The Brand Referral Bonus calculator adds it to your return.

• Using Meta's reported revenue. Meta counts purchases its pixel or Conversions API sees, and Amazon doesn't send its purchases back. The attribution gap calculator compares the reports.
• Mismatched dates or campaigns. Spend and Attribution results must cover the same campaigns and days.
• Reading a scenario as a result. The what-if rows show what would happen at each multiple. They don't say which multiple is yours.
• Margin without fees. Your margin before ads comes after product cost and Amazon's fees, not just product cost.

The number that decides whether Meta ads pay is the Amazon sales they cause. Amazon Attribution can't give you that, and neither can Meta. A holdout test can: pause the ads in some regions and compare Amazon orders there with everywhere else.

Size a test with the holdout test calculator. Once you have a result, enter the extra orders it found above, or turn it into a return with the incremental ROAS calculator.