Your MER is wrong if you sell cash on delivery

Marketing Efficiency Ratio became popular for a good reason. The ad platforms grade their own homework, they have been grading it generously since 2021, and MER was the answer: ignore what the channels claim, take the revenue the business actually made, divide it by everything you spent on marketing, and look at one honest number.

It is a better number than ROAS. It also has two problems that almost nobody writing about it mentions. The first is embarrassing. The second will quietly bankrupt a store in this region while its dashboard looks healthy.

Problem one: half the industry is using it upside down

Triple Whale, whose analytics product made the term popular, defines it as blended ad spend divided by order revenue, reported as a percentage. Their published 2026 median across their customers is 39%.

Almost everyone else defines it as total revenue divided by total marketing spend, reported as a multiplier. On that definition a MER of 4.0 means four dollars of revenue per dollar spent.

These are reciprocals of each other. A 39% MER in the first convention is a 2.56x MER in the second. Both describe the same business. Neither article tells you the other exists.

So the first rule of MER is unglamorous: when someone quotes you a MER, ask which way up it is. If the number is under 1, it is a cost ratio. If it is above 1, it is a multiplier. We use the multiplier version throughout this article because it is the more common one, and because a number that goes up when things improve is harder to misread in a meeting.

What a good MER is, and why nobody can tell you

There is no universal good MER, and the benchmarks that circulate are mostly vendors publishing medians from their own customer base. Triple Whale's industry medians run from 25% for automotive to 55.5% for health and wellness, on their own data, with no methodology published.

The only benchmark that means anything is the one you calculate yourself:

Break-even MER = 1 ÷ contribution margin

At a 30% contribution margin you need about 3.3x before marketing has paid for itself. At 40%, about 2.5x. This is why "is 4.0 good" is an unanswerable question. For a 40% margin business it is comfortable. For a 25% margin business, where break-even is 4.0x exactly, it is a business running to stand still.

Notice what this means: you cannot use MER without knowing your contribution margin, which means MER does not save you from the harder accounting. It just stops the ad platforms lying to you while you do it.

Problem two: in a cash on delivery market, the revenue number is not real

Here is the part that matters here and does not appear in any of the guides, because they are written for markets where the customer has already paid before the parcel moves.

MER needs a revenue figure. You take it from Shopify, or from your dashboard, and it tells you what was ordered. In a prepaid market that is very close to what was collected, so the distinction does not matter and nobody draws it.

In a cash on delivery market those are different numbers, and the gap is large. An order placed is not revenue. It is a request for revenue, which becomes money only if someone opens the door and pays. Across the Lebanese stores we run, 15 to 25 percent of cash on delivery orders are refused at the door. That is our own figure from operations, not a published statistic, and yours will differ.

So your MER is computed on a revenue number that includes orders which came back. It is overstated by roughly the refused share, every month, consistently, in the direction that makes you feel fine.

What that does to the maths

Worked on an illustrative store, not a client. The assumptions are stated so you can run your own.

A store does $60,000 in placed orders in a month and spends $15,000 on marketing. Reported MER is 4.0x. The dashboard looks healthy.

Sixty percent of its orders are cash on delivery and 20 percent of those are refused, so 12 percent of all orders come back. Delivered revenue is $52,800, and the delivered MER is 3.52x.

If that store runs a 30 percent contribution margin, its break-even MER is 3.33x. It reported 4.0x and comfortable. It is actually at 3.52x, which is thin, and it does not know it.

And that is before the refused orders cost anything. At an average order value of $40 that store placed 1,500 orders, 180 of which came back. At $5 out and $2.50 back, those refusals cost about $1,350 in courier fees alone for revenue that never existed, plus whatever share of the stock returns unsellable. Those costs sit below the MER line and push the break-even ratio up at the same time as the real MER is coming down. The two numbers move toward each other from both directions.

The fix, which is not complicated

Compute MER from delivered revenue. Not orders placed. Money collected, in the period you spent the marketing.

If your platform cannot separate delivered from placed, your courier's settlement statements can, because that is the document that says what actually reached you. It is less convenient than reading a dashboard and it is the only version of the number that is true.

Then keep both. Reported MER tells you how efficiently you are generating demand. Delivered MER tells you how efficiently you are generating money. The gap between them is your refusal problem expressed as a ratio, and watching it move is a better operational signal than either number alone.

The timing problem MER cannot see at all

One more thing, which no ratio fixes. A delivered order is not collected money until your courier settles, and that cycle can run weeks behind the spend that produced the order.

MER is a ratio of two numbers from the same period. It has no opinion about when the cash arrives. A store can post a strong delivered MER every month and still run out of money, because the marketing was paid for in advance and the revenue arrives on the courier's schedule. That is a cash flow question, and it needs a cash flow answer rather than a better ratio.

What to actually do with it

• Work out your contribution margin first, because without it your break-even MER is unknown and the ratio means nothing.
• Compute delivered MER, from money collected, and treat the reported one as a demand signal rather than a financial one.
• Track the gap between reported and delivered. Widening means your refusal rate is moving and you will feel it in cash two months from now.
• Do not set MER as a team target on its own. It is a ratio, and the fastest way to improve a ratio is to stop spending, which also stops the business growing.
• Keep ROAS for the thing it is good at, which is comparing campaigns against each other inside one platform. It is a steering tool, not a scoreboard.

What we do not know

• There is no published MER benchmark for MENA or for cash on delivery markets that we could find. Every benchmark in this article is from a US and European DTC customer base and should not be read across to a store here.
• The vendor benchmarks quoted below come from companies selling analytics products, drawn from their own customers, with no methodology published. They indicate rough magnitude and nothing more.
• Our 15 to 25 percent refusal range is our own operational figure across the Lebanese stores we run. It is not a published statistic and it is not a claim about the market as a whole.

Frequently asked questions

Sources

• Triple Whale, Marketing Efficiency Ratio, for the spend divided by revenue convention reported as a percentage, the 2026 median of 39% across Triple Whale customers, and the industry medians from 25% for automotive to 55.5% for health and wellness. Triple Whale sells the analytics product that popularised the metric and these figures are from its own customer base, with no sample size or methodology published, so they indicate magnitude rather than a benchmark you can be held to.
• Omniconvert, Marketing Efficiency Ratio: what ROAS hides, for the opposite convention of total revenue divided by total marketing spend as a multiplier, the break-even MER formula of 1 divided by contribution margin with the 3.3x at 30% and 2.5x at 40% worked examples, the typical DTC clustering of 3.0x to 5.0x, and the point that platform-reported ROAS overstates results and that a large share of Meta conversions have been modelled rather than observed since iOS 14.5. Omniconvert sells tools in this category, and the statistics it attributes to Polar Analytics, Ask-Luca and Eightx are cited in its article rather than reproduced, so we have not verified them at source and have not repeated their specific percentages as fact here.
• Our own COD versus prepaid analysis, and the calculator methodology, for the cash on delivery share and refusal rate figures used above, with their ranges and how they were derived.

Sources read on 25 October 2026. The two formula conventions, the break-even formula and the vendor benchmarks are sourced. The refusal rate is our own operational figure and is labelled as such. The worked example is modelled on the stated assumptions and describes no real store. The argument that MER must be computed on delivered rather than placed revenue in a cash on delivery market is our own, and we could find no published treatment of it, which is either an opportunity or a sign that we have missed something: if you know of one, tell us and we will link it.