ROAS is not enough: how to run Meta Ads by profit, not by credited revenue

ROAS is useful for orientation, but it is not a profit metric. It shows how much revenue an ad platform attributed to the money spent. It does not tell you whether the orders had strong margin, whether the customers were new, or whether the ad actually created demand. High ROAS can therefore be dangerously reassuring.

Why ROAS often misleads

ROAS is based on credited revenue. If Meta Ads shows an ad to someone who already knows the brand, has a product in cart or would have purchased anyway, the platform may receive credit for the sale. The report looks efficient. The business may not actually be growing.

The second issue is margin. An ad with ROAS 8 may sell a low-margin product with high return rates or expensive fulfilment. An ad with ROAS 3 may bring new customers with higher average order value and repeat purchase potential. ROAS hides these differences.

What to track instead of ROAS alone

The core question is: how much money is left after advertising? For ecommerce, it is better to work with gross profit per order, CPA and total profit volume. For lead generation, replace order value with qualified lead value, deal value or closed revenue.

• Gross profit per order: Formula / logic - Order value × gross margin; purpose - Shows the economic value of the order before ad cost.
• Profit after ad cost per order: Formula / logic - Gross profit per order - CPA; purpose - Shows whether a specific acquisition makes economic sense.
• Profit volume: Formula / logic - Total gross profit from orders - ad spend; purpose - Shows whether the channel grows the business, not just the report.
• New customer share: Formula / logic - New customers / all orders; purpose - Helps separate growth from recycled demand.

How to classify ads by profit and acquisition cost

A practical approach is to map ads by two dimensions: how much profit they bring and how expensive the acquisition is. This helps you identify which ads to scale, which ones to fix and which ones only look good because they generate cheap conversions.

• High profit, low CPA: Signal - Valuable customers at a reasonable cost.; recommended action - Scale gradually, protect engagement and create variations of the same angle.
• High profit, high CPA: Signal - Customer quality is strong, but acquisition is expensive.; recommended action - Improve hook, format, landing page or offer. Do not kill it without analysis.
• Low profit, low CPA: Signal - Efficient on paper but sells low-value orders.; recommended action - Monitor as a supporting asset. Do not mistake efficiency for growth.
• Low profit, high CPA: Signal - Expensive and low value.; recommended action - Reduce carefully, especially if it holds a meaningful share of spend.

How to build the reporting

The ad account usually does not contain the full economics. Meta may know order value, but it may not know margin, returns, fulfilment cost or whether the customer is new. That is why Meta Ads data should be combined with ecommerce, CRM or warehouse data. Even a simple export is useful if it connects campaign/ad ID, order value, margin and customer type.

• Track spend, purchases, CPA, revenue, average order value and estimated gross profit per ad.
• Separate new and returning customers whenever your ecommerce or CRM data allows it.
• Do not compare ads only by one day. Most accounts need a window of several days or weeks.
• For expensive products, track assisted signals such as product page visits, checkout, email sign-up or inquiry quality.

When high ROAS becomes a problem

High ROAS can be dangerous when it comes mostly from retargeting, brand-aware users or cheap low-margin orders. The ad looks great in the dashboard, but the company may not be getting bigger. The goal is not to maximise ROAS at any cost. The goal is to spend more while increasing total profit.

FAQ

Frequently Asked Questions

Should we stop using ROAS?

What if we do not know exact product-level margins?

How do we recognise an ad that only takes credit?