The ROAS Formula, Explained With Real Numbers

A ROAS of 3:1 feels like a win.

It can also be a business quietly losing money on every sale it makes.

The ROAS formula is revenue from advertising divided by the cost of that advertising. Spend $1,000, get $3,000 back, and the ratio reads 3:1, or 300%, or "3x": three labels for the same division problem.

That part takes ten seconds on a phone calculator. Knowing whether the number it produces is actually good takes longer, because a 3:1 return is comfortable for a software company and dangerously close to a loss for a retailer running on 20% margin.

The averages everyone repeats, 2.87:1, "aim for 4x," skip that step entirely. Below is the formula worked with real numbers, the break-even math that actually answers the question, and the benchmarks with their sources attached instead of passed around as folklore.

The ROAS Formula, Worked With Real Numbers

Revenue from ads, divided by the cost of running them. That's the whole formula, and Corporate Finance Institute's January 2025 guide to the metric breaks it into three moves anyone can run this afternoon:

1. Add up every dollar tied to the campaign: media spend, platform fees, and anything the ad platform charged beyond the raw bid.
2. Pull the revenue your analytics platform attributes to that campaign, over the same window the spend covers.
3. Divide revenue by cost. Write the result as a ratio (4:1), a multiple (4x), or a percentage (400%).

Liftoff's April 2026 guide for app marketers makes a good point about those three labels: they all mean the same thing, and mixing them inside one spreadsheet is how a budget meeting turns into an argument about vocabulary instead of money.

Run CFI's own numbers through the formula. Spend $15,000 on a campaign, generate $60,000 in revenue from it, and the ROAS is 4, or 4:1: four dollars back for every dollar spent. Spend $10,000 and generate $11,000, and the ROAS is 1.1.

That second campaign didn't lose money outright. It came close enough that most finance teams would flag it, not celebrate it.

The formula breaks the moment a cost gets left out. CFI's guide names the usual suspects: agency fees, creative production, and platform surcharges rarely make it into the "cost" side of the equation, which inflates every ROAS calculated without them. The true cost of a website visitor by channel is where most of those missing dollars hide. Pull that number before trusting a ROAS that looks too good.

Raw ROAS Lies Until You Divide by Margin

ROAS is not profit. Treating it like profit is the single most expensive habit this metric encourages.

Foundry's April 2026 benchmark report spells out why with one example: $1 spent to generate $4 in revenue nets $0.80 in gross profit at 20% margin, and after subtracting that dollar of ad spend, the campaign is down $0.20. The same 4:1 ROAS at 50% margin nets $2 in gross profit, or $1 after spend.

Identical ROAS. Opposite outcome.

Break-even ROAS is 1 divided by gross margin, expressed as a decimal. At 20% margin, the floor is 5:1. At 30%, it's 3.3:1. At 50%, it's 2:1. At 70% margin, roughly where a SaaS company sits, break-even drops to 1.43:1, according to the same Foundry analysis.

Here's what that looks like on a real campaign. A retailer running 35% margin spends $1,000 on a Facebook campaign and generates $3,000 back: a 3:1 ROAS that reads green on almost any dashboard. Gross profit on that $3,000 is 35%, or $1,050. Subtract the $1,000 it cost to earn it, and the campaign cleared $50.

Fifty dollars of actual profit, dressed up as a "strong" 3:1 return.

Work out the real number this afternoon. Divide 1 by your gross margin, as a decimal. Anything below that result is losing money, whatever colour the ad platform paints it.

What Counts as a Good ROAS (It Depends on Three Things)

Ask five people what a good ROAS looks like and count the different answers.

Triple Whale's guide, last updated in December 2025, reports that the median ROAS across brands using its platform was 2.04 in 2024, less than half the "aim for 4x" figure that gets repeated everywhere. The same guide cites Google's own Economic Impact report putting the average Google Ads ROAS around 2:1, and Amazon's advertising team publishing a similar average with a "good" range of 3 to 4.

None of those numbers agree, and CFI's guide explains why: the floor moves with the business.

Source: Corporate Finance Institute, January 2025.

Neither table means much without the margin sitting next to it. A retailer at 20% margin needs a 5:1 floor just to break even, which makes CFI's "4:1 is strong" advice a losing target for that business and an easy one for a SaaS company clearing 70% margin.

Subscription businesses have a fourth wrinkle: the first invoice is a fraction of what the customer is actually worth. A ROAS read off day one looks weak next to a one-time purchase business, even when the second and third renewal make it the better campaign by month six.

ROAS also only tells half the story on its own. Pairing it with what it actually costs to acquire that customer shows whether a campaign is profitable once, or every time that customer buys again. Chasing the highest reported ROAS instead of the highest reported profit is the same trap cost-effective marketing keeps circling back to: the cheapest-looking number and the most profitable number are rarely the same number.

Average ROAS by Industry and by Channel, With Sources

Benchmarks are only useful once they're specific enough to argue with. Foundry's April 2026 report breaks blended ROAS out by industry. These are blended averages across channels, and they hide real variation inside each row.

Source: Foundry, April 2026, aggregated from Databox, Nest Scale, White Label Agency and Store Growers.

Financial services looks broken until the model behind it gets explained. That industry runs a loss-leader approach: a $500 acquisition cost against a customer worth $15,000 over five years is a strong investment even though the first-touch ROAS sits below 1:1.

Channel matters as much as industry. The same report breaks out average ROAS by ad channel:

Source: Foundry, April 2026.

WebFX's 2025 analysis of paid search campaigns, cited in Liftoff's April 2026 guide, found an average return of 226%, ranging from 70% in financial services up to 686% in heavy equipment: the same industry spread the table above shows, measured a different way.

Find the row that matches the business, then check the channel table against it. If a channel's benchmark clears the break-even number from the section above by a wide margin, that's where the next incremental dollar belongs, not wherever the budget has always gone out of habit.

Why Your ROAS Keeps Sliding This Year

Average ROAS across industries fell 10% year over year in 2026, and the causes are structural, not seasonal, according to Foundry's report.

CPCs rose 10% to 25% across nearly every industry that year. Conversion rates fell 9.28% year over year on Google Ads, declining in 13 of the 14 industries tracked.

Both moving the wrong direction at once compounds the damage. A 15% CPC increase paired with a 10% conversion-rate drop produces roughly a 28% ROAS decline, worse than either factor would do alone.

AI Overviews add a third drag. The same report found that AI-powered search results cut paid click-through rates by 58% to 68% on the queries where they appear. The searches still happen. The click-through rate on them increasingly doesn't.

Some of that decline is attribution loss rather than actual performance loss, cookie limits and cross-device tracking gaps hiding conversions that did happen. The number on the screen is still lower than last year either way, and a budget meeting doesn't care which cause is behind it.

Pull last year's cost per click and conversion rate for the same month before blaming the creative. If CPC climbed double digits while conversion held flat, the auction moved against the account. If conversion dropped while CPC held flat, the landing page did. More of what drives or stalls that traffic sits on the Traffic shelf.

Three Ways to Lift ROAS Without Spending Another Dollar

The ROAS formula has three inputs on the revenue side: traffic, conversion rate, and average order value, all divided by ad spend. Three levers move that number, and only one of them requires spending more.

1. Lift the conversion rate. Ten thousand visits at a 2% conversion rate and an $80 average order convert into $16,000 of revenue on a $1,000 campaign: a 16:1 ROAS. Lift the conversion rate to 3%, change nothing else, and revenue rises to $24,000, a 24:1 ROAS. That's a 50% jump from one number on a landing page, in line with the same 2%-to-3% lift Foundry's report credits with a 50% ROAS improvement at zero additional spend.
2. Raise the average order value. The same math works from a different angle. Push the average order from $80 to $120 at the same traffic and conversion rate, and revenue climbs from $16,000 to $24,000, the identical 50% lift, without touching a bid or a targeting setting.
3. Stop paying a price that resets every month. Both levers above assume the cost side of the formula stays put, and for most paid channels it doesn't. CPCs rose 10% to 25% in 2026 alone, so the same campaign needs a bigger revenue number just to post an identical ROAS next quarter.

A fixed-price placement skips that third problem, because the cost side of the formula is set once and stays there. The Board is a public leaderboard where a brand pays a fixed amount for a rank, disclosed as a paid placement, and that amount never gets re-bid against an auction next month. Taking the top spot right now costs $118, a number that only rises when another brand pays more for it, never because an algorithm decided the same audience got more expensive overnight. Impressions and clicks on every rank are counted and public, which is more than most ad platforms hand over for free.

Test one lever this week. Pull the landing page's current conversion rate, work out what a single point of improvement does to this month's ROAS with the formula above, then go change the page instead of the bid.