What Is a Good ROAS on Amazon?

What ROAS means on Amazon

ROAS is Return on Ad Spend: total ad-attributed sales divided by total ad spend. Amazon's own ad console reports it for you next to ACOS, which is the same relationship expressed as a percentage instead of a ratio. Spend $200, get $800 in attributed sales, and your ROAS is 4x.

On its own, that number tells you nothing about whether the campaign made money. A 4x ROAS is excellent on a product with thin margins and a disaster on one with almost no margin at all after Amazon's referral and fulfillment fees. That's the gap every generic benchmark misses.

Why 'the industry average is 2x to 4x' isn't the answer

You'll see this range quoted a lot, and it's not wrong exactly — it's just the wrong question. A ROAS of 3x could be deeply profitable for a supplement brand with 70% margin and disastrous for a private-label kitchen tool with 20% margin after COGS and Amazon fees. The average tells you what other people got. It doesn't tell you what you need.

What you need is your break-even ROAS: the point at which ad spend exactly eats your profit and no more. Below it, every sale loses money. Above it, every sale contributes.

How to find your own minimum ROAS

Start from gross profit before advertising, not the sale price. Take a product that sells for $40. Cost of goods is $12. Amazon referral and FBA fees run $10. That leaves $18 of profit before you spend a cent on ads — your break-even point.

Minimum ROAS = sale price ÷ profit before ad spend. Here that's $40 ÷ $18 = 2.22x. Spend enough to hit exactly 2.22x and you've broken even. Anything above it is real margin; anything below it, you're funding growth out of your own pocket — sometimes worth it for a launch, but you should know you're doing it.

• Higher margin products can tolerate a lower ROAS and still be profitable.
• Lower margin products need a much higher ROAS just to break even — sometimes 6x, 8x, or more.
• Your target ROAS is a business number, not an ad-platform default. Amazon's dashboard won't tell you what yours should be.

What a good ROAS looks like once you're spending at scale

Single-campaign ROAS gets noisy fast, especially once you add upper-funnel DSP alongside Sponsored Products. Awareness placements — online video, non-endemic display, prospecting audiences — are supposed to run a lower ROAS than bottom-of-funnel search. That's not underperformance, that's the job the placement is doing. Judging it against a Sponsored Products benchmark is comparing a fishing net to a spear.

Across 30 advertisers we manage, July 2026 delivered a 6.04x return on ad spend measured across the whole portfolio — not the best line item, the whole book. That same book ran 78.4 million impressions at a $4.00 CPM and a blended $1.42 CPC, a $5.49 cost per acquisition across 57,137 attributed purchases, and 20.1% of those purchases came from a shopper new to the brand. The $0.41 CPC you'll see quoted elsewhere in this category is online-video only — not the blended number, and not comparable to a search-heavy account.

The point isn't that 6.04x is the good number. It's that a portfolio number mixing prospecting and conversion will always read differently than a single search campaign, and any 'good ROAS' figure that doesn't say which mix it's measuring is telling you less than it looks like.

When your ROAS is bad news, here's what to actually check

A ROAS below your break-even number isn't automatically a crisis, and one above it isn't automatically a win. Check these before you touch bids:

• Attribution window mismatch. Sponsored ads and DSP can both claim credit for the same sale if you're reading dashboards separately instead of reconciling them — you'll double-count and think you're doing better than you are.
• Last-click isn't incrementality. A display impression with a last-click sale attached doesn't prove the display caused the sale — the shopper might have converted from an organic search click regardless. We've made this exact mistake: reading a prospecting line item's last-click ROAS as proof it drove the purchase, when a holdout test later showed most of that volume would have happened anyway.
• Cutting spend to defend ROAS. Pulling back to only your highest-converting, most branded keywords will lift the ratio and shrink the business at the same time. A rising ROAS on a shrinking sales number is not a win.
• Wrong benchmark for the placement. Judging a retargeting or prospecting campaign against a branded-search ROAS will always look like failure, because it's answering a different question.