What Is ROAS? Formula, Break-Even & Worked Example

Return on Ad Spend (ROAS) is the revenue an advertising campaign generates for every unit of currency you spend on it. The formula is ROAS = Revenue from Ads / Cost of Ads. Spend $3,000 and take $12,000 in revenue, and your ROAS is 4, usually written 4:1. It is the fastest way to tell whether a paid campaign is pulling its weight, and it is one of the few metrics every ad platform reports natively.

It is also the metric most often read wrong. ROAS measures revenue, not profit. On its own it cannot tell you whether a campaign made money, because it knows nothing about your cost of goods, your shipping, your returns, or whether the customer ever buys again. This article covers the formula and a worked example, how ROAS differs from ROI and from Amazon's ACoS, how to derive your own break-even and target ROAS from your margin, and where ROAS stops being useful.

What ROAS is

ROAS answers one narrow question: how efficiently did this campaign turn money into revenue? That narrowness is the point. Because it is scoped to a campaign, an ad set or even a single keyword, it lets you compare like with like, pause what is not working, and shift budget without waiting for a quarterly P&L.

It sits alongside click-through rate, conversion rate, cost per acquisition and ROI in most eCommerce reporting. Each measures a different stage. Cost per acquisition tells you what one customer cost. ROAS tells you what the spend returned. Neither tells you what you kept.

One more distinction worth fixing early, because it is regularly muddled online: ROAS and ROI are not the same metric with different names. ROI measures the profitability of an overall investment. ROAS measures the revenue generated by a specific campaign. They use different arithmetic and they answer different questions, and the section below shows both on the same numbers.

The ROAS formula, worked

Worked example

Last month's campaign had a budget of $3,000. It generated $12,000 in revenue.

ROAS = 12,000 ÷ 3,000 = 4. For every dollar the store spent on that campaign, it took $4 back in revenue.

What belongs in the denominator

The number your ad platform shows uses media spend alone, which is why the platform's ROAS is almost always higher than your real one. A denominator that reflects reality includes:

• Media spend paid to the platform.
• Agency retainers or freelancer fees for the campaign.
• The salary share of the in-house paid media specialist who ran it.
• Creative production: photography, video, copy, design.
• Affiliate commissions and platform selling fees tied to those orders.
• Tooling that exists only to run ads: feed management, bid software, creative tools.

The size of that gap is not trivial, and it varies enormously between advertisers. In a large-scale field experiment on Meta's platforms, Tadelis and co-authors (NBER, 2023) found wide variation in advertisers' ability to generate returns from the same advertising, with patterns consistent with learning by doing and with differences in advertiser sophistication. Two stores buying the same impressions do not get the same ROAS, and neither of them gets the ROAS the dashboard promises.

ROAS vs ROI vs ACoS

Run the same campaign through all three formulas and the differences become obvious. Revenue $12,000, cost $3,000:

Two practical notes. First, ROAS and ACoS carry exactly the same information, so there is nothing to gain by tracking both: if you sell on Amazon and elsewhere, convert with ROAS = 1 / ACoS and standardize on one. TACoS is the genuinely different one, because it puts ad spend against total sales and therefore shows whether advertising is building organic demand or replacing it.

Second, ROI is only as honest as the costs you feed it. If "cost" means media spend, ROI is just ROAS minus one and adds nothing. ROI earns its keep when the denominator is the full investment, and the numerator is contribution rather than revenue. For that comparison in detail, see ROAS vs true profit vs contribution margin.

Break-even ROAS and target ROAS

The margin arithmetic, worked

Say a campaign promotes one product: a newly launched yoga starter kit. It sells for $120 and the cost of goods is $60, so the gross margin is 50 percent.

Break-even ROAS = 1 ÷ 0.5 = 2. To break even, the campaign must return $2 of revenue for every $1 spent, which is the same as saying you cannot spend more than $60 to win a $120 order.

Now suppose the campaign reports a ROAS of 1.6. On the dashboard that looks respectable: $1.60 back for every $1 spent. Do the arithmetic on a single order:

• Order revenue: $120
• Ad cost to win it at a ROAS of 1.6: 120 ÷ 1.6 = $75
• Cost of goods: $60
• Result: 120 − 75 − 60 = −$15

The campaign loses $15 on every order it generates, before shipping, payment fees, returns or overhead. Nothing on the ad platform reports that, because the platform does not know your cost of goods. This is the single most common way stores lose money while reporting a positive ROAS.

Setting a target ROAS

Break-even is the floor, not the goal. Target ROAS is the figure you set above it, and where you set it is a strategic choice rather than a calculation:

1. Calculate the margin on the products you actually advertise. Not the blended store margin. If your ads push a discounted entry product, use that product's margin. Our guide to calculating eCommerce profit margin covers the inputs.
2. Divide 1 by that margin. This is your break-even ROAS, and it is the number every campaign must clear on a single-order basis.
3. Add the operating costs advertising has to cover. Shipping subsidies, payment processing, expected return rate, and the share of overhead you want paid ads to carry. Each one raises the real floor.
4. Decide what the campaign is for. A young store that needs cash sets a target well above break-even. An established brand buying market share can run closer to it, deliberately and with a limit.
5. Split the target by campaign job. Prospecting to new customers, retargeting, and branded search do not deserve the same target. A single blended target quietly subsidizes the easy campaigns with the hard ones.

Two situations make this harder, and it is worth naming them rather than pretending otherwise. A campaign promoting a mixed basket of products has no single margin, so you need order-level cost of goods to get a true figure. And a brand awareness campaign will not produce the order in the session, so the ROAS you see depends entirely on your attribution window and model, not on the campaign.

What counts as a good ROAS

The honest answer to "is 4 a good ROAS?" is that it depends on your margin, your category, your customer's repeat behavior and what the campaign is for. A ROAS of 4 is excellent at a 20 percent margin and unremarkable at an 80 percent margin. This is why comparing yourself to another store's reported ROAS tells you nothing: you are comparing two numerators against two entirely different denominators of cost.

Your own break-even also moves. It changes when supplier costs change, when you discount, when shipping subsidies change, when the product mix in a campaign shifts, and when return rates rise. Treat break-even ROAS as a number you recalculate each quarter, not one you set once.

"No one should be doing manual bidding. In the last few years, ROAS bidding and automated bid systems have become so good that users should not manually bid, especially on thousands of SKUs." — Brad Geddes, Co-founder at Adalysis

That point has only got stronger. Most platform bidding is now target-based rather than manual, which moves the work from setting bids to setting the right target. Which brings the whole thing back to the margin arithmetic above: an automated bidding strategy will faithfully chase whatever target you give it, including a wrong one.

Why ROAS alone misleads

Consider two campaigns with an identical ROAS of 2.5. The first acquires customers who buy once and never return. The second acquires customers who go on to buy three more times over the following year. The ad platform reports them as equally good. They are not remotely equally good, and a store that reallocates budget on ROAS alone will keep funding the wrong one.

The distortion runs both ways:

The fix is not a better ad metric. It is connecting acquisition to what happens afterwards. Once you can see, per campaign, how many acquired customers came back and what they were worth, the target ROAS stops being one number and becomes a set of them: lower for the campaigns that reliably produce repeat buyers, higher for the ones that produce a single discounted order and nothing else.

This is what a customer intelligence platform is for. Nexus by Omniconvert segments customers with RFM (recency, frequency, monetary), tracks cohorts by acquisition source, and predicts customer lifetime value, so you can compare acquisition campaigns on the whole relationship rather than the first checkout. It also works in the other direction: high-value segments can be pushed back to Meta Ads and Google Ads as source audiences, which is the most reliable way to raise ROAS on prospecting, since the targeting improves rather than the bid.

How to improve a low ROAS

When ROAS drops, look at what fell with it. Low click-through rate points at targeting, offer or creative. Healthy click-through with a low conversion rate points at the landing page or the checkout. Changing bids first, which is the common reflex, does nothing for either.

Build audiences from your own customer data

Your customer data is the most valuable targeting asset you have, and it is the one your competitors cannot copy. Use your best existing customers as the seed for lookalike audiences on acquisition campaigns. Build custom audiences to re-engage loyal customers, to reach dormant ones before they churn, and to push newly acquired customers toward a second purchase. Segment-based audiences beat interest-based targeting for exactly the reason ROAS misses: they are built on what people actually bought.

Fix the landing page before you raise the bid

Paid traffic sent to a mediocre landing page is money handed back to the platform. If click-through is high and conversion is low, the creative did its job and the page did not. Test the headline, the offer framing, the proof, and the mobile layout with Omniconvert Explore, which has run more than 70,000 experiments across 7,000+ websites. Conversion rate sits directly in the ROAS numerator, so an uplift there raises ROAS without a cent of extra spend.

Take the friction out of checkout

Cart abandonment is where paid traffic goes to die. Offer multiple payment options, allow guest checkout, and show delivery costs, taxes and fees before the final step rather than at it. Unexpected costs revealed late are a leading reason people abandon a checkout they had every intention of completing.

Then look past the campaign

A ROAS problem often turns out to be a retention problem wearing a disguise. If repeat purchase rates are low, every acquisition campaign has to pay for itself on a single order, which is the hardest possible test. Raising repeat rate lowers the ROAS every campaign needs to clear. For the tactical layer, our guide to increasing ROAS covers 24 specific tactics across ads, landing pages and lifetime value, and reducing customer acquisition cost attacks the same problem from the cost side.

Frequently Asked Questions