ROAS vs MER vs blended CAC: which number should actually drive your ad decisions

Use platform ROAS to diagnose channels, not to judge the business. Use MER to know whether you are actually making money, because it runs on real store revenue with no attribution games. And never treat one platform's ROAS as truth, because every platform claims credit for the same sale, so the numbers add up to more revenue than your bank account ever saw.

That is the whole answer. The rest of this page explains why, and gives you a rule for which number to reach for depending on the question in front of you.

The three numbers, plainly

Platform ROAS is revenue attributed by a platform, divided by the spend on that platform. Meta reports its ROAS, Google reports its ROAS. Each one is measuring its own slice, using its own attribution window and its own view of who caused the sale.

Blended ROAS is total revenue over total paid spend across every channel. No attribution. You take the money that came in and divide it by everything you spent on ads. This is your daily dashboard number, the fast read on whether today was healthy.

MER, marketing efficiency ratio, is total revenue divided by total marketing spend, again using real store revenue and no platform attribution. This is your weekly P and L view. It is slower and blunter than blended ROAS, but it is the number your accountant would recognise. A commonly cited healthy DTC range for MER is roughly 3.0 to 5.0, depending on your margins.

Blended CAC is the all-in cost to acquire a customer: total spend divided by customers acquired. It answers a different question again, which is what you are paying to bring someone in.

The inflation trap

Here is the thing that breaks most reporting. If you add up the revenue Meta claims and the revenue Google claims and the revenue your email tool claims, you get a number far bigger than your actual sales. Every platform counts the same purchase. A customer sees a Meta ad, clicks a Google search ad, opens an email, then buys. All three claim the sale.

This is not a bug you can fix by picking a better attribution setting. It is structural. Platform ROAS is systematically inflated because each platform is paid to prove its own worth. A commonly cited figure is Meta overstating ROAS by roughly a quarter, and platform ROAS routinely reads far above the blended ROAS you can verify in Shopify.

So platform ROAS is not a lie, but it is not truth either. It is a channel's self-report. Useful for comparing ad sets inside that channel, dangerous the moment you use it to decide how the whole business is doing.

Which number answers which question

Think of it as three different questions, not three competing answers.

Question one: is this channel or ad working relative to my other channels and ads? Use platform ROAS, but only to compare like with like inside the same account. Ad set A at 3.1 versus ad set B at 1.4 is a real signal about which creative and audience is pulling, even if both numbers are inflated by the same attribution bias. You are comparing two inflated numbers on the same scale, so the comparison holds.

Question two: is the business actually making money? Use MER, and cross-check with blended ROAS day to day. These use real revenue. If MER is sliding while every platform still reports a glorious ROAS, believe MER. The platforms are claiming credit for sales that were going to happen anyway.

Question three: what does it cost to get a customer, and can I afford it? Use blended CAC against your margin and your customer lifetime value. This is the acquisition question, and it is where scaling decisions live. One caution: blended CAC mixes new and returning buyers, so it can look great simply because loyal customers keep repurchasing. For growth you want to isolate the cost of genuinely new customers. We cover that in why nCAC beats blended CAC for growth.

| Metric | What it measures | What it is good for | Its blind spot |

| --- | --- | --- | --- |

| Platform ROAS | Revenue a platform attributes to itself, over spend on that platform | Comparing ad sets and creatives inside one channel | Inflated by double-counting, unsafe as business truth |

| Blended ROAS | Total revenue over total paid spend, no attribution | Daily read on overall efficiency | Does not separate new from returning, or channel from channel |

| MER | Total revenue over total marketing spend, real store revenue | Weekly P and L view of whether you are profitable | Too blunt to tell you which ad or channel to fix |

| Blended CAC | All-in cost to acquire a customer | Judging affordability against margin and LTV | Hides new-customer cost behind repeat buyers |

How not to steer the business off an inflated number

The common failure looks like this. A media buyer sees Meta reporting a 4.0 ROAS, pours budget in, and watches MER quietly fall. Meta kept claiming sales that email and organic were already going to close. Spend went up, real profit went down, and the dashboard looked green the whole time.

The fix is a hierarchy. Let MER and blended ROAS set the ceiling on how much you can spend and still make money. Let platform ROAS guide allocation inside that ceiling, choosing which ad sets get the budget you have already decided is safe. Never let platform ROAS set the ceiling itself. It is a steering wheel, not a fuel gauge.

When a channel's reported ROAS and your blended numbers disagree, that gap is the signal, not the noise. It usually means the channel is harvesting demand that already existed rather than creating new demand. That is also often where a funnel is leaking, and you can find the funnel step that is leaking rather than guessing.

Reconciling all of this by hand, across Meta, Google, and your store, every day, is where most teams give up and just trust the platform number. This is the gap AdBrain is built for: it reconciles platform-reported figures against what the account actually did, so scale, refresh, and kill decisions run on reconciled numbers instead of inflated ROAS.

Pick the metric that matches the question. Diagnose channels with platform ROAS. Judge the business with MER. Price acquisition with blended CAC. And keep one rule above all of them: no single platform's ROAS is the truth, so never let it drive the whole ship.