ACOS vs ROAS and Why They're the Same Number Wearing Different Clothes
ACOS vs ROAS explained with formulas, a conversion table, break-even math, and when to use each. Learn why Amazon flipped to ROAS and how to set a real target.
On this page
ACOS and ROAS are the same two numbers pointed in opposite directions. One divides spend by revenue. The other divides revenue by spend. That is the entire difference, and yet the two labels quietly push people toward different decisions.
So let me settle the confusion first, then show you when each one earns its keep.
The formulas, side by side
ACOS (advertising cost of sales) tells you what percentage of your ad-driven revenue got eaten by ad spend. Amazon Ads defines it as ad spend divided by ad revenue, times 100.
ACOS = (ad spend ÷ ad revenue) × 100
ROAS = ad revenue ÷ ad spend
Spend $50, earn $100, and your ACOS is 50%. That same campaign has a ROAS of 2. Same inputs. Amazon Ads calls ROAS the inverse of ACOS, and that is precisely what it is.
Google uses ROAS exclusively. Its Google Ads Target ROAS strategy has you set a percentage goal, and their own example is clean: $5 in sales for every $1 spent equals a 500% target ROAS. Amazon historically reported ACOS. So if you run both platforms, you are already translating between the two whether you notice it or not.
If you want the full definition of return on ad spend, including benchmarks and break-even math, our pillar on what ROAS is covers that ground. This piece stays narrow: how the two metrics differ in practice and which one to reach for.
The conversion table you should tape to your monitor
Because the two are inverses, you can read either one off the other. This is where the psychology gets interesting.
| ROAS | ACOS | Plain reading |
|---|---|---|
| 100% (1x) | 100% | You broke even on revenue, lost on cost |
| 200% (2x) | 50% | Half of ad revenue went to ads |
| 250% (2.5x) | 40% | |
| 400% (4x) | 25% | |
| 500% (5x) | 20% | Google's example target |
| 1000% (10x) | 10% |
Adchieve makes a point worth repeating: ROAS has an exponential feel that misleads people. A jump from 1500% to 2000% ROAS looks huge. In ACOS terms it is a move from 6.7% to 5%, barely a nudge. Meanwhile going from 250% to 500% ROAS sounds modest but halves your ACOS from 40% to 20%. That is a massive efficiency gain hiding behind a small-sounding percentage change.
This is the real reason to know both. ROAS flatters improvements at the high end and undersells them at the low end. ACOS keeps the scale honest.
When ACOS reads better
ACOS answers a profitability question in one glance: what share of my sales am I handing back to advertising? That framing lines up directly with margin.
Amazon Ads ties this to break-even ACOS, which equals your product profit margin. Keep ACOS below that margin and you make money on the marginal sale. Cross it and you are buying sales at a loss. If your margin is 35%, a 30% ACOS is comfortable and a 40% ACOS is bleeding. You do that comparison in your head instantly because both numbers are percentages of revenue.
Try doing the same with ROAS. A 35% margin implies a break-even ROAS of roughly 2.86x. Correct, but nobody eyeballs that during a Monday standup. ACOS wins on plain readability whenever the conversation is about margin.
When ROAS reads better
ROAS is the language of bidding automation and portfolio decisions. Google's Smart Bidding literally sets max CPC bids to hit an average ROAS target, predicting the value of each auction in real time. To use Target ROAS on Search and Shopping, Google Ads requires at least 15 conversions in the past 30 days, and it recommends the strategy specifically when your conversions carry different values.
ROAS also plays nicely with revenue-first thinking. When you are deciding whether to pour another $10k into a channel, "this returns 4x" is a cleaner mental model than "this runs 25% ACOS." And the moment you widen the lens to whole-account efficiency, ROAS scales up into MER (marketing efficiency ratio), which ACOS does not do gracefully.
A note Amazon Ads makes and I will underline: do not treat either number as your only KPI. New campaigns run high ACOS purely because they are new and still learning. Killing them for a bad early number is a classic own goal.
A worked example (hypothetical)
Say you sell a $60 kitchen gadget. Costs of goods, fulfillment, and fees run $39, leaving a $21 margin, so your margin is 35%. That makes your break-even ACOS 35% and your break-even ROAS about 2.86x.
Here is one month of a hypothetical Sponsored Products campaign:
- Ad spend: $4,000
- Ad-attributed revenue: $16,000
- ROAS: 16,000 ÷ 4,000 = 4x
- ACOS: (4,000 ÷ 16,000) × 100 = 25%
At 25% ACOS against a 35% break-even, you have 10 points of margin cushion. Two ways to read that. The conservative operator says "good, we are profitable, bank it." The growth operator says "we have room to spend more aggressively and let ACOS drift toward 32% if it buys meaningful incremental volume."
Both readings come from the same two numbers. The metric you display just tilts which instinct fires first. When you push spend harder, protect efficiency deliberately rather than by accident; our guide on scaling paid ads without wrecking ROAS walks through marginal ROAS and how to add budget during the learning phase.
The trap both metrics share
ACOS and ROAS are last-click, platform-reported numbers. They tell you what the ad platform claims it drove. Neither tells you what would have sold anyway.
That gap matters most for branded search and remarketing, where a chunk of "attributed" revenue was heading to checkout regardless. A gorgeous 8x ROAS on branded terms can be mostly demand you already owned. The only way to know the true lift is to test it, which is the whole point of incrementality testing rather than trusting the dashboard number at face value.
Profit also lives outside these ratios. A strong ROAS with an ugly customer acquisition cost and no repeat purchase can still sink a P&L. ACOS and ROAS measure ad efficiency on a single order. They say nothing about lifetime value.
So which do you use?
Use whichever your team reads fastest, then hold the line. Amazon-native sellers usually think in ACOS because break-even against margin is instant. Google and multi-channel teams usually think in ROAS because it matches Smart Bidding and rolls up to MER.
Just do not run both dashboards in both formats and let people quietly optimize against different ceilings. Pick one as your primary, keep the conversion in your back pocket, and translate at the borders. If you want the arithmetic done for you across scenarios, our ROAS calculator handles the math, and if you would rather hand the whole profitability question to operators who live in it, that is what our paid media team does.
The metric war is fake. The margin question underneath it is the only thing that was ever real.