ROAS gets treated as a leaderboard number. Whichever campaign, channel, or SKU shows the highest Return on Ad Spend gets more budget; whichever shows the lowest gets cut. A 4x looks like a win next to a 2x, and budget moves accordingly. That comparison only works if both numbers are being measured against the same floor and they rarely are. Revenue is not profit, and a return multiple that says nothing about margin says nothing about whether the spend actually made money. As with CAC, the raw number matters less than where it sits relative to a calculated line specific to the business. For CAC that line is a ceiling. For ROAS it’s a floor.
What ROAS Actually Measures?
ROAS is revenue generated divided by ad spend. Two blind spots show up constantly:
- Gross ROAS and net ROAS answer different questions. Gross ROAS uses total revenue at the point of sale. Net ROAS strips out returns, discounts, and payment processing costs before dividing by spend. A campaign built around a steep discount code can post an impressive gross ROAS while barely breaking even net of the discount — the spend “worked,” in the sense that revenue came in, without the business keeping much of it.
- Blended ROAS hides margin mix. A channel or campaign selling a mix of high-margin and heavily discounted or low-margin SKUs will show one average ROAS that doesn’t represent either SKU well. This is the same distortion blended CAC creates — an average masking two different economic realities sitting next to each other.
The Floor, Not the Ranking
Every ROAS number has a breakeven point below which the ad spend loses money, and that point is set by gross margin, not by what a “good ROAS” is supposed to look like. The formula is simple: breakeven ROAS = 1 ÷ gross margin. A product line running at 25% gross margin needs a 4x ROAS just to reach breakeven. Meaning a 4x on that line is not a win, it’s the starting line. A product line at 50% margin breaks even at 2x, so a 2.5x there is already generating real profit, ahead of the 4x return sitting on a thinner-margin line.
Ranking campaigns by raw ROAS, without this floor, systematically favors thin-margin lines that need a bigger multiple just to tread water, and can quietly starve a thicker-margin line that’s actually the more profitable place to spend.
Floors Move With the Product, Not the Channel
Because the floor is set by margin, it changes with what’s being sold, not with which channel is selling it. The same channel running two product lines can have two different breakeven points within the same reporting period, exactly as different channels can carry different CAC ceilings depending on the LTV of the customers each one brings in. A single target ROAS applied across an entire account, the way a single CAC target gets applied across channels, makes the same mistake in the other direction. It rewards spend on whichever line has the lowest floor to clear, rather than the line generating the most actual profit per rupee spent.
Case Study: The Channel That Won the Leaderboard and Lost Money
A D2C brand ran the same performance channel across two product lines, a hero SKU sold near full price, and a secondary line that moved primarily through a standing discount code. The discounted line consistently posted a higher gross ROAS and, on a leaderboard view, looked like the better-performing half of the account. Budget was shifted toward it over several months, following the ROAS ranking.
Once margin was applied per line, the picture reversed. The hero SKU’s true margin put its breakeven ROAS well under 2x, meaning nearly all of its return was profit. The discounted line’s margin, after the standing discount, put its breakeven closer to 5x, a level its “higher” ROAS was barely clearing. The budget shift toward the leaderboard winner had been moving spend toward the line with less profit per rupee, not more. Reallocating by floor instead of by raw ROAS improved account-level profitability without changing total spend.
Reading ROAS Correctly
ROAS sits in the Quality dimension of the Cost, Quantity, Quality framework and toward the Transaction stage of the journey-stage framework, both of which treat it as one input among several rather than unpacking it on its own. On its own, a raw ROAS number is a ranking with no reference point. The question worth asking isn’t which campaign has the higher ROAS. It’s which campaign is further above its own floor.
Setting margin-adjusted ROAS floors per product line or channel takes real cost and margin data, not just ad platform reporting. If you’re working through what that looks like for your business, get in touch with 8 Spades.