Are Facebook ads worth it for Amazon brands?

"Are Facebook ads worth it?" mixes two separate checks. The first is economics: does each caused sale leave enough after variable costs to pay for the ad? The second is measurement: do you know how many sales the ads actually caused?

Platform dashboards answer the second question with attribution rules, not experiments. For an Amazon-first brand, they also miss a large share of outcomes. Facebook ads for Amazon covers why Meta can't place its pixel on Amazon and what Amazon Attribution still misses.

ROAS is credited revenue divided by spend. "Worth it" is a profit question, so compare incremental ROAS to break-even ROAS:

break-even ROAS = 1 ÷ margin before ads
worth it when incremental ROAS > break-even ROAS

Use the margin left after product cost, Amazon referral and FBA fees, and any per-order costs you can't avoid. Contribution margin explains the idea; the break-even ROAS calculator turns your numbers into a threshold. There's no universal "good ROAS" without your costs.

Meta counts purchases its systems see within the ad set's attribution window. That includes buyers who would have reordered anyway, and it excludes buyers who saw an ad and searched your brand on Amazon. Researchers comparing large Facebook experiments with common non-experimental methods found those methods often misstated the true effect (Gordon and others, 2019).

A made-up supplement brand spends $8,000 a month on Meta. Ads Manager credits $32,000, a reported ROAS of 4. At a 35% margin before ads, break-even ROAS is 2.86. On the report alone, the ads look profitable.

A geo holdout test pauses Meta in matched states and compares Amazon orders by ship-to state. It finds about $14,000 of caused sales: incremental ROAS 1.75, below break-even. The ads destroy margin even though the dashboard glows green. Read what is ROAS for reported vs. incremental, and holdout groups for the test design.

Pick US states that normally sell in step, pause Meta ads in one group with location targeting for several weeks, and compare Amazon orders in those states with the rest, adjusted for how they differed before the test. That estimate is incremental sales. Divide by spend for incremental ROAS.

Size the test first. A few weeks of orders are noisy; a small brand may not detect a realistic lift unless the holdout is large enough. The holdout test calculator checks power before you spend. The Facebook ads Amazon ROI calculator turns a test result into profit after fees and product cost.

Facebook ads for ecommerce covers pixel and catalog setup; this page is the Amazon-first "worth it" check once ads are running.

• New products or markets where organic rank and brand search are still thin, and incremental orders clearly exceed break-even.
• Prospecting that reaches people who wouldn't have searched you, confirmed by holdout lift on Amazon, not just rising attributed ROAS.
• Creative that educates before the Amazon click, when the test shows incremental orders rise in holdout-off states while retargeting-heavy campaigns don't.
• Room above break-even after a honest test, so you're not scaling on credit from loyal buyers who would have reordered anyway.

Spillmark runs geo holdout tests for brands that advertise on Meta and sell on Amazon. It plans matched regions from your order history, locks the design before the test starts, compares Amazon sales there with everywhere else, and takes off Amazon fees and your product costs to show you the real profit.

Read how lift is calculated, then use the break-even ROAS calculator and holdout test calculator to see whether a test can answer "worth it" for your spend level. We tell you before a test starts if the design is too small to see a realistic lift.