In the high-stakes world of digital advertising, Return on Ad Spend (ROAS) has long been the North Star. It is the metric that marketing directors present to board members, the number that agencies plaster on their case studies, and the primary lever that media buyers pull to justify their existence.
On the surface, it is a seductive equation: for every dollar you put into the machine, how many dollars does it spit back out? If the machine gives you four dollars for every one you insert, the logic follows that you should keep feeding it until the bank runs dry.
However, as we move deeper into an era of privacy-first browsing, fragmented user journeys, and algorithmic modeling, this logic is beginning to fracture. The obsession with “Average ROAS” has become a trap; a comfortable delusion that often obscures the true health of a business. The difficulty lies not just in calculating the number, but in trusting it.
When you peel back the layers of attribution windows, view-through conversions, and marginal utility, you realize that Average ROAS is rarely telling you the truth about your growth. It is often telling you exactly what you want to hear, right up until the moment your profitability vanishes.
Table of Contents
The mathematics of deception: average vs. marginal
The single most dangerous misunderstanding in performance marketing is the confusion between Average ROAS and Marginal ROAS. Most dashboards show you the average: total revenue divided by total spend. If you spent £1,000 and made £5,000, your dashboard proudly displays a 5.0 ROAS. This looks healthy. It signals that you should scale up.
The trap is assuming that the next £1,000 you spend will also yield a 5.0 return. In economic reality, returns diminish as volume increases. Your first batch of ad spend usually captures the “low-hanging fruit”; the high-intent customers who were already searching for you or were ready to buy. These conversions are cheap and easy, inflating your initial efficiency. As you scale, you have to work harder to convince “colder” audiences.
Consider a scenario where your first £1,000 generates £5,000 (5.0 ROAS). Excited, you spend a second £1,000, but this tranche only captures harder-to-reach customers, generating just £1,000 in revenue (1.0 ROAS). Your “Average ROAS” for the total £2,000 spend is now 3.0.
A 3.0 ROAS still looks acceptable to most businesses, so you might keep spending. But you shouldn’t. That second tranche of spending was essentially a wash, you broke even on the ad spend and likely lost money once you factor in the cost of goods sold (COGS).
By looking only at the average, you masked the inefficiency of your scaling. You are effectively using the profits from your best customers to subsidize your losses on the worst ones, all while the “Average ROAS” metric tells you everything is fine.
The attribution issue for "Tech Giants"
If the math of averages is tricky, the source of the data is even murkier. We are currently living through an “Attribution Civil War” where the major advertising platforms such as Google, Meta, TikTok are fighting for credit over the same transactions. Each platform operates as a “Walled Garden,” unable (and unwilling) to see what happens outside its borders.
Imagine a customer journey in 2025. A user sees an Instagram ad for a surveyor on Monday but doesn’t click. On Tuesday, they remember the company name and search for it on Google. They click a Google Search ad and browse the site. On Wednesday, they are retargeted by a display ad, and finally, on Friday, they type the URL directly into their browser to purchase a survey.
Who gets the credit? Facebook claims it because the user saw the ad (View-Through Attribution). Google claims it because they clicked the search link. The display network claims it for the retargeting view. If you sum up the revenue reported by each platform’s dashboard, you might see three sales recorded for a single transaction. This “double-counting” inflates your ROAS artificially.
This problem has been exacerbated by the death of the third-party cookie and the introduction of Apple’s App Tracking Transparency (ATT). Platforms can no longer track users across the internet with the precision they once had. To compensate, they have shifted from “deterministic” tracking (knowing exactly who bought what) to “probabilistic” modeling (guessing who likely bought based on data patterns).
When a platform reports a ROAS of 4.0 today, it is partly a calculation and partly a statistical estimation. Trusting this number implicitly is risky because the platform is grading its own homework. It has a financial incentive to show you a high ROAS so that you continue to spend money.
Pro Tip: The "Break-Even" calculation. Don't start spending until you know your "Zero Line."Before you launch any campaign, calculate your Break-Even ROAS for that specific offer. This is the point where you neither make nor lose money. Break- even ROAS = 1/Gross Margin %
The profitability gap: ROAS vs. POAS
Perhaps the most critical oversight in the “Average ROAS” mindset is the disconnect between revenue and profit. ROAS is a revenue metric. It does not care about your margins, your shipping costs, your agency fees, or your overheads. It is entirely possible to have a high ROAS and still go bankrupt.
This is why sophisticated marketers are shifting their focus to POAS (Profit on Ad Spend). Let’s look at a practical example involving two different products. Product A costs £50 and sells for £100 (50% margin). Product B costs £10 and sells for £100 (90% margin).
If you run a campaign for Product A and achieve a ROAS of 2.0 (generating £200 revenue from £100 spend), you have actually made zero profit. The £200 revenue minus £100 ad spend leaves £100, which is entirely eaten up by the £100 cost of goods for the two items sold. However, if you achieve the same ROAS of 2.0 for Product B, you are highly profitable because the cost of goods is so low.
The “Average ROAS” metric flattens these nuances. If your campaigns shift from selling high-margin digital products to low-margin physical goods, your ROAS might stay the same, but your bank balance will shrink. Relying on a blended average across a catalog of products with different margin profiles is a recipe for disaster. You need to know the break-even ROAS for every single SKU or service line you offer. Without this context, “Average ROAS” is just a vanity metric, a number that looks good in a weekly report but fails to pay the bills.
The incrementality conundrum
The final, and perhaps most difficult, aspect of measuring true advertising performance is “Incrementality.” This concept asks a simple but uncomfortable question: Would this customer have bought from us anyway, even if we hadn’t shown them the ad?
Average ROAS is notoriously bad at answering this. It loves to claim credit for “Brand Search” campaigns. If a user types “Trusted Surveyors Wales” into Google, they already know who you are. They are at the bottom of the funnel, credit card in hand. If you run an ad on your own brand name, the ROAS will be sky-high, often 20x or 30x. This inflates your overall account average, making your marketing look incredibly efficient.
But did that ad actually generate new revenue? Likely not. If you had turned the ad off, that user would probably have clicked the organic listing right below it and bought anyway. By mixing this “Brand” ROAS with your “Cold Traffic” ROAS (where you are targeting people who have never heard of you), you create a muddy average that hides the performance of your growth efforts.
Real growth comes from incremental lift, conversions that would not have happened without the ad interaction. Measuring this requires rigorous testing, such as holdout groups (where a segment of your audience is deliberately blocked from seeing ads) to compare their purchase rate against the exposed group. Most small to medium businesses do not have the budget for these scientific tests, leaving them dependent on the flawed “Average ROAS” metric that overvalues remarketing and brand defense while undervaluing the hard work of top-of-funnel discovery.
The solution: Triangulation and MER
If Average ROAS is a trap, what is the alternative? The answer lies in abandoning the search for a single “source of truth” and instead using a method called “triangulation.” This involves looking at three distinct data points to form a complete picture:
-
Platform Data: Use the ROAS reported by Facebook or Google as a directional signal for day-to-day optimization (e.g., is Ad A performing better than Ad B?), but do not treat the revenue number as absolute cash in the bank.
-
Marketing Efficiency Ratio (MER): This is often called “Blended ROAS.” It is the total revenue of the business divided by the total marketing spend from all channels combined. It ignores the squabbling between platforms and looks at the ecosystem as a whole. If your MER is healthy, your marketing mix is working.
-
Customer Feedback: Post-purchase surveys (e.g., “How did you hear about us?”) provide qualitative data that algorithms miss. Often, a customer will click a Google Ad to buy, but tell you they first heard about you on a podcast or from a friend. This “Zero-Party Data” helps you understand where the demand is actually being created, rather than just where it is being captured.
By moving away from a blind reliance on Average ROAS and embracing a more holistic view of contribution margin, incrementality, and blended efficiency, businesses can step out of the trap. They can stop optimizing for a number on a dashboard and start optimizing for the only metric that truly matters: the sustainable growth of bottom-line profit.
