Deividas here. ROAS is the number almost every Meta advertiser watches, and it's also the number that quietly misleads more brands than any other metric on the dashboard. This post covers what it actually means, how to calculate it correctly, and when to stop trusting it.
ROAS stands for Return on Ad Spend. It measures how much revenue a campaign generates for every pound or dollar spent on advertising. The formula is straightforward: total revenue attributed to the ad campaign divided by the total ad spend for that campaign. A campaign that spends £10,000 and generates £35,000 in attributed revenue has a ROAS of 3.5x.
That number tells you something, but not as much as it looks like it does. ROAS is a ratio of revenue to spend — it says nothing about whether that revenue was profitable, whether those customers would have found you anyway, or whether the attribution model is giving the campaign credit it doesn't deserve.
There is no single good ROAS because it depends entirely on your margins. A brand with 60% gross margins can operate at a 2x ROAS and be healthy. A brand running 20% margins needs a 5x ROAS just to cover cost of goods before overheads. The number that matters is your break-even ROAS — the point at which ad revenue covers ad spend plus cost of goods. Everything above that is contribution to overhead and profit.
Across the e-commerce brands we work with at Triple Scale, a blended ROAS of 3–4x is the range where healthy accounts tend to operate. Below 2x is a flag worth investigating. Above 5x on a scaling account often means you're spending too conservatively and leaving growth behind.
The biggest problem with ROAS is that it measures revenue attributed to ads, not revenue generated by ads. Meta's default attribution window (7-day click, 1-day view) captures a lot of purchases that would have happened regardless — returning customers who already knew the brand, email subscribers who clicked a retargeting ad before buying, organic searchers who happened to see an ad in the same window. The ad gets the credit; the ROAS looks great; the account isn't performing as well as the number suggests.
The second problem is that blended ROAS treats new customers and repeat purchasers identically. A £30,000 campaign result looks the same whether it came from 300 new customers at £100 each or 300 returning customers who were already going to buy. For a brand building long-term growth, those outcomes are completely different in value.
New Customer ROAS (NC ROAS) strips out repeat purchasers and measures the return from first-time buyers only. It's harder to pull from the standard dashboard — you need to set up custom conversions or use a third-party attribution tool — but it tells you something blended ROAS cannot: whether your prospecting is working.
At Triple Scale, we treat an NC ROAS of 2x as the floor for healthy prospecting. That sounds low against a blended ROAS of 3–4x, but it makes sense when you account for LTV. If your average customer buys three times over their lifetime and each purchase carries the same margin, acquiring them at 2x ROAS on the first sale still generates strong lifetime profit. Optimising purely for blended ROAS often means pulling budget from prospecting that looks "inefficient" but is actually building your most valuable long-term revenue.
Marketing Efficiency Ratio (MER) is total business revenue divided by total ad spend — across all channels, all campaigns, everything. It's the clearest view of how efficiently your entire marketing engine is working, because it's immune to attribution fraud and cross-channel credit disputes. A brand with £500K in revenue and £100K in total ad spend has an MER of 5x, regardless of what any individual campaign's ROAS says.
We target MER in the 3–4x range for the brands we manage. When MER is healthy and blended ROAS looks low, the attribution model may be undercounting. When ROAS is healthy and MER is weak, there's a channel cost or organic contribution problem worth investigating. ROAS and MER together give you a picture that neither can provide alone. For a full breakdown of how MER works and how to set targets, read the MER guide here.
No attribution model is perfect. Meta's native attribution overcounts. Google Analytics undercounts. Third-party tools like triple attribution or MMM (media mix modelling) get you closer but still involve assumptions. The practical answer is to use multiple signals in parallel: platform ROAS as a directional in-account metric, NC ROAS for prospecting health, MER as the business-level truth, and week-over-week revenue trends as a sanity check on all of them.
Don't optimise for one number in isolation. ROAS is a useful input to the dashboard, not the output that determines decisions.
Use ROAS for: comparing relative performance between campaigns or ad sets within a single account, over a consistent attribution window. It's a valid directional signal when the conditions are held constant. Don't use it for: deciding whether to scale overall ad spend, evaluating the profitability of a campaign, or comparing performance across brands with different margin structures.
The brands that scale well know their break-even ROAS, track NC ROAS separately, and check MER weekly. ROAS is one of several inputs — not the metric that runs the account.
For the full measurement framework we run at Triple Scale, see how to read your Meta ads reporting numbers, and what a good Facebook ads conversion rate looks like alongside your ROAS targets.
