ROAS is among the most used and most misunderstood metrics in marketing. Agencies put it in reports, platforms calculate it automatically, and business owners compare their figure to someone else's — and most of those comparisons are meaningless. This article covers the correct calculation and where the traps are.
The basic equation
ROAS means return on ad spend, and it's calculated like this:
ROAS = revenue from advertising ÷ ad spend
If you spend SAR 10,000 on ads and generate SAR 40,000 in sales, your ROAS is 4×. Straightforward.
The problem is that this number describes revenue and says nothing at all about profit. And not all revenue is profit.
The number that actually matters: break-even ROAS
Before asking "is 4× a good number?", you need your own break-even point:
Break-even ROAS = 1 ÷ gross margin
Gross margin = (selling price − cost of goods) ÷ selling price.
A worked example
A product you sell for SAR 200 that costs you SAR 120:
- Gross margin = (200 − 120) ÷ 200 = 40%
- Break-even ROAS = 1 ÷ 0.40 = 2.5×
So anything below 2.5× loses money and anything above it makes money. At 4× you genuinely are profitable — just not by the margin you probably imagine.
Now add the forgotten costs
The equation above ignores things you actually pay:
- Shipping and packaging
- Payment gateway fees (typically 2–3%)
- Returns (10–25% in some categories)
- Agency management fees
- Platform and hosting costs
Add these to the example above and suppose they come to SAR 25 per order:
- Real profit per order = 200 − 120 − 25 = SAR 55
- Real margin = 27.5%
- Real break-even ROAS = 3.6×
The gap between 2.5× and 3.6× is the gap between a profitable business and one that believes it's profitable.
The better metric: POAS
Because ROAS is calculated on revenue, we ask clients to track a second figure alongside it:
POAS = net profit ÷ ad spend
Above 1 means profit. Below 1 means loss. No mental break-even arithmetic required each time.
In the example above: SAR 55 profit from an order that cost SAR 50 in advertising gives POAS = 1.1 — barely profitable.
Where the false numbers hide
1. Double counting across platforms
Add up Snapchat, TikTok and Google sales and you'll get a figure larger than your actual revenue. That's expected: each platform claims a sale if the customer saw its ad within the attribution window.
We've seen accounts where the platforms collectively claimed 320% of real sales. The rule: your store data is the source of truth; platform numbers are for relative comparison only.
2. Brand campaigns stealing the credit
A campaign bidding on your own brand name shows a fantastical ROAS — 15× or higher — but those customers were going to buy anyway; they searched for you by name. Don't blend it with acquisition campaigns in one report, or the picture disappears.
3. A pixel sending bad data
If the purchase event double-fires, your displayed ROAS is twice reality. If the pixel doesn't pass order value, the platform weights every sale identically and optimises you toward the cheap ones. These faults are extremely common, and the fix is getting the tracking right before anything else.
4. Ignoring returns
The platform dashboard records the order the moment it completes. A return two weeks later is never deducted. In categories like fashion and footwear, return rates can reach 25% — meaning your real ROAS is a quarter lower than the figure on screen.
Practical steps to take today
- Calculate your real margin after every cost, not just cost of goods.
- Work out your break-even ROAS by dividing 1 by that margin.
- Reconcile platform numbers against your store dashboard for a full month and record the variance as a percentage.
- Separate brand campaigns from acquisition campaigns in every report.
- Track POAS monthly alongside ROAS.
The bottom line
ROAS is a useful metric but an inherently incomplete one. The figure alone, without context, cannot tell you whether you're making money. Know your break-even, clean up your tracking, and treat your store data as the reference. And if your numbers don't reconcile and you don't know where to start, check your tracking first — nine times out of ten that's where the problem is.