ROAS vs MER: why platform ROAS is lying to you

Most ad accounts are managed toward a single number: return on ad spend as reported by the platform. It's convenient, it's in every dashboard, and it's quietly steering budgets in the wrong direction. The ROAS vs MER debate is really a question about which number you trust when deciding whether to spend more. Platform ROAS tells you what an ad platform claims it drove. MER tells you what your whole marketing budget actually produced.

Both have a place. This guide explains where each one belongs, why platform ROAS misleads when it's used for the wrong decisions, and the scorecard we use instead.

ROAS vs MER: the definitions

• ROAS (return on ad spend) = revenue attributed to ads ÷ ad spend, usually per campaign or channel, as reported by each ad platform.
• MER (marketing efficiency ratio) = total revenue ÷ total marketing spend, across every channel, from your store or finance data.
• New-customer MER (sometimes called aMER) = revenue from first-time customers ÷ total marketing spend. It strips out repeat buyers who may have purchased anyway.

ROAS is a channel-level optimization metric. MER is a business-level health metric. Problems start when the first is used to make decisions that belong to the second, like setting total budget or deciding whether marketing is working at all.

A hypothetical week shows how far apart they can drift. A store spends $10,000 across Meta and Google. Meta reports $30,000 in purchase value and Google reports $25,000, so each platform shows a healthy ROAS. But the store's own records show $40,000 in total revenue, and $12,000 of that came from first-time customers. MER is 4.0, new-customer MER is 1.2, and the combined platform claims add up to more revenue than the business actually took in. Which number should decide next month's budget? The illustrative numbers make the answer obvious: the one that matches the bank account.

Four reasons platform ROAS misleads

• Attribution overlap: Meta, Google and TikTok each claim credit for the same purchase. Add their reported revenue together and it can easily exceed what actually landed in your bank account.
• It ignores margin: a 4x ROAS on a product with a 25% gross margin loses money once fulfillment, fees and returns are counted.
• It rewards retargeting: campaigns that show ads to people who were already about to buy look brilliant on ROAS while creating little new demand.
• It mixes observed and modeled conversions: platforms estimate conversions they can't observe directly and count view-through conversions inside their default windows. Those are useful signals, not an audit.

None of this means the platforms are being dishonest. Each one reports its own view of the world using its own rules. The problem is treating that view as the ledger.

Where MER falls short too

MER is much harder to game, but it's a blunt instrument. It can't tell you which campaign to cut, it reacts slowly when you test a new channel, and it can be flattered by seasonality, price increases or a strong email calendar. A rising MER while new-customer counts fall is a warning sign, not a win.

• Use MER and new-customer MER to set and defend the overall budget.
• Use channel metrics, including platform ROAS, to optimize within each channel.
• Use incrementality tests to settle arguments between the two.

Manage to contribution margin instead

Start with what a customer is worth to the business after cost of goods, shipping, payment fees and returns. That's the ceiling on what you can pay to acquire one. Then work backward into a target cost per new customer for each channel.

A hypothetical example: say a new customer generates $50 of contribution margin on their first order and the business wants first orders to break even. Your ceiling for blended new-customer CAC is $50. If finance can fund a 90-day payback and your own cohort data shows another $20 of margin in that window, the ceiling rises to $70. The numbers are made up; the method isn't.

If you can't say what a new customer is worth in dollars, you can't say whether your ads are working.

From the blended ceiling, set channel targets. Each platform's CPA or ROAS target should be calibrated against what the business actually sees: if Meta consistently claims more new customers than show up in your store data, its in-platform target needs to be stricter than the blended ceiling to compensate. Recheck that calibration whenever tracking, attribution settings or the channel mix change.

A simple weekly scorecard

• Total marketing spend (all channels, including agency and creative costs if you want the full picture)
• New customers acquired (from your store or CRM, not ad platforms)
• Blended new-customer CAC = spend ÷ new customers
• MER and new-customer MER
• First-order contribution margin
• 90-day repeat revenue per cohort

Read the scorecard as trends, not single weeks. Rising spend with flat new-customer counts means marginal CAC is climbing: the extra budget is buying less. Rising MER with falling new customers usually means you're harvesting existing demand rather than creating it. And a strong week followed by a weak one is often just timing: promotions, paydays and email sends all move weekly numbers.

Pull these numbers from your store, CRM and finance tools rather than ad platforms. If those sources don't agree with each other, fix that first. Our guide to conversion tracking setup walks through how. Platform metrics still matter for in-channel optimization. They just shouldn't be the number leadership uses to decide whether to spend more.

How to test what's actually incremental

Neither ROAS nor MER proves cause and effect. To know whether a channel creates sales that wouldn't have happened otherwise, you have to test it:

• Geo holdout tests: pause or reduce spend in a set of comparable regions and compare sales against regions where spend continued.
• Platform lift studies: Meta and Google both offer conversion lift tests that compare exposed and held-out groups, for accounts that meet their requirements.
• Planned spend pauses: for smaller channels, a deliberate pause with a clear before-and-after read can be enough to make a decision.
• Marketing mix modeling: open-source tools such as Google's Meridian and Meta's Robyn estimate each channel's contribution from historical spend and sales, once you have enough data history.

Start with one test on your largest or most-questioned channel. The result usually changes the budget conversation more than months of dashboard debate.

Switching from ROAS to MER: a step-by-step plan

• 1. Calculate first-order contribution margin per customer from store data.
• 2. Agree a payback window with finance and derive a target blended new-customer CAC.
• 3. Build the weekly scorecard from first-party sources.
• 4. Set channel-level ROAS or CPA targets that are consistent with the blended target, not independent of it.
• 5. Run one incrementality test per quarter on the channel you're least sure about.
• 6. Review monthly and move budget based on the scorecard, using platform metrics to optimize inside each channel.

This is how we run PPC management, with marketing analytics and attribution providing the source of truth. When it's time to set the annual number, the same inputs feed a marketing budget you can defend to your CFO. If you want to know what that would cost for your account, see how our pricing works.

Frequently asked questions

Is a higher ROAS always better?

No. A very high ROAS often means spend is concentrated on retargeting and branded search (people who were likely to buy anyway), and that you may be under-investing in finding new customers. The right ROAS is the one that delivers your target new-customer CAC at the scale you need.

What is a good MER?

It depends on your gross margin, repeat purchase rate and growth goals, so there's no universal target. Work out the MER at which marketing spend breaks even on contribution margin, then decide how far above that line you need to operate.

Should I stop looking at ROAS completely?

No. Platform ROAS is still useful for comparing campaigns, ad sets and creatives within a single platform. Just don't add ROAS figures across platforms, and don't use ROAS alone to decide your total budget.