Use ROAS when
You are steering inside one channel: bids, audiences, creatives, keywords. Label it every time with the platform and the attribution window.
ROAS vs MER vs POAS: which number for which decision
ROAS vs MER vs POAS depends on the decision. Use ROAS to steer campaigns inside a channel. Use MER to see whether marketing as a whole is getting more or less efficient. Use a profit-based measure such as POAS to decide whether growth is worth paying for. Write down which number each decision uses.
Analytics & measurement
Most arguments about these three are really arguments about which decision is on the table. Settle that first and the metric usually picks itself.
What does each number count?
ROAS (return on ad spend). Conversion value attributed to ads, divided by ad spend. Vendor glossaries define it this way12. "Attributed" is the key word: the platform decides what counts.
MER (marketing efficiency ratio). Total revenue divided by total marketing spend, with no attempt to split credit between channels2. Sources disagree on the denominator. One vendor's MER definition divides by media spend rather than all marketing3.
Blended ROAS. Often used as another name for MER: Shopify's blog says MER is also called blended ROAS4. Our convention keeps them apart. Blended ROAS is total revenue over paid media spend across all platforms. MER is total revenue over all marketing spend, including things like creative production, tools and agency costs. The only difference is the denominator, so say which one you mean.
POAS (profit on ad spend). Profit from ad-attributed sales, divided by ad spend. Ask which costs were taken off to get to "profit" before you trust the figure.
ROAS vs MER vs POAS, side by side
Metric
Formula
Question it answers
Decision it suits
What it hides
ROAS
Attributed revenue ÷ ad spend
How much revenue did this platform credit to these ads?
Steering bids, campaigns and creatives inside a channel
Costs, returns and sales that would have happened anyway
Blended ROAS (our convention)
Total revenue ÷ paid media spend
Is paid media, taken together, getting more or less efficient?
Direction of the paid media budget
Which channel is doing the work; non-media marketing costs
MER
Total revenue ÷ total marketing spend
Is marketing as a whole getting more or less efficient?
Overall budget direction, month by month
Which channel to cut; seasonality and promotions move it too
POAS
Ad-attributed profit ÷ ad spend
Do the sales these ads were credited with leave profit?
Whether a campaign or product line is worth scaling
Still attributed, so it inherits attribution's bias
No row is the best metric. Each one is a good answer to one question.
Why is ROAS not enough?
ROAS tells you how much revenue a platform credited itself with. It does not tell you what was left after discounts, delivery and returns. It does not tell you what would have sold anyway.
A measurement vendor argues that platform ROAS can overstate some channels and understate others5. Either way, the number is the platform's opinion of the platform.
There is also the margin problem. A Shopify Community thread asks why an ad with a ROAS of 3.8 can still lose money6. The short answer: revenue is not profit.
Break-even ROAS is the ROAS at which you stop losing money. A vendor guide from Polar Analytics puts it at 1 divided by gross margin1. Our view is that contribution margin before marketing is the better base: what is left per order after product, delivery, payment fees, returns and discounts, before any ad spend. So, in our convention, which is not an industry standard, break-even ROAS is 1 divided by contribution margin before marketing.
Example (illustrative). One month: total revenue €200,000. Paid media spend €40,000. All marketing spend €50,000. Blended ROAS is 5.0. MER is 4.0. Meta and Google together claim €220,000 of attributed revenue, more than the shop sold. Contribution margin before marketing is 30 per cent, so break-even ROAS is about 3.3. A campaign at a ROAS of 3.0 looks fine in the panel and loses money on the orders it brings.
Check your own break-even point with the Break-even ROAS & contribution margin calculator.
Does MER solve attribution?
No. MER tells you whether marketing as a whole is getting more or less efficient. It can't tell you which channel to cut. Useful number, wrong job.
In a 2021 PPC practitioners' chat on MER vs ROAS, one participant described moving clients to MER for overall direction while setting channel-level ROAS targets7. That matches our view. MER is also easy to flatter. A big promotion or a seasonal peak lifts it without marketing doing anything cleverer.
Here is a question worth asking in your next review: what is your MER month by month, and does it move when channel ROAS says it should? If channel ROAS climbs and MER stays flat, someone may be collecting credit for sales that came anyway.
When to choose which
Use MER or blended ROAS when
You want to know whether total spend is pulling its weight. Use it month by month, next to promotions and seasonality, so you don't celebrate a sale period.
Use POAS or another profit measure when
You are deciding whether to scale a product line or a campaign. Revenue that leaves no margin is not worth buying more of.
Use none of them alone when
The decision is "where should the next euro go?". All three are averages. One measurement vendor argues that neither ROAS nor MER answers that, and points to marginal ROAS: the return on the next unit of spend2. That is a vendor view, but the question is fair.
Our take
Write down which number each decision uses before the meeting, not during it. Our default:
- ROAS steers the work inside a channel.
- MER, or blended ROAS if you want media only, shows the direction of the whole.
- Contribution margin decides whether growth is worth paying for.
- A test or an incremental ROAS estimate checks whether the attributed numbers hold up when a big budget decision is coming.
None of this is an industry standard. It is our convention, and we label it as ours.
For ZFD: A good ROAS in the panel does not close the question. We follow the money to what the business keeps. For the other half of the argument, read what is a good ROAS for ecommerce.
FAQ
What is the difference between MER and ROAS?
ROAS divides attributed revenue by ad spend for a channel or campaign. MER divides total revenue by total marketing spend, with no attribution. ROAS steers channels. MER shows overall direction.
Is MER the same as blended ROAS?
Some sources treat them as one metric; Shopify's blog says MER is also called blended ROAS4. In our convention, blended ROAS uses paid media spend as the denominator and MER uses all marketing spend. Always say which you mean.
Is POAS better than ROAS?
It is closer to the money, because it takes costs off. It is still based on attributed sales, so it shares attribution's blind spots.
Our agency already reports ROAS. Why do we need anything else?
ROAS shows what a platform credited itself with. It does not show what was left after costs or what would have sold anyway. Keep it for steering. Don't run the business on it.
What is a good MER?
It depends on your margin and on what you include in marketing spend. Compare your MER with your own history and your contribution margin, not with someone else's number.
“Their engagement is impressive.”
Sources
- ROAS Explained (Polar Analytics)Vendor
- MER vs ROAS: The Third Number Both Metrics Hide From You (BlueAlpha)Vendor
- What is MER? (Admetrics)Vendor
- Marketing efficiency ratio (Shopify blog AU)Vendor
- Why do platform ROAS and incremental ROAS tell such different stories? (Sellforte)Vendor
- Why can an ad with a ROAS of 3.8 still lose money? (Shopify Community)Forum
- PPCChat: MER vs ROAS (Karooya, 7 XII 2021)Forum
Next step
Put your own numbers through the Is your ROAS real? calculator and see how much of your ROAS survives your margin.