Google Ads says 4.2x. Meta says 3.6x. TikTok says 3.0x. Every channel cleared its target, and then month-end lands and the business made less money than it did in a quieter month on less spend.
Nobody is lying and nothing is broken. The dashboards and the bank account measure two different things. Once you see why, the argument changes from “whose number is wrong” to “which number answers which question.”
Three numbers settle it, not two. ROAS says whether a channel is pulling its weight. MER, the marketing efficiency ratio, meaning total revenue divided by total marketing spend, says whether the whole operation is. Contribution margin, meaning what is left from a sale after the costs of making and delivering it, says whether any of it made money.
What ROAS measures, in plain terms
Return on ad spend is one channel’s attributed revenue divided by what you spent on that channel. Our guide to what ROAS is covers the formula, the breakeven maths, and why 1.0x is not the same as breaking even.
The word doing the work is attributed. ROAS measures revenue that one system decided to credit to one set of ads, using its own rules about who gets credit and for how long. Change the rules and the number changes without a single extra order arriving. That makes it a useful steering signal at the level it was built for, and a poor description of your business.
Why three platforms can all be right and the total still be wrong
The mechanism has a name: the attribution window, meaning how long after someone sees or clicks an ad the platform will still take credit for a sale.
Google’s own advertising help documentation states the default plainly: unless you customize the click-through conversion window when you create a conversion, it is 30 days. Meta’s documented standard is tighter, seven days for a click and one day for a view, and since a change Meta made in January 2026 that one-day view is the longest view-through window Ads Manager offers. Google’s default reach back is roughly four times longer.
Neither window is adjustable jointly, and there is no shared ledger between the platforms. A customer clicks a search ad on the 1st, sees a Meta ad on the 12th, and buys on the 20th. That is one order in your system, and a full conversion in two platforms’ reports, each correctly applying its own rules.
Search Engine Land, reporting on how ad platforms count conversions differently, puts it in a line: there are only so many real sales in a period, and several platforms will each claim the same one. Our post on tracking ROAS by channel shows what that looks like once you add the dashboards up. That is not a bug. It is window maths, and a reason to stop summing ROAS.
What MER measures, and why no setting can inflate it
MER is total revenue divided by total marketing spend over the same period. HubSpot’s guidance gives it in exactly that form, and Shopify’s states it identically: revenue in, marketing money out, one ratio.
Where the numbers come from matters more than the formula. The numerator is your actual revenue, out of your books, the figure your accountant would recognize. There is exactly one of it, and it was never assembled out of platform claims. Lengthen every window on every platform tomorrow and your channel ROAS figures climb while MER does not move by a cent. It is immune to double counting by construction, not by cleverness. HubSpot draws the same line: ROAS focuses on the return of specific campaigns, MER gives a blended view of marketing as a whole.
Its one soft spot is the denominator. Blended ROAS usually means total revenue over paid media alone, while MER more often means total revenue over everything you spend to market, agency retainers and creator fees and production and software included. Those are not the same number. Pick a definition, write it down, and hold it constant.
A worked example: three healthy channels, one unprofitable month
One month, one hypothetical direct-to-consumer brand, in US dollars.
| Channel | Spend | Revenue the platform claims | Reported ROAS |
|---|---|---|---|
| Google Ads | $40,000 | $168,000 | 4.2x |
| Meta | $30,000 | $108,000 | 3.6x |
| TikTok | $10,000 | $30,000 | 3.0x |
| Total | $80,000 | $306,000 | 3.83x |
Every channel cleared its target. Nothing there would trigger a review.
Actual revenue in the order system that month, across every source including organic search, direct traffic, and email, was $232,000. That is $74,000 less than the platforms claim between them, and the real gap is wider still, because some of that $232,000 is revenue no platform claimed credit for at all.
Marketing costs beyond media: agency retainer $9,000, creative production $7,000, creator fees $6,000, software $4,000. Total marketing spend, $106,000.
| Ratio | Calculation | Result |
|---|---|---|
| Platform-reported ROAS, summed | $306,000 ÷ $80,000 | 3.83x |
| Blended ROAS, media spend only | $232,000 ÷ $80,000 | 2.90x |
| MER, all marketing spend | $232,000 ÷ $106,000 | 2.19x |
Same month, same business, three answers. Only the last one reconciles to the bank.
The third number: what you actually keep
All three are still ratios of revenue, and revenue is not money you keep. Contribution margin is what survives a sale once you subtract every cost that rises with it.
On that $232,000: cost of goods $92,800, fulfillment and inbound freight $23,200, payment processing $6,960, and an allowance for returns and discounts of $11,600. Those variable costs total $134,560, leaving $97,440 before any marketing cost, or 42 percent of revenue. Subtract the $106,000 of marketing spend and the month finishes at minus $8,560, before rent, salaries, or anything else fixed.
Contribution margin also hands you a target. Breakeven ROAS is one divided by your gross margin, which the ROAS guide works through, and the same arithmetic extends to the blended number: breakeven MER is one divided by your contribution margin percentage. At 42 percent, that is 2.38x.
MER came in at 2.19x, under the bar, which is why the month lost money. Blended ROAS on media alone was 2.90x, over it, which is why a team scoping its denominator to media only would report a profitable month in good faith. Every channel cleared 2.38x with room to spare. Every reading but one is above breakeven, and the business still lost money.
When the channel number should drive the decision
ROAS is the right instrument when the question is narrow and the comparison is like for like.
- Creative and audience tests inside one platform. Both variants sit under identical rules, so the comparison is fair even when the absolute number is inflated.
- Bid and budget moves within a channel. Relative performance between two campaigns beats either one’s absolute figure.
- Diagnosing a bad month. When the blended number drops, channel ROAS is where you find out which part moved.
Use ROAS to compare things measured the same way. Never to add up things measured differently.
When the blended number should drive the decision
MER is the right instrument when the question is about the whole system and the money is real.
- Monthly and quarterly spend envelopes. What this business can afford is a question about total spend against total revenue.
- Board, lender, and investor reporting. It reconciles to the accounts, so nobody has to take a platform’s word for anything.
- New channel bets. A new channel looks superb in its own dashboard for two months because it is harvesting demand other channels created. MER catches whether total revenue actually moved.
There is no universal MER target, and be wary of anyone who hands you one. The reason is mechanical: a business on 60 percent margins and one on 25 percent need different ratios simply to break even, and the breakeven formula above gives each of them their own number.
The trap: cutting your way to a better blended number
Because MER divides revenue by spend, the fastest way to improve it is to spend less. Switch off the channels that do not visibly close sales and the ratio jumps immediately: the denominator drops this week, while the revenue those channels were seeding does not disappear until next quarter. It reports like a win, and it is a delayed cut to the top of the funnel.
The exposed channels are the ones that start journeys rather than finish them: prospecting, video, upper-funnel search, creators. They rarely look impressive on last-click reporting, and their absence surfaces months later as thinner branded search and fewer returning customers.
A single blended ratio has no channel dimension by design, so MER cannot tell you which channels are the exposed ones. The way past it is to test instead of assume: hold a channel out, or run matched markets, and watch whether total revenue moves. Our guide to reported versus incremental ROAS explains what that distinction is worth, and the incrementality testing guide covers how the tests are run.
Reading all three in the same room
- MER first. It ties to the books, so it sets the tone.
- Breakeven MER next to it. One divided by your contribution margin percentage. Without it, MER is a number with no verdict attached.
- Channel ROAS underneath, as diagnosis, never as a total. Show trend and relative movement, and say out loud that the platform figures overlap.
- Contribution dollars last. Revenue ratios describe efficiency; this one says whether the month made money.
The alternative is a deck of healthy channel numbers and a finance lead who has already read the P&L and stopped believing it by page two. Leading with the number that reconciles buys you the credibility to explain the ones that do not.
None of this needs software to understand. It needs software to maintain: the month-end version is somebody reconciling three platform exports, an order export, and a spend sheet by hand. Attriqs exists to make that view standing rather than manual, one independent record of where revenue came from beside what you spent to get it.
Frequently asked questions
Is MER the same thing as blended ROAS? Close, and the terms get swapped in casual use. Both divide total revenue by spend, but blended ROAS usually counts paid media only while MER counts every marketing cost. That makes MER the stricter of the two. Pick whichever suits your business, write the definition down, and do not change it mid-year.
Which one should I report to my board? MER, with your breakeven MER beside it and contribution dollars underneath. MER reconciles to your accounts, which channel ROAS does not, and no settings change inside an ad platform can move it. Keep channel ROAS in the appendix, as diagnosis.
What is a good MER? There is no universal answer, and any number offered without reference to your own margins is closer to marketing than to measurement. HubSpot’s guidance on MER declines to publish a target and points businesses at their own historical performance instead. Compute your own: one divided by your contribution margin percentage is the ratio you have to beat. Independent benchmarks are thinner than they look. The most methodologically transparent figure nearby is Duke University’s CMO Survey: 308 US marketing leaders, fielded January 2026, putting marketing budgets at 9.0 percent of company revenues. That is budget share on a largely business-to-business panel, not an efficiency ratio, so do not invert it into an implied MER.
Can I fix my attribution and use ROAS on its own? Better attribution narrows the gap, but it cannot close it. Any model still divides credit for sales that already happened, and it cannot tell you what would have happened had you run no ads at all. That is what incrementality testing is for, meaning does revenue actually change when the ad stops running. It is also why MER stays on the page: it is the one ratio nobody can adjust their way into.
Ready to see one picture instead of three versions of it?
Attriqs gives you an independent, channel-level view of revenue that reconciles to your books, so ROAS, MER, and margin can be read on one page rather than assembled by hand every month end. Get in touch and we will set it up around the way you already report.