Four metrics that describe the same relationship from different angles. What each one measures, how to convert between them, and why the ROAS your platform reports is not the one that decides whether you make money.
ROAS, ACoS, CPA and CPC are not four different things: they are four ways of looking at the same relationship between what you spend on ads and what you get. Each platform picked one and turned its users into native speakers of that dialect: Meta and Google speak ROAS, Amazon speaks ACoS, acquisition teams speak CPA, and whoever manages bids lives in CPC.
The trouble starts when you have to compare channels, or when someone decides a campaign is profitable using the wrong metric.
How much revenue each unit of spend generates. Expressed as a multiple: 4x means 4 of revenue for every 1 spent.
What share of revenue advertising took. Amazon's language. A 25% ACoS means that of every 100 sold, 25 went to ads.
What it cost you to win one sale. The most intuitive metric, because it compares directly against your margin.
What you pay per visit. It does not tell you whether you make money: it tells you whether the entry price is bearable.
All four are joined by two bridges: average order value (AOV) and conversion rate (CR).
The first line is worth internalising: ROAS and ACoS are inverses. When someone from Amazon says “my ACoS is 20%” and someone from Meta says “my ROAS is 5x”, they are saying exactly the same thing.
And the third line explains why CPC misleads: lowering CPC only improves profitability if conversion does not fall with it. Cheaper, worse traffic can worsen your CPA even as CPC drops.
None of the four says anything on its own. “3x ROAS” is neither good nor bad: it depends on your margin. Break-even is always calculated from the contribution margin.
With a 60 order and a 20.55 contribution margin: maximum CPA 20.55, break-even ROAS 2.92x, maximum ACoS 34.24% and, at a 2% conversion rate, maximum CPC 0.41. Four numbers for the same boundary.
A store on 60% margins can live with a 1.7x ROAS. A store on 15% margins needs 6.7x. Comparing two businesses' ROAS without knowing their margins tells you nothing.
This is where most money gets lost. The ROAS in your Meta, Google or Amazon dashboard divides attributed revenue by spend. And both “revenue” and “attributed” hide decisions:
The useful contrast is three figures side by side: the break-even ROAS from your margin, the reported ROAS from the platform, and the business ROAS: total actual revenue divided by total ad spend for the same period. If reported is 5x and business is 2x, the platform is claiming sales that were arriving anyway.
Rule of thumb: use break-even ROAS to decide how much to bid, reported ROAS to compare campaigns with each other, and business ROAS to decide whether the channel as a whole is worth it.
| Metric | Actual | Break-even | Reading |
|---|---|---|---|
| ROAS | 3.00x | 2.92x | Above it: profitable, but only just |
| ACoS | 33.33% | 34.24% | Below the ceiling: fine |
| CPA | 20.00 | 20.55 | 0.55 of headroom per sale |
| CPC (CR 2%) | 0.40 | 0.41 | One cent of slack in the bid |
All four rows say the same thing: the campaign makes 0.55 per sale, 27.50 in total. It invoiced 3,000 and earned under 30. At those numbers a 3% rise in product cost puts it into loss.
And note: that 27.50 is contribution, not profit. The month's fixed costs have not entered yet.
| Situation | Metric | Why |
|---|---|---|
| Setting a campaign target | ROAS or ACoS | They compare directly against break-even |
| Deciding what to pay for a sale | CPA | It compares in money against your margin: hardest to misread |
| Adjusting keyword bids | CPC | It is the unit you actually bid in |
| Comparing channels | CPA | It does not depend on each channel's average order value |
| Talking to someone from Amazon | ACoS | It is that platform's native language |
| Wide price ranges | ACoS or ROAS | A single CPA cannot serve 15 and 300 baskets |
There is no universally good ROAS. The only valid target is the one that comes from your margin plus the profit you want to keep.
They overlap. The total usually exceeds actual turnover, and budgeting from that sum allocates money by who claims best, not who sells.
Cheaper, lower-quality clicks worsen CPA. CPC is only useful next to the conversion rate.
A 3x can be excellent on one product and ruinous on another. Compare each against its own break-even.
If your pixel sends tax-inclusive amounts, your target ROAS has to be calculated on the same base. Mixing them inflates apparent profitability.
Platforms do not deduct them. At an 8% return rate, your effective CPA per kept sale is noticeably worse than the dashboard says.
Run your numbers
The calculator turns your costs into all four at once: break-even ROAS, maximum CPA, maximum ACoS and maximum CPC.
Calculate my break-even ROASImpossible to say without your margin. At a contribution margin of 25% of price, break-even is 4x: a 4x would be earning exactly nothing. At a 50% margin, break-even is 2x and a 4x is excellent.
They are inverses: ROAS = 100 / ACoS. A 25% ACoS is a 4x ROAS; a 50% ACoS is a 2x ROAS.
It is ad spend over total sales, not just the attributed ones. It shows whether advertising is pulling organic sales along: if ACoS holds steady while TACoS falls, the product is gaining traction on its own.
What matters is consistency. If your platform records the tax-inclusive value, calculate your target ROAS on the same base (the calculator has that option). To analyze real profitability, work tax-exclusive throughout.
You can still get ROAS, ACoS and CPA; only CPC needs it. Conversion is orders divided by sessions, and it is worth measuring per channel: brand traffic converts far better than discovery.
Usually because the average order value of ad-driven sales is lower than your overall average. ROAS compensates with revenue volume, but CPA per sale gets worse. Check AOV segmented by channel.
Find out how much you can pay for a sale or a click without losing money.
Work out how much returns cut from your monthly profit.
Strip out the fees, the tax, the materials, the shipping and the ads.
Work out your profit per unit after fees, fulfillment and advertising.
What each order really leaves you, and how much of that you can spend winning the customer.
What each imported unit really costs once it is sitting in your warehouse, ready to sell.
The order value above which you can give shipping away without losing money.
How much you can afford to pay for a customer without giving up your profit.
How many months it takes to recover what you paid to acquire a customer.
The stock level at which you have to place the order to avoid running out.
How many buffer units you need so a deviation does not leave you out of stock.
What a full warehouse costs you per year, and what share of your inventory that is.
What running out of stock really cost you, in lost contribution and extra spending.
Whether the bundle you are building leaves more money than selling the same items separately.
How far you can cut the price without falling below the margin you want to keep.
How many more units you have to sell so a discount does not leave you worse off.
The price below which you lose money, and the one your target margin needs.
How many days pass between paying your supplier and collecting from your customer.
Whether the supplier's minimum order pays off once you have paid to hold it.
Which sales channel leaves most per order for the same product and its real costs.