A 4x ROAS sounds like a business that prints money. Then the month closes and the bank balance hasn't moved. Nothing is broken. ROAS in Ads Manager measures revenue Meta can connect to an ad, and two things sit between that and profit: what each order costs you, and how many of those sales the ads really caused.
- ROAS is revenue divided by ad spend. Product cost, shipping, card fees and refunds all come out before profit.
- Your break-even ROAS is 1 divided by the share of each order you keep. At 42% kept, it's about 2.4x.
- Meta counts sales after views and engagements as well as clicks, so its ROAS includes sales that would have happened anyway.
- Judge profit on blended numbers from your store: total revenue divided by total ad spend.
- Then cut spend that only claims sales, and raise what each order is worth.
Where a 4x ROAS goes
Take an illustrative store with an $80 average order. These aren't benchmarks, just a typical shape. Swap in your own figures.
| Per $80 order | Amount | Share |
|---|---|---|
| Order value | $80.00 | 100% |
| Product cost | - $32.00 | 40% |
| Shipping and packing | - $8.00 | 10% |
| Card fees | - $2.40 | 3% |
| Returns and refunds | - $4.00 | 5% |
| Left for ads and profit | $33.60 | 42% |
| Break-even ROAS | 2.38x | 80 / 33.60 |
Illustrative figures. Your break-even ROAS is your order value divided by what's left after every cost except ads.
At a real 4x, this store spends $20 of ads per order and keeps $13.60. That's a good business. Now suppose only six in ten of the sales Meta reports were caused by the ads, and the rest would have happened anyway.
- ROAS: 4.0x in Ads Manager. 2.4x if only six in ten sales were caused by the ads.
- Ad cost per sale: $20.00 in Ads Manager. $33.33 for each sale the ads really caused.
- Profit per sale, before fixed costs: $13.60 in Ads Manager. $0.27 in the bank.
- Illustrative figures. Fixed costs such as staff, software and management fees still come out of that last line.
The same ad account, the same spend, and the profit has gone. That's the gap most store owners feel and can't name. What is a good ROAS for Facebook ads covers the break-even formula and benchmarks in more detail.
Why Meta's ROAS is higher than your bank balance
It counts views and engagements, not just clicks
Meta's standard attribution can credit a sale within 1 day after an impression of your ad, and within 1 day after a non-link engagement, which for video includes a 5-second play. Someone who scrolled past your ad and bought from a Google search an hour later can count as a Meta sale.
It claims sales from people who were already buying
Retargeting past visitors and past customers produces the best ROAS in most accounts, because those people were close to buying anyway. Some of that is real persuasion, some of it is the ad standing near a sale that was coming. Meta offers an incremental attribution model that predicts whether a conversion is caused by an ad, and the gap between it and the standard model tells you how much of your ROAS is credit-taking.
Every channel claims the same sale
Meta, Google and your email tool each count a sale by their own rules. One order can appear in all three reports. Meta itself says that if you use an outside analytics tool, evaluate performance in those tools.
Revenue isn't margin
ROAS treats a $80 order the same whether you keep $34 of it or $10. Discount codes, free shipping thresholds and low-margin products all lower the share you keep without changing the ROAS.
Refunds don't come off
Ads Manager counts the purchase when it happens. A return two weeks later stays in the ROAS unless your setup sends it back. Check what value your Purchase event sends, too: Shopify purchase tracking covers the common setups.
The latest days aren't finished
Click-through attribution runs up to 7 days, so a recent week keeps gaining sales as late conversions arrive. That makes this week look worse than it is, and last month look better than the day you judged it. Why ROAS drops when nothing changed explains the reporting lag.
How to find your real number
- Work out what's left per order. Product, shipping, fees and refunds, as a share of the average order. That share sets your break-even ROAS.
- Measure blended ROAS from the store. Total store revenue divided by total ad spend on every channel, same dates. It can't be inflated by any one platform.
- Split first orders from repeat orders. Ads that mostly reach returning customers look brilliant and add little. Your store's order export shows which customers are new.
- Compare Meta's settings. Look at 7-day click only, and at the incremental model if your account has it. If ROAS halves, half of it was credit-taking.
- Test before you trust. Meta's A/B tests split audiences so nobody sees both versions. Turning campaigns off and on by hand doesn't tell you much.
If blended ROAS sits below break-even while Ads Manager shows 4x, the ads aren't the only problem, and spending more won't fix it. Why ROAS drops when you raise the budget covers what happens if you try.
What to change
- Cut spend that only claims sales. Retargeting and past-customer audiences that show a high ROAS but barely move blended revenue. Cap them rather than feed them.
- Put budget where new customers come from. Prospecting usually reports a lower ROAS and brings more first orders. Judge it on new-customer revenue in the store.
- Raise what each order is worth. Bundles, a threshold for free shipping that sits above your average order, and fewer blanket discount codes. Each one lowers break-even ROAS.
- Fix tracking before judging anything. Missing or doubled Purchase events distort every number above. The Conversions API closes most of the gaps the Pixel leaves.
- Refresh creative on a schedule. Tired ads push costs up slowly enough to miss. Creative testing for ecommerce shows a cadence that works.
When it's worth a second pair of eyes
When Ads Manager says the ads work and the accounts say they don't, someone has to reconcile the two, campaign by campaign, before the next budget decision. That reconciliation is the first thing we do in our Meta ads management for ecommerce: blended numbers, new-customer revenue and break-even ROAS before any change to spend.
Questions people ask
Why is my ROAS good but I'm not making money?
Usually two things at once. ROAS counts revenue, not profit, so product cost, shipping, fees and refunds still have to come out of it. And Meta's ROAS includes sales it would have got anyway, from people who saw an ad and were already going to buy. Together they can turn a reported 4x into break-even.
What ROAS do I need to make a profit?
More than your break-even ROAS, which is 1 divided by the share of each order left after product, shipping, fees and refunds. If 42% of an order is left, break-even is about 2.4x. Measure it against the revenue in your store, not the number in Ads Manager.
Why does Meta show more sales than Shopify?
Meta counts a sale if it happened within its attribution window after a click, a view or an engagement with an ad, so it can claim sales other channels also claim. Your store counts each order once. Add up what Meta, Google and your email tool each report and the total is often more than you sold.
Should I change the attribution setting to 7-day click?
Report on it first. Meta lets you compare attribution settings, and click-only numbers are a stricter view of what the ads did. Changing the setting on a live ad set also changes what Meta optimizes for, so compare before you change anything.
What is blended ROAS, or MER?
Total store revenue divided by total ad spend across every channel, for the same period. It can't be inflated by any one platform's attribution, which makes it the number to judge profit on.
When is it worth getting help?
When the store's blended numbers and Ads Manager tell different stories and you can't tell which campaigns earn their spend. We're not the right fit if margins can't cover ads at any realistic ROAS, or if the store has no sales yet without ads.
ROAS looks fine and the profit isn't there? Book a strategy call and we'll reconcile your Ads Manager numbers against the store with you.
