The ROAS formula, break-even ROAS, and why your dashboard disagrees with your bank

The ROAS formula

The ROAS formula is revenue attributed to ads divided by what you spent on those ads:

ROAS = attributed revenue ÷ ad spend

Google's glossary defines it the same way, as total conversion value divided by total spend, and shows it as a percentage. A ROAS of 4 and a ROAS of 400% are the same number: four dollars of attributed revenue per dollar of spend.

Hypothetical example: a month with $20,000 in ad spend and $70,000 in attributed revenue has a ROAS of 3.5 (350%).

Keep it separate from ROI. Google defines ROI as total profit divided by total spend. ROAS uses revenue, which is why a campaign can show a strong ROAS and still lose money. The rest of this article is about closing that gap.

How to calculate ROAS without fooling yourself

The formula is trivial. The inputs are where the mistakes live. Before comparing ROAS across months or channels, decide these once and write them down:

• Revenue: gross or net? Use revenue after discounts. Decide whether tax and shipping charged to the customer are in or out, and keep it consistent.
• Returns. Platform revenue is counted at purchase. If a meaningful share comes back, your real ROAS is lower than the dashboard by roughly that share.
• Spend: media only or fully loaded? Media-only ROAS is what platforms report. If you pay creative production or management fees, a fully loaded version tells you what the channel really costs. Report both, labelled.
• Which revenue? Platform-attributed, or revenue you can tie to the channel in your own store data. They will not match; more on that below.

None of these choices is wrong. Mixing them between reports is.

Break-even ROAS: the number that actually matters

A ROAS figure means nothing until you know the ROAS at which you stop losing money on an order. That is break-even ROAS:

Break-even ROAS = 1 ÷ contribution margin %

Here contribution margin is what is left of each revenue dollar after every variable cost except advertising: product cost, shipping and fulfilment, payment fees and expected returns.

Read it this way: below a ROAS of about 2.19, this hypothetical store loses money on every ad-acquired order. At exactly 2.19 it breaks even on the first order. Above it, it earns contribution.

Target ROAS for a profit goal. If you want to keep, say, 15% of revenue as contribution after ads, subtract that from the margin before inverting:

Target ROAS = 1 ÷ (contribution margin % − desired margin %) = 1 ÷ (0.4575 − 0.15) = 3.25.

The same store at a ROAS of 3.0 is profitable but below its own goal. At 2.0 it is paying to ship product. Two brands with the same ROAS can be in completely different situations because their margins differ, which is why comparing your ROAS to someone else's tells you very little.

What a discount does to break-even ROAS

Promotions are where break-even ROAS gets ignored most often, because the media buyer sees the ROAS and finance sees the margin, and nobody puts them on the same page. Run the same hypothetical store through a 20% sitewide discount:

• Average order value drops from $80 to $64.
• Product cost stays at $28 and shipping at $9: the box costs the same to fill and send.
• Payment fees fall with the price, to about $1.92. Keep the returns allowance at $4.
• Contribution before ads: $64 − $28 − $9 − $1.92 − $4 = $21.08, or about 32.9% of the order.
• Break-even ROAS: 1 ÷ 0.329 ≈ 3.04.

The promotion moved break-even ROAS from 2.19 to about 3.04. A campaign that reports a ROAS of 2.8 during the sale looks fine against last month's target and is losing money on every order. If the discount also lifts conversion rate enough, the trade can still be worth it, but you only know if the new break-even is calculated before the sale starts.

Three other ROAS mistakes we see often:

1. One target for every product. A heavy, low-margin SKU and a light, high-margin SKU need different break-even ROAS. One account-wide target overspends on the first and starves the second.
2. Free shipping not counted. If you absorb shipping, it belongs in the cost lines. Leaving it out flatters contribution margin and lowers break-even ROAS on paper only.
3. Targets never updated. Supplier prices, carrier rates and return rates change. A break-even calculated last year may describe a business you no longer run.

The cost per acquisition formula, and how it connects

The same logic works in cost per order instead of ratios. Google's glossary defines cost per action as total cost divided by total conversions:

CPA = ad spend ÷ conversions

And the break-even version comes straight from the table above:

Break-even CPA = contribution before ads per order = $36.60 in the example.

The two views agree: average order value ÷ break-even ROAS = $80 ÷ 2.19 ≈ $36.60. Use whichever your team reads faster. Media buyers often think in ROAS; finance usually thinks in CPA.

Two refinements, both measured in your own data rather than assumed:

• New versus returning customers. If most ad-attributed orders come from people who already bought from you, the CPA to acquire a new customer is much higher than the blended figure suggests.
• Repeat purchase. If first-time customers reliably reorder, you may choose to accept a first-order CPA above break-even. That is a financing decision. Make it on your measured repeat rate and payback window, never on a hope.

Why platform ROAS is not your real ROAS

Every ad platform reports the revenue it can claim under its own attribution rules. That is useful for comparing campaigns inside one platform and unreliable for deciding how much the business should spend. A few documented reasons:

• Timing. Google's primary conversion columns are reported by the time of the click, not the time of the conversion. A click last week that buys this week is credited to last week, so a recent period can look weaker than it will end up.
• Credit rules. Google lets you choose how much credit each click gets, and models like data-driven attribution produce fractional conversions. Change the model and the reported ROAS changes, although nothing in the business did.
• Overlap. Each platform counts the same order if it touched it. Add up what every channel claims and the total can exceed your store's revenue.
• Demand you already had. Branded search and retargeting often collect orders from people who were going to buy anyway.

The cleanest correction is incrementality. Google's Conversion Lift documentation defines incremental ROAS as incremental conversion value (treatment minus control) divided by ad spend, counting only revenue the ads caused. It is a lower number than platform ROAS, and it is the one that tells you whether to spend the next dollar. We go further into that in blended ROAS vs contribution margin.

A monthly ROAS routine

Put this on the calendar once a month. It takes about an hour once the inputs exist.

1. Recalculate contribution margin from actual COGS, shipping, fees and returns for the month. Margins drift; break-even ROAS drifts with them.
2. Update break-even ROAS and target ROAS from that margin.
3. Compare each channel's platform ROAS against break-even, and note the gap between platform-claimed revenue and store revenue.
4. Check the new-customer share of ad-attributed orders.
5. Decide: scale, hold or cut each channel, against the target, and write down why.

The written reason is the part teams skip and later regret. Three months from now, when someone asks why spend on a channel doubled, "ROAS was 3.4 against a break-even of 2.2 and a target of 3.1" is an answer. "It was doing well" is not. Keep the decisions in the same sheet as the inputs, so the history of your margin and your targets travels together.

To skip the spreadsheet, use the free calculator: enter order value and costs, and it returns your break-even ROAS, target ROAS and break-even CPA. For a brand-level view of the same math, see how we approach ecommerce performance marketing.