A 4.0x ROAS proposal can pass the agency’s test and fail your finance lead’s. One counts revenue Meta credited to ads; the other wants to know what survives refunds, costs and the agency invoice.

The gap starts in the definitions, not the dashboard. A ROAS target means nothing commercially until both sides agree what counts as revenue, which costs are excluded and who gets credit for the sale. A precise-looking number does not rescue vague accounting.

Free tool: the ROAS Calculator works out your return on ad spend, your break-even ROAS and the budget a revenue goal takes.

I treat these 12 terms as a translation sheet, not a vocabulary test. Each one settles a buying decision you should make before granting account access. Get the definitions into the proposal now, while everyone is still enthusiastic about answering questions.

Platform-attributed ROAS

Platform-attributed ROAS is the revenue credited to ads divided by media spend.

Meta calculates it as purchase conversion value divided by amount spent, using the purchase data sent through your connected tools. That makes the revenue feed part of the definition, not a technical footnote someone can sort out after onboarding.

The misuse starts when an agency presents that credit as proof the ads caused the revenue. Attributed revenue is not the same as revenue caused by advertising. It is revenue assigned to ads under the platform’s rules.

When someone says, “We average 4.5x,” I want three answers: over what spend, using what revenue feed, and under which attribution settings? Without those, 4.5x is decoration.

Ask before signing: “What exactly goes into the numerator and denominator of your promised ROAS?”

Attribution window

The attribution window is the time boundary that decides which purchases can receive credit for an ad interaction.

Meta’s 7-day click and 1-day view settings allow different routes into the reported total. A purchase following a click can qualify over a longer period than one credited after a view.

A wider window can credit more purchases to the same spend without the ads performing any better. Same store, same ads, larger net. That is why two agencies can show different ROAS figures without either having changed a campaign.

Fix the attribution window before comparing proposals. Otherwise, you are comparing different-sized nets and calling the difference skill.

I would put the setting beside the target in the agreement, not bury it in reporting notes. If the setting changes later, the comparison needs to say so.

Ask before signing: “Which window does this target use, and will it stay fixed?”

Click-through conversion

A click-through conversion is a credited purchase that follows a click on your ad within the attribution window.

The important distinction is between a click that qualifies for this bucket and other engagement. Meta’s reporting separates link-click attribution from other clicks and qualifying video engagement. “They engaged with the ad” is not a precise enough explanation of what the report counts.

The more expensive confusion is treating click-through as another word for incremental. A click establishes an interaction, not proof that the sale would otherwise disappear. The buyer touched an ad and purchased inside the window. That is useful information. It is not the whole commercial argument.

Ask before signing: “Do you report click-through and engage-through separately, and which does your ROAS target include?”

View-through conversion

A view-through conversion is a credited purchase after someone sees an ad without clicking it.

A 1-day view setting allows a purchase to receive credit after an impression alone. Dismissing every such purchase as fake revenue is too easy. Seeing an ad and buying later are not mutually exclusive.

But the reverse shortcut is just as weak: include view-through credit, show a healthy total, then claim that number proves cold audiences convert because of the ads. The report has not established that.

Keep view-through visible rather than hiding it inside the headline. I want to know how much of the target depends on it, especially when the proposal makes strong claims about prospecting.

Ask before signing: “What share of reported ROAS is view-through, and would you still hit the target on click-only reporting?”

Gross revenue versus net revenue

Gross revenue is the sales figure before the agreed deductions; net revenue is what remains after deductions such as discounts, cancellations and refunds.

Those deductions need naming. “Revenue” on its own leaves too much room for the agency and finance to use different numbers while believing they agree.

Meta does not automatically subtract refunds from the original reported purchase value. That gap matters after the reporting screenshot has already done its job in the review meeting.

Take $100k in credited revenue on $25k in spend: reported ROAS is 4.0x. If $15k is subsequently refunded, revenue after those refunds is $85k and the corresponding ratio is 3.4x. The original platform figure can still show 4.0x.

Do not base the agency’s reward on gross revenue while finance plans around net revenue without explicitly accepting that difference.

Ask before signing: “Do you report Shopify net sales or Meta purchase value, and how do refunds change the result?”

Contribution margin

Contribution margin is the share of revenue left after the variable costs you explicitly include.

Product cost, fulfillment, payment fees, platform fees and a refund reserve all belong in the conversation. Name what is included and what is kept separate. Otherwise, margin becomes another reassuring word with no agreed calculation underneath it.

An agency that quotes a ROAS target without asking about margin has skipped the mechanism that makes the target useful. A 3.0x can look attractive while failing to cover the costs of the product being sold.

Margin turns a reporting target into a commercial threshold. I want that threshold built from your costs, not borrowed from a slide about another store.

Ask before signing: “Which costs sit inside this margin, and which have you kept outside it?”

Break-even ROAS

Break-even ROAS is one divided by your contribution margin before advertising, expressed as a decimal.

Using that method, a 60% margin needs 1.67x, a 40% margin needs 2.50x, and a 25% margin needs 4.0x. Fixed costs remain separate from that calculation. Break-even is not permission to forget them.

Say you spend $20k a month and have a 35% contribution margin before ads. Your break-even ROAS is about 2.86x. A proposal presenting 2.5x as a win is promising a result below that threshold.

I cover the full method for calculating break-even ROAS from contribution margin separately. The procurement test is simpler: make the agency show your calculation before showing its target. An industry benchmark cannot do that work.

Ask before signing: “What is our break-even ROAS under these cost assumptions?”

New-customer acquisition cost

New-customer acquisition cost is acquisition spend divided by genuinely new customers, identified from store order history.

The phrase new customer needs its own agreement. Meta’s new-versus-existing split depends on the audience definitions supplied; it is not automatically the same as the store’s first-time-buyer record. Stale exclusion lists can misclassify customers and flatter acquisition reporting.

A 4.0x blended ROAS also does not tell you whether the agency is acquiring customers efficiently or collecting orders from returning buyers. If 70% of orders come from returning customers, I want that visible before anyone presents the total as an acquisition success.

Define “new” using the customer record you agree to trust. Then keep the acquisition-cost calculation tied to it.

Ask before signing: “Will you report nCAC against Shopify first-time buyers rather than relying only on a Meta audience segment?”

Blended marketing efficiency ratio

Blended marketing efficiency ratio, or MER, is total revenue divided by total ad spend across all channels.

MER is attribution-agnostic by design. It does not ask each platform to claim its share of a purchase. That matters because channel reports can claim credit for the same sale.

I use platform ROAS to help steer decisions inside Meta and MER to assess the broader commercial picture. Neither number needs to impersonate the other.

The reporting trick to watch for is a mid-meeting substitution: platform ROAS leads when it looks strong; blended MER takes over when platform ROAS dips. Suddenly the goalposts have wheels.

Name both metrics and keep their sources consistent. MER still needs an agreed revenue figure and spend definition; being attribution-agnostic does not make it definition-free.

Ask before signing: “Will you show Meta ROAS and blended MER separately, using the same sources every month?”

Incrementality

Incrementality is the result advertising causes beyond what would have happened without it.

Attribution asks which ads receive credit. Incrementality asks whether the advertising changed the outcome. Conversion Lift uses test and holdout groups to investigate that second question.

The distinction matters most when the people receiving ads already have a strong reason to buy. A retargeting report can look excellent while leaving the harder question unanswered: how many of those purchases needed the ad?

A high attributed ROAS does not settle the incrementality question. I would rather hear an agency explain that limitation plainly than stretch a platform report into evidence it cannot provide.

Ask before signing: “Which of your claimed wins have you tested with a holdout, and how do you decide when to run one?”

Conversion lag

Conversion lag is the delay between an ad interaction and the purchase being credited, which leaves recent reporting incomplete.

Purchases happen later, and event reporting can arrive late too. Reporting delays make a fresh result a moving number rather than a finished verdict.

My practical rule is to treat the last 48 hours as provisional. That does not mean every number becomes final at hour 49. It means I do not want a Monday review turning incomplete week-to-date ROAS into a confident diagnosis.

The same problem appears in Google Ads reporting affected by conversion lag: a short window can punish whatever spent most recently.

Agree when you will judge a period, then revisit it consistently. Otherwise, timing becomes another variable disguised as performance.

Ask before signing: “What reporting cutoff do you use, and will you show the same week again after seven days?”

Fully loaded acquisition cost

Fully loaded acquisition cost is media spend plus management and creative costs, divided by the number of customers acquired.

Leave out fees and you are comparing only part of the bill. Agency pricing can include percentage-of-spend charges, flat retainers and separate creative costs. The proposal needs to show which apply, not leave you to discover the denominator invoice by invoice.

On $20k in media spend, a $4,000 retainer adds 20% to the cost before creative. That is a 20% increase in cost, not a 20% reduction in ROAS. The distinction matters when someone starts translating fees into “true efficiency.”

Keep acquisition cost and revenue-to-cost ratios clearly labelled. One divides costs by customers; the other divides revenue by costs. Both can expose the fee burden, but they are not interchangeable.

Ask before signing: “What is the acquisition cost including your fee and expected creative bill, and what happens to the revenue-to-cost target when those costs are included?”

Put the definitions into a one-page measurement agreement

I would make every finalist accept the same math before touching spend. The agreement should make different proposals comparable, not create another reporting document nobody reads.

One page needs these decisions:

  • Revenue source: Shopify net sales or Meta purchase value, with explicit treatment of discounts, refunds and cancellations.
  • Attribution settings: The agreed click and view windows, with click-through, engage-through and view-through credit shown separately where applicable.
  • Included costs: Media, management and expected creative, with exclusions named rather than implied.
  • Customer definition: The record used to identify first-time buyers and calculate new-customer acquisition cost.
  • Reporting cutoff: The provisional period, the weekly review point and when the same period gets revisited.
  • Metric labels: Platform ROAS, blended MER and fully loaded acquisition cost, each with its own agreed calculation.

Then ask every shortlisted agency to restate its target against that page. I would apply the same standard to groas: no exemption for the provider you already like. Finance wants to know what counts, what is excluded and when a number is ready to judge.

Good ROAS

Good ROAS is a judgment that only means something after revenue, costs and attribution rules are agreed.

This is the term I would retire from proposal decks. “Good” sounds like a benchmark and behaves like fog.

Good on gross or net? Click-only or with view-through folded in? Before fees or after them? From new customers or returning buyers who might have purchased anyway?

Until those choices are written down, good means whatever the seller needs it to mean that week.

Send the one-page agreement before granting ad account access. The agency that restates its target against your revenue source, your window and your fully loaded cost is showing you how it will report when results get uncomfortable. Hire that one.