MER vs ROAS vs Profit

Three metrics, endlessly debated, and the debate is usually about the wrong thing. Each answers a real question. None of them answers “did we make money”, which is generally the question being asked.

What each one measures

ROAS divides attributed revenue by channel spend. Its input is the platform’s own view of which purchases it caused, inside its own click and view windows.

MER divides total store revenue by total advertising spend. No attribution model is involved. Both numbers come from your own records.

Profit after advertising takes contribution margin for the period and subtracts total advertising cost. Both inputs are money, and the output is money.

They form a sequence, each fixing a flaw in the one before and introducing a different limitation.

What MER fixes

ROAS has a structural problem: platforms attribute independently and generously, and none deducts what another claimed. Sum them and you routinely get more revenue than the store actually took — $14,800 more in the example above, on a month where net sales were $111,500.

MER removes that entirely. Total revenue is a fact from your own admin. Total spend is a fact from billing. The ratio cannot be inflated by an attribution window, cannot be double-counted across channels, and does not change when a platform updates its methodology.

For deciding whether total spending is sustainable, this makes MER strictly better than blended ROAS.

What MER still cannot tell you

Two things, and they matter.

It includes revenue advertising did not buy. Organic search, direct traffic, email to existing customers, repeat purchases. All of it sits in the numerator. A store with strong repeat business has a healthy MER partly because of customers it acquired two years ago.

This creates a specific trap: MER improves as the returning-customer base grows, so advertising can be getting steadily less efficient while the ratio holds steady or rises. The metric moves for reasons that have nothing to do with the spending it is supposed to judge.

It does not know your margin. MER is a revenue ratio. A 4.5x MER at a 45% contribution margin is comfortable. The same 4.5x at a 22% margin is losing money. Nothing in the ratio distinguishes the two.

Where each belongs

ROAS — inside a single channel, comparing creatives, audiences or campaigns against each other. The attribution bias is at least consistent within an account, so relative comparisons hold even when absolute figures do not.

MER — as a top-line health check on total spending. Watch the trend rather than the level, and watch it alongside the share of revenue coming from new customers, because that is what tells you whether an improving MER is real.

Profit after advertising — for the decision about whether to spend more. It is the only one of the three denominated in money that has actually paid for the goods.

The number that connects them

Any revenue ratio can be converted into a profitability judgement with one input: contribution margin rate.

Break-even ROAS is one divided by that rate. At 45% you break even at 2.2x; at 25% you need 4.0x. The same conversion works for MER — a target MER is just the ratio at which contribution margin covers advertising plus fixed costs.

This is why the ratio debate is usually beside the point. A store that knows its contribution margin rate can make any of these metrics useful. A store that does not cannot make any of them useful, and will spend its energy arguing about attribution windows instead.

The fourth option nobody lists

There is a metric missing from the usual three, and it is frequently the most useful of the set: contribution margin per order against acquisition cost per new customer.

Unlike ROAS and MER it is not a ratio of revenue to spend. Both sides are denominated in money that has already paid for the goods, the shipping and the fees. If margin per order is $45 and a customer costs $38, the order works, and no attribution model is required to interpret that.

It also scales down cleanly in a way period-level ratios do not. A single month’s MER is affected by seasonality, by a stock-out, by a viral post. Margin against acquisition cost is a property of the unit economics, and it moves only when something structural changes — pricing, product mix, delivery costs, or what the market charges for a customer.

The reason it appears less often is mundane: it needs cost data at order level, and ratios need only revenue and spend. That is the entire reason stores default to arguing about attribution — the easier metrics are the ones available without doing the costing work first.

A practical arrangement

Run all three, each for its own job, and keep them on one screen.

MER weekly, with new-customer share beside it, so an improving ratio can be read correctly.

Channel ROAS weekly, used only for within-channel comparisons and never against a fixed target imported from someone else’s business.

Profit after advertising monthly, alongside advertising as a percentage of contribution margin. That percentage is the most decision-ready number of the set: it says how much of every marginal dollar earned is going to acquisition, and therefore how much is left to carry everything else.

When the ratios look stable and that percentage is climbing, the ratios are measuring something that has stopped mattering.