Blended ROAS is a comfort blanket with a number on it
Total revenue divided by total spend feels like the honest, platform-agnostic metric. It is also incapable of telling you which channel to cut, hides the performance of everything inside it, and improves when you stop spending on the things that grow you.
Contents
- What blended ROAS actually is
- The baseline problem
- It cannot answer the question you have
- The margin problem nobody adjusts for
- Where it genuinely earns its place
- The three-layer arrangement that actually works
- The window problem
- New customers versus everybody else
- What to do first if this is all you have
- How to present it without misleading anybody
Blended ROAS has become the sophisticated person's metric, and there is a good reason for that. Everybody has noticed that the numbers inside the ad platforms add up to more revenue than the business actually made, and blended ROAS is the obvious response: ignore all of them, take the revenue in the bank, divide by everything you spent. It cannot be gamed by a vendor because no vendor is consulted.
Blended ROAS is unbiased and nearly useless. Platform ROAS is precise and structurally self-serving. Choosing between them is choosing which failure you prefer.
The problem is that businesses do not face a decision called overall marketing. They face decisions about specific channels, specific campaigns and specific budgets, and blended ROAS is silent on every one of them. It is a thermometer in a building with no zone controls: it will tell you the temperature is wrong and nothing about which room to adjust.
What blended ROAS actually is
Blended ROAS is total revenue in a period divided by total marketing spend in that period. It makes no attribution claims, requires no tracking, and cannot be affected by cookie loss, ad blockers, or platform reporting changes. That robustness is its entire appeal.
The robustness is real and worth respecting. Over the last several years a great deal of marketing measurement has quietly degraded — tracking prevention in browsers, mobile platform changes, consent requirements, and the ordinary attrition of identifiers. Metrics built on tracking got worse without announcing it. Blended ROAS did not move, because it never depended on any of that in the first place.
It is also the metric a chief financial officer will reach for unprompted, because it maps onto how they already think: money in, money out, over a period, reconcilable to the accounts. That shared vocabulary is genuinely valuable and is the reason the metric deserves a place rather than dismissal.
What it is not is a management tool for a marketing team. Knowing that the business returned 4.1x on marketing last quarter is interesting; it does not tell you whether to move sixty thousand pounds from display into trade press, which is the actual question sitting on somebody's desk.
The baseline problem
A large share of revenue in most businesses would have occurred without any marketing in that period: repeat purchases, contract renewals, referrals and demand created by previous years of investment. Blended ROAS puts all of it in the numerator, so the ratio measures your history as much as your current spending.
This is the single biggest reason the metric misleads, and it is not a subtle effect. A business with a strong installed base and a decade of brand equity will post an impressive blended ROAS while running mediocre campaigns, because the baseline is doing the work. A young business with excellent marketing and no baseline will look far worse on the same measure.
The consequence is that blended ROAS cannot be compared between companies, and can only be compared within a company over periods where the baseline is stable. Any benchmark you read of the form good businesses achieve a blended ROAS of X is comparing organisations whose baselines differ by an order of magnitude.
It also creates a genuinely perverse short-term incentive. Cutting marketing spend improves blended ROAS immediately, because the denominator drops today and the numerator only drops later. A team under pressure to improve the ratio has a lever available that works within one reporting period and damages the business over three.
It cannot answer the question you have
Allocation decisions are between channels, and blended ROAS has no channel dimension. Two businesses with identical blended ROAS can have completely different underlying distributions, one wasting half its budget and one not, and the metric reports them as identical.
| Channel | Company A spend | A revenue | Company B spend | B revenue |
|---|---|---|---|---|
| Paid search | 100,000 | 600,000 | 100,000 | 900,000 |
| Paid social | 100,000 | 500,000 | 100,000 | 300,000 |
| Display | 100,000 | 400,000 | 100,000 | 60,000 |
| Trade and events | 100,000 | 500,000 | 100,000 | 740,000 |
| Total | 400,000 | 2,000,000 | 400,000 | 2,000,000 |
Both post a blended ROAS of 5.0x. Company B has a hundred thousand pounds in display returning sixty thousand, which is a fire that has been burning for some time, and a trade programme quietly returning 7.4x that deserves more money. Company A is reasonably balanced. Nothing in the blended figure distinguishes them, and nothing in it ever will.
This is why the metric tends to be popular with executives and unpopular with practitioners. It is exactly the right level of abstraction for asking whether marketing overall is working, and exactly the wrong level for doing anything about the answer.
The margin problem nobody adjusts for
ROAS is computed on revenue, not on profit, so a business with mixed margins can improve its ROAS while reducing its contribution. A channel selling high-revenue, low-margin lines will always look better than one selling the reverse.
This matters most in businesses with genuinely different economics across product lines, which is most businesses of any size. Selling an installation package at forty thousand pounds with a twelve percent margin generates more revenue and less profit than selling a service contract at eight thousand with a sixty percent margin, and a revenue-based ratio ranks them in the wrong order without hesitation.
The fix is to compute the ratio on contribution rather than revenue wherever the margin data exists. It is more work, it requires finance to share numbers marketing does not normally see, and it changes conclusions often enough to be worth the friction. Where full margin data is unavailable, even a crude split into two or three margin bands is a substantial improvement on ignoring it.
It is worth being blunt about why this rarely happens: the margin conversation surfaces things some teams would rather not surface, including that the campaigns generating the most impressive revenue numbers are attached to the least profitable products. That is a reason to have the conversation, not to avoid it.
Where it genuinely earns its place
Blended ROAS works as a guardrail: a slow-moving check that the total picture is sane, computed the same way every period, sitting alongside the channel-level numbers rather than replacing them.
Used this way it does something no attribution system can do, which is to catch the case where all your channel numbers look good and the business is not growing. That situation is more common than it should be, and it is the signature of measurement that has drifted away from reality — usually through double counting, or through attributing revenue that would have arrived anyway.
It is also the right number for a board pack, provided it is presented with its inputs. Total spend, total revenue, the ratio, and a note on what is included in spend. Boards are well equipped to interpret a ratio like that and poorly served by channel detail they cannot act on.
The discipline that makes it work is consistency of definition. Blended ROAS computed with agency fees included in one quarter and excluded in the next is not a trend, it is two unrelated numbers on the same axis, and this happens constantly because nobody wrote the definition down.
The three-layer arrangement that actually works
Use blended ROAS as the sanity check, channel-level attribution from your own closed-sales data as the allocation map, and periodic incrementality tests as the audit on the largest channels. Each covers the others' weakness.
Blended ROAS is unbiased but blind. Attribution is granular but depends on data completeness and cannot establish causation. Incrementality establishes causation but is slow, expensive, and answers one narrow question at a time. Running one alone means accepting its specific failure with no correction available.
The layers also fail in different directions, which is what makes the combination useful. If attribution says a channel is performing and incrementality says it is not, you have learned something important. If they agree, you can spend with more confidence than either would justify alone.
CloseRev covers the middle layer: it matches closed sales to your own source records and shows channel-level revenue with the unmatched portion stated plainly. It does not attempt incrementality, and it will not tell you that your blended figure is fine.
The window problem
Blended ROAS divides revenue in a period by spend in the same period, but the revenue and the spend do not correspond to each other. In any business with a sales cycle longer than a few days, this quarter's revenue was substantially caused by last quarter's spending.
The mismatch is usually ignored because it is invisible when spending is flat. Hold budget steady for a year and the lag cancels out — each period inherits roughly what it contributes — and the ratio behaves sensibly. The trouble starts the moment spending changes, which is precisely when somebody is looking at the metric to decide whether the change worked.
Increase spend sharply and blended ROAS falls immediately, because the denominator moves this month and the numerator moves in three. A team that reads that as failure will reverse the increase before the revenue arrives, and will then see the ratio recover, which confirms the wrong conclusion very convincingly. This is one of the most reliable ways to talk a business out of growth.
The mitigation is to lag the comparison deliberately: compare revenue in a period against spend in the period that plausibly produced it, and say that is what you are doing. It is cruder than it sounds and much better than pretending the two columns refer to the same customers.
New customers versus everybody else
Splitting blended ROAS into new-customer revenue and existing-customer revenue removes most of the baseline distortion and turns the metric into something that can actually be trended. It is the single most valuable modification available.
Most marketing spend in most businesses is aimed at acquiring customers, while a large share of revenue comes from customers already acquired. Mixing the two produces a ratio whose movements are dominated by retention dynamics that the marketing budget barely touches. Separating them means the acquisition ratio responds to acquisition spending, which is the relationship you were trying to observe.
It also surfaces something uncomfortable and useful: businesses frequently discover that their acquisition economics are much worse than the blended figure suggested, and that the overall picture was being carried by a base built years ago. That is important information about durability, and the blended metric is specifically designed to conceal it.
The split requires knowing whether each sale was to a new or existing customer, which the sales system already knows. It is one additional column on the export, and it changes the metric from a comfort blanket into a diagnostic.
What to do first if this is all you have
If blended ROAS is your only measurement today, the highest-return next step is not a better ratio. It is producing one channel-level revenue table from your own closed-sales data, once, to find out whether the distribution is even roughly what you assumed.
The reason to do this before anything else is that the answer is usually surprising, and surprising in a way that changes a budget immediately. Businesses reliably discover a channel returning far less than believed and another returning far more, and those two findings together are worth more than a year of refining the overall ratio.
It also settles whether further investment in measurement is worthwhile. If the first channel table comes back looking exactly as expected, you have learned that your instincts are calibrated and you can spend your attention elsewhere with confidence. That is a genuinely valuable outcome and it is rarely the one that occurs.
What you should not do is spend six months building a measurement platform before producing any answer at all. The first table can be made from two exports in a week, and the appetite for everything that follows depends entirely on whether that first table told anybody something they did not know.
How to present it without misleading anybody
Report blended ROAS with its definition, its inputs, and at least one channel-level view beside it. A single ratio presented alone invites the reader to assume it decomposes evenly, which it never does.
The specific harm to avoid is a board concluding that because the blended figure is healthy, the marketing mix is healthy. Those are different claims and the second one does not follow. A business can post a strong overall ratio while a third of its budget returns nothing, and that third is exactly what a board would want to know about.
In practice a single slide handles this: the ratio, the two inputs, the definition in one line, and a small table of channel-level revenue with the unattributed share shown. It takes no longer to read and it removes the inference that causes the damage.
Report the blended number so nobody can accuse you of hiding behind attribution. Report the channel numbers so somebody can actually do something on Monday.
Questions people actually ask
- What is blended ROAS?
- Total revenue divided by total marketing spend across all channels, with no attempt to attribute revenue to a particular source. It is the simplest possible efficiency measure and it is immune to platform self-reporting, because it never asks any platform anything.
- Is blended ROAS better than platform-reported ROAS?
- It is more honest and less useful. Platform ROAS is precise about a claim that is structurally biased in the platform's favour. Blended ROAS is unbiased and cannot tell you anything about any individual channel, which is usually the decision you actually face.
- What is a good blended ROAS?
- There is no universal figure, because it depends entirely on gross margin, cycle length and how much of your revenue comes from existing customers. A 3x blended ROAS is comfortable at seventy percent margin and loss-making at twenty. Any benchmark quoted without a margin is meaningless.
- Why does blended ROAS improve when I cut spend?
- Because a meaningful share of revenue would have happened anyway — repeat customers, brand demand, word of mouth — and that baseline stays in the numerator while the denominator falls. Cutting all marketing produces an infinite blended ROAS and, eventually, no business.
- Should I include brand spend in blended ROAS?
- Include everything you consider marketing investment, and state what is in it. The frequent temptation is to exclude the slow-returning items to make the ratio look better, which turns the metric into a presentation rather than a measurement.
- What should I use alongside blended ROAS?
- Channel-level attribution based on your own closed-sales data, so you can see which sources produced revenue, plus periodic incrementality tests on the largest channels. Blended ROAS is the check; attribution is the map; incrementality is the audit.