How do agencies artificially raise reported ROAS while growth is actually slowing down?

They can raise the reported ratio by cutting prospecting efforts and focusing only on retargeting past buyers. This strategy harvests existing demand instead of creating new sales opportunities. This method often leads to total sales sliding even when the ratio looks good in reports.

How does an agency's fee structure potentially encourage poor performance decisions?

Agencies using a percentage-of-spend model are financially incentivized to keep ad spend stable on high-performing audiences. This model discourages managers from taking necessary risks, such as spending more to reach strangers, which is often required for business growth.

What calculation should be used to determine if an ad campaign is actually profitable?

The break-even ROAS should be calculated as one divided by the gross profit margin. Without this number, a ROAS target is just theoretical theater. This calculation ensures the business is profitable on each individual sale regardless of the ad performance ratio.

Why should a business look at blended ROAS instead of only the Meta platform number?

Blended ROAS is total revenue net of returns divided by total ad spend across all platforms, providing an honest picture. This metric does not claim which ad caused a sale, but it accurately reflects what the business actually took in against what it spent.

Beyond overall ROAS, what metric should agencies prioritize to ensure business growth?

Managers should focus on new-customer ROAS and new-customer CPA measured against lifetime value. This ensures the agency is successfully acquiring genuinely new customers, which is the primary indicator of sustainable business growth.

Don’t Hire the Meta Agency That Promises Higher ROAS

Q: How accurate is Meta's reported ROAS given the platform's attribution windows?

Meta's reported ROAS can significantly overestimate true returns due to its specific measurement rules. The platform uses a seven-day click plus one-day view window and includes modeled conversions for untracked traffic. This creates a gap between the reported ROAS and the actual incremental ROAS.

If a Meta agency promises higher ROAS on the sales call, do not hire them. Hire the one that asks permission to lower it for a while.

I spent years running accounts where ROAS was the number the client wanted to see go up. I know how fast I could make it happen: cut prospecting, push budget into retargeting and past buyers, then report through the window that flatters the result. Reported ROAS climbs. The business may not grow by a dollar. Any manager who leads with “we improve ROAS” is either selling you that trick or does not know it is one.

Here is the mechanism. Meta will happily take credit for people who were already going to buy. Show an ad to someone who abandoned a cart yesterday. They click and check out; Meta reports a 6x return on that ad set. In documented lift tests, roughly 60% of those retargeting conversions would have happened anyway, and true incremental ROAS often runs 40 to 70% below what Ads Manager shows. That gap is where ROAS-polishing lives.

I used to run a version of this play without meaning to be dishonest. Say you inherit an account spending $20k a month, with tired prospecting at 1.6x. Shift 70% of the budget to warm audiences, kill the broad ad sets that look inefficient on a seven-day window, and the account suddenly reports 3.4x. The client is thrilled. But a better test asks what would have happened without the ads. One brand reporting 4.1 ROAS inside Meta ran a three-week geo holdout; revenue in the paused markets fell by only a third of what Meta claimed it drove. Two-thirds of those sales were going to happen with the ads off.

The fastest way to raise reported ROAS is to shrink the business that produces it. Cut prospecting and you cut the expensive work that creates future retargeting pools. Lean on existing customers and you harvest demand instead of making it. Let attribution claim sales driven by email, search or habit. The ratio rises while revenue stalls or falls.

Three ways to make ROAS rise while growth stalls

First, cut prospecting. It looks terrible on a weekly report. You pay to reach people who have never heard of you. Most do not click; most who click do not buy. That ad set prints at 1.2x while retargeting prints at 5x. So the ROAS-first manager pauses prospecting to “reallocate to what works.” For two weeks, the blended number in Ads Manager looks excellent. Then the retargeting pool starts to starve because fewer new people enter it. Total sales slide even as the ratio stays pretty. You bought a better-looking fraction by shrinking the business underneath it.

Second, live in the cart. Retargeting past site visitors, video viewers, email openers and past buyers beats cold traffic on reported return because much of the intent was already there. Push 80% of spend toward them and you can get a beautiful 4x that your accountant cannot find. That is not media buying. That is standing by the checkout register and asking for commission.

I used to tell clients this was “efficiency.” I was wrong. It was harvesting. If your manager cannot tell you what share of Meta revenue came from new customers, ask why before you celebrate the ROAS.

Third, let attribution take the credit. Meta reports on a seven-day click plus one-day view window and includes modeled conversions for traffic it cannot directly observe. Someone sees your video on Monday, Googles you on Thursday and buys; Meta may claim the sale. An email flow brings someone back; Meta may claim a share. Refunds happen, but the platform number does not subtract them. That is how you get the split where Ads Manager shows 4.2x while Shopify math, net of returns and across all spend, shows about 1.8x. Same sales. Different counting rules, and only one set pays your suppliers.

My rule after getting burned on this: reported ROAS tells you what the platform credits itself for. Growth tells you how many extra dollars showed up. If your manager reports the first number every week and never checks it against the second, they are grade-grubbing on a test they wrote themselves.

The fee model can reward the same bad decision

A percentage-of-spend agency faces a separate incentive problem. If it finds $2,000 a month in wasted spend and switches it off, a typical 15% deal gives the agency a $300 monthly pay cut for doing the job well. Agencies often charge 10 to 20% with a minimum. Their fee climbs with spend even when the return on the next dollar falls.

Pair that fee with a ROAS-first promise and the easy path is clear: keep spend on the audiences that produce the safest reported ratio. Avoid the awkward conversation about spending more to reach strangers, or spending less because the account is saturated. A polished platform number makes either conversation easier to postpone.

This is what groas gets right structurally. groas runs Meta Ads for a flat $1,499 a month, with no cut of your ad spend. A flat-fee manager can recommend a smaller budget next month without taking a smaller fee. I am not saying every percentage agency cheats. I am saying its fee model works against one of the decisions you are hiring a manager to make. Pay for the judgment, not for keeping spend on the books.

Set break-even before you set a ROAS target

Start with break-even ROAS, written down before anyone touches the account. The basic math is one divided by gross profit margin. Shopify’s example is clean: $100 of revenue minus $60 in product, shipping and fees leaves a 40% margin, so break-even ROAS is 2.5x. On truly incremental first-order sales, below 2.5x loses money; above it makes money. Without that number, a ROAS target is theater.

Say you sell skincare at 70% margin after product, shipping and transaction fees. Break-even is 1.43x. A 2.0x ROAS that looks weak in an agency deck can be profitable growth. Sell furniture at 25% margin after freight and returns, and break-even is 4.0x. A 3.5x ROAS that looks handsome in a case study loses money on each order under those assumptions. If a manager does not ask for your margin and write down break-even on the first call, they are not managing profit. They are managing a ratio in Ads Manager.

Then demand a new-customer guardrail. Blended ROAS rewards selling to people who already bought. Manage new-customer ROAS and new-customer CPA against lifetime value, and use total revenue as a check on the program. One agency setup I respect uses explicit triggers: pull prospecting back around 1.7 new-customer ROAS, push spend around 2.1, and judge the program on merchant-reported revenue rather than Meta’s claim. Evestar describes that approach. The point is not to copy its thresholds into your account. It is to make the manager say when buying strangers is worth a lower reported return.

Check the plumbing before trusting that new-customer number. If purchase events are missing from Pixel or CAPI, past buyers can appear to Meta as new customers. An agency that promises to grow new-customer share but never asks about event coverage may be reading a broken gauge and billing you for it. Ask for the check first.

Finally, put blended ROAS beside the platform number: total revenue net of returns from Shopify or your order system, divided by total ad spend across platforms. It does not decide which ad caused a sale. That is its value here. It shows what the business took in against what it spent without accepting Meta’s claim to every order. If Meta reports 4x and blended sits at 1.9x, you did not get 4x across the business. You got a 4x platform report and a 1.9x blended result.

I check this monthly rather than weekly because weekly numbers jump with paydays, email drops and weekends. Pull net revenue, pull all ad spend, divide, and put the result next to Meta’s report. If revenue climbs while reported Meta ROAS dips, do not punish the manager for the uglier ratio before you understand what the extra spend bought. The manager worth paying points to the business line first and lets the platform line sit underneath it.

Five sales-call questions that expose a ROAS polisher

Bring these to the call. You do not need to know how to build an ad set to hear the difference. A polisher defends the ratio. A grower explains what the ratio costs.

What is my break-even ROAS, and will you put it in the proposal? A useful answer starts with margin, shipping, fees and returns, then writes down a number such as 2.5x. A bad answer promises 4x without asking what you keep from a sale.
What share of Meta revenue will come from new customers, and how will you check it? Listen for a target and a CAPI coverage check. “We’ll report blended ROAS” does not answer the question.
If we raise prospecting spend 30% next month, what happens to reported ROAS? A manager willing to buy growth can tell you it may drop for two to three weeks while the retargeting pool refills. A promise that it will rise on schedule is a sales answer, not a plan.
How will you reconcile Meta with Shopify and total spend each month? Look for net revenue divided by all-platform spend, shown alongside platform numbers. An Ads Manager screenshot is not a reconciliation.
What would vanish if we turned Meta off for three weeks? Ask for a holdout, geo split or pause-test plan and a prediction the manager is willing to get wrong. “The pixel proves it” is not a test.

The one reason to hold ROAS high

There is one owner who should hold the line on ROAS and mean it: the one who cannot afford to buy growth yet. If cash is tight, margin is thin after returns, and a bad month of prospecting means missing payroll, keep to the smaller, profitable base. A brand with 25% margins and no repeat purchase has no business paying 1.8x to acquire a stranger. That is not cowardice. That is arithmetic. The honest manager says it plainly: we need cash now, so we will keep ROAS high and accept that growth waits.

Everyone else pays for that caution whether they notice or not. Hire the agency that promises a higher number and you may get it, printed neatly in Ads Manager each Monday while Shopify revenue sits flat and the customer list ages. Hire the one that names break-even, checks new-customer share, reconciles revenue and asks to spend below today’s ROAS to buy strangers. The report may look worse before the business gets bigger. Ignore that distinction, and the ratio becomes the thing you grow.