I run three channels for one ecommerce client, and Meta’s reported return on ad spend sits anywhere between 1.11 and 1.99 depending on which cut of the account you read. Not one of those figures tells the owner whether the business made money last month. A marketing efficiency ratio does something different: it divides total revenue by total marketing spend and ignores which platform claimed the sale. That figure is harder to argue with, because nothing inside it depends on an attribution window.
This is for you if two dashboards disagree and somebody has to pick one.
What the marketing efficiency ratio actually measures
Total revenue for a period, divided by total marketing spend for the same period. That is the whole of the marketing efficiency ratio, and MER is how most people abbreviate it.
No pixel, no attribution model, no conversion window. Both inputs come from places that are hard to fake: the store’s own revenue report and the invoices you paid.
Some finance teams run it upside down, as spend expressed as a percentage of revenue. Same information, different habit. Pick one and stay with it, because half of every argument about this metric is two people quoting reciprocals at each other.
Scope is what makes it useful. Platform ROAS answers a narrow question: what did this channel claim it caused? The account level version answers a wider one: for every dollar the business spent on marketing, how much revenue arrived, from anywhere at all.
Those are different questions, and an account can look healthy on the first while the second quietly slides.
Why platform ROAS and the marketing efficiency ratio disagree
They count different revenue and different spend, so agreement was never on the cards.
Platform ROAS counts revenue the platform believes it caused, inside the window it chose, for the spend that ran through it. Your bank counts everything: paid, organic, email, direct, and repeat buyers who needed no advert at all.
So blended ROAS usually lands below the sum of what the platforms claim, and it moves whenever non-paid revenue moves. A good month for email flatters it. A press mention flatters it too.
That gap is not always a platform overstating its case. Haus, publishing on 13 August 2025 across 640 Meta experiments, found that for every $100 in platform-attributed DTC revenue Meta generated $115 in incremental revenue. Haus sells geo experimentation, so weigh their interest, but the direction of that finding cuts against the usual assumption.
Both numbers can be wrong in opposite directions at once. Only one of them is auditable.
The account that made me stop quoting one channel figure
A US flag and banner ecommerce client, Meta sales campaigns, June to August 2026.
Four defensible ROAS figures came out of the same three months of data. Mixed attribution windows, as Meta blends them: 1.11. Every campaign including awareness: 1.25. All six sales ad sets: 1.58, from 64 purchases at $67.24 each. Sales ad sets with the failed state test removed: 1.99, from 61 purchases at $53.72.
None of those is a lie. Each answers a slightly different question, and I could defend any one of them in a meeting.
I also run Google Ads and Klaviyo for the same client. So even the most flattering Meta figure describes roughly a third of their marketing, and the owner has never once asked me what Meta thinks its own campaigns did. The question I actually get asked is simpler: for everything we spent, what came back?
What a marketing efficiency ratio cannot tell you
It cannot tell you which ad set to switch off, and that limitation gets expensive.
One ad set in that flag account targeted a neighbouring state. It returned 0.25, three purchases at roughly $341 each, and about $1,027 went into it before I stopped it.
Now the part that should worry anyone steering by a single account level number. That ad set had the highest reach in the account, 17,086 people, and the most link clicks, 542. Every traffic metric said it was the winner. Revenue said it was the only loser.
Across a three month period, $1,027 of waste barely moves the top-line ratio. The account-wide figure stays respectable while a slice of the budget goes nowhere.
Steer the business by the ratio. Diagnose the account one ad set at a time.
Email changes the number without changing the ads
Klaviyo is email, not an advertising platform, and that distinction does more damage to this metric than anything else.
Send a well-timed flow to an existing list and revenue rises with almost no matching spend. The ratio improves. Meanwhile the ad account is identical to how it was yesterday.
That is not cheating. It is genuinely better economics, and it is one of the few reliable ways to improve the number. It does mean the figure has to be quoted with its denominator spelled out.
My rule is to include everything that counts as marketing: media spend, email tooling, agency or freelancer fees. If a line is excluded, name it in the same sentence as the figure. A marketing efficiency ratio is only ever as honest as the cost list sitting behind it.
This is also how one account went from $10,000 to $50,000 a month without the Meta budget doing the heavy lifting. More objectives, awareness campaigns, custom audiences and lookalikes, then channels the client was not using, including TikTok and owned email.
Setting a target when nobody will tell you the margin
Most ecommerce owners will not hand over a gross margin figure, and the flag client never has.
Across my ecommerce clients gross margin runs between 20% and 50%. That puts break-even return on ad spend somewhere between 2.0x and 5.0x, a range wide enough to be useless as a target for any single business.
Without margin, an absolute target is guesswork. What still works is direction: this month against last month, at equal or higher revenue.
A figure of 3.0 means nothing on its own. Falling from 3.4 to 2.6 while revenue stayed flat means the marketing got more expensive, and that conclusion needs no margin disclosure at all.
Ask for margin once, in writing. If it never arrives, say plainly that you are optimising a proxy, and put that sentence in the report rather than in your head.
How to read the marketing efficiency ratio week to week
Weekly readings are noise. A rolling 28 day figure is the shortest window I trust on a mid-sized ecommerce account.
Read it beside revenue, never alone. Rising efficiency on falling revenue is a business contracting, and on a chart that looks identical to a business getting better at buying.
Log every structural change with its date. Adding a channel raises the denominator immediately and the numerator weeks later, so the number dips before it climbs, and somebody will panic in the gap.
One more habit repays the trouble: note which cut of the data you quoted last time. Moving from “Meta sales campaigns” to “all Meta campaigns” changed the answer by 0.33 in the account above, and nobody in the meeting will spot that the definition moved.
Where the published benchmarks stop being useful
Two sources are worth knowing, and both arrive with their limits attached.
IRP Commerce reports UK and EU ecommerce running at a cost of acquisition of 9.34% of revenue in June 2026, alongside a 2.03% conversion rate and a £127.06 average order value. IRP describes its own numbers as informed estimates rather than audited values, built on last-click data from SME and mid-market stores covering roughly 30% of the UK market.
Triple Whale puts platform median return on ad spend at 2.04 for 2024, with Meta accounting for 68.3% of the ad spend it tracks. That 2.04 carries no stated sample size or date range on the page where it appears, which is worth saying out loud before anyone builds a target on it.
Use both to bound a conversation. Neither one tells you what marketing efficiency ratio your own business should be hitting.
What to change this week
Three things, in this order.
Calculate one marketing efficiency ratio for the last 28 days: total revenue divided by total marketing spend. Then write down exactly what you included, especially email tooling and fees.
Then compare that against the previous 28 days at a similar revenue level, and treat the direction as the signal rather than the value itself.
Last, open the ad sets underneath it and find the one with the best traffic metrics and the worst revenue. Almost every account has one, and no account level figure will ever point at it.
Cost per lead by channel covers the same comparison problem in lead generation, the two purchase costs I can date and evidence sets out the volumes behind these numbers, published return tables and where they come from explains why industry averages will not help you here, and reconciling blended against platform figures walks the same flag account in more detail. Search usually carries the recoverable revenue in accounts like this, which is what my Google Ads work is for. To have your own numbers reconciled properly, book a teardown.