What return on ad spend tells you — and what it hides
ROAS expresses revenue per dollar of media. A 4.0 means every dollar into the platform came back as four dollars of tracked revenue. It is the fastest read on whether a campaign is working, and it is also the metric most likely to be quoted without the context that makes it meaningful.
The hidden variable is gross margin. ROAS counts revenue, not profit, so the same 3.0 is comfortable for software and fatal for a business reselling hardware at a 25% margin. Two companies can report identical ROAS and sit on opposite sides of profitability. That is why the break-even calculation below is the part of this page worth bookmarking.
Break-even ROAS: the only benchmark that is yours
Break-even ROAS is simply 1 ÷ gross margin. Run it once and you have a floor that no industry average can give you:
- 25% margin → you need 4.0x before you have covered cost of goods.
- 40% margin → 2.5x.
- 70% margin → 1.43x.
Note what this implies: a high-margin business can profitably run at a ROAS that would bankrupt a low-margin one. When a competitor appears to be outbidding you irrationally, this is usually the explanation — they are not being irrational, they are working from a different floor.
Planning with a target ROAS: revenue and budget
ROAS is usually quoted backwards — a number reported after the quarter closed. The same formula is far more useful before it opens, because it turns a target into the two figures a plan actually needs:
- Revenue needed = ad spend × target ROAS. $10,000 of media at a 4.0 target has to return $40,000.
- Spend supported = revenue ÷ target ROAS. A $200,000 revenue goal at a 4.0 target allows $50,000 of media.
Read those two together and the efficiency-versus-volume trade-off stops being abstract. Hold the revenue goal fixed and every increase in target ROAS shrinks the budget you are allowed to spend to reach it — a 6.0 target against that same $200,000 goal permits only $33,000 of media. Teams routinely set a high target and a high growth number in the same meeting without noticing they have asked for both.
The check that matters most takes one number. Run your target against break-even before anything is booked: at a 25% margin, break-even is 4.0x, so a 3.0 target is under water before the first impression is served. A target below break-even does not fail slowly — it loses more money the better the campaign performs.
Is a 3x ROAS good? Only your margin can answer that
It is the most common question about this metric and it has no general answer, which is why published benchmarks mislead so reliably. A 3.0x is comfortable at a 40% margin, precisely break-even at 33%, and a steady loss at 25%. The grid below runs that arithmetic across the range most businesses sit in, so you can find your own row rather than borrow somebody else’s average.
| Gross margin | 1.5x | 2.0x | 2.5x | 3.0x | 4.0x | 5.0x | 6.0x |
|---|---|---|---|---|---|---|---|
| 10%(10.00x to break even) | −$850 | −$800 | −$750 | −$700 | −$600 | −$500 | −$400 |
| 20%(5.00x to break even) | −$700 | −$600 | −$500 | −$400 | −$200 | $0 | +$200 |
| 25%(4.00x to break even) | −$625 | −$500 | −$375 | −$250 | $0 | +$250 | +$500 |
| 30%(3.33x to break even) | −$550 | −$400 | −$250 | −$100 | +$200 | +$500 | +$800 |
| 40%(2.50x to break even) | −$400 | −$200 | $0 | +$200 | +$600 | +$1,000 | +$1,400 |
| 50%(2.00x to break even) | −$250 | $0 | +$250 | +$500 | +$1,000 | +$1,500 | +$2,000 |
| 60%(1.67x to break even) | −$100 | +$200 | +$500 | +$800 | +$1,400 | +$2,000 | +$2,600 |
| 70%(1.43x to break even) | +$50 | +$400 | +$750 | +$1,100 | +$1,800 | +$2,500 | +$3,200 |
Platform ROAS versus blended ROAS
Every ad platform is incentivised to claim credit for the same sale, and modern attribution windows let them. Add up the revenue Meta, Google, and TikTok each report and you will frequently exceed what the business actually booked — sometimes by a wide margin.
Blended ROAS — total revenue ÷ total media spend, across everything — cannot be double-counted. It is a blunter instrument and it will look worse than any individual platform number, which is precisely why it is the one to steer by. Use platform ROAS to decide what to change inside an account; use blended ROAS to decide how much money the channel gets.
Why the highest ROAS is rarely the right target
ROAS is trivially maximised: spend only on people already searching your brand name. You will post a spectacular number and grow very little. The trade-off between efficiency and volume is real, and the correct point on that curve is wherever marginal contribution is still positive — not wherever the ratio peaks.
Because of this, ROAS is a poor sole input to a budget decision. Pair it with absolute new-customer volume and with marketing ROI, which works in profit rather than revenue and accounts for the costs sitting outside the ad account.
Improving a ROAS you control
Ranked roughly by how often they move the number in practice: fix conversion tracking first, because a ROAS built on bad data cannot be optimised. Then work average order value — bundles, thresholds, and post-purchase offers raise the numerator without touching the auction. Then creative, which drives the click-through rate that in turn lowers your cost per click. Bid and budget tuning comes last; it is the lever most teams reach for first and the one with the least headroom.