High ROAS Can Still Hide Losing Orders

High ROAS can still hide losing orders. That sounds backwards until you compare the ads dashboard with the business math behind each sale.

On September 8, Search Engine Land published a useful warning about an ecommerce account reporting 11x ROAS while losing money on the average order. The issue was not that ROAS was fake. The issue was that the metric ignored returns, tax, cost of goods, fulfillment, payment fees, inventory timing, and brand demand that would have converted anyway.

That is the paid ads trap for ecommerce brands heading into Q4: the campaign can look efficient while the business gets less healthy. Before increasing spend, audit what ROAS is actually measuring.

1. Rebuild ROAS Around Contribution Margin

ROAS answers one narrow question: how much reported revenue did the ad platform connect to spend?

It does not answer the better question: how much money did the business keep after the sale?

Start with a contribution view. For each major product category, document average order value, expected return rate, discounts, tax treatment, cost of goods, outbound shipping, return shipping, payment fees, pick-and-pack cost, and any marketplace or app fees tied to the order.

Then calculate break-even ROAS by category. A product with a 50% gross margin can tolerate a different paid media target than a product with a 22% gross margin. A campaign selling low-margin clearance inventory should not be judged by the same dashboard target as a full-price hero product.

This is where many ecommerce accounts get into trouble. Discounts improve conversion rate, which can lift platform ROAS. But if the discount pushes the order below contribution break-even, the campaign is celebrating a sale the finance team has to absorb.

At Emarketed, we have seen the other side of that math when paid media, merchandising, and conversion work together. Sector 9 Skateboards grew sales by 165% and increased marketing-attributed revenue by 254%, while Nutcase Helmets grew sales by 306% and sales attributed to marketing by 774%. The lesson from those campaigns is not that higher spend fixes everything. It is that paid ads strategy has to be tied to the economics of the sale, not only the platform return.

2. Separate Brand Demand From New Demand

Blended ROAS is where weak decisions hide.

Brand search, remarketing, returning customers, and email-supported purchases often report strong returns because the buyer already knows the company. Those campaigns may still deserve budget, especially when competitors bid on the brand or when returning customers need a reminder. But they should not be mixed with nonbrand prospecting when the team is deciding whether paid media is creating growth.

Create separate views for:

• branded search
• nonbrand search
• Shopping and Performance Max by new versus returning customer
• remarketing
• prospecting paid social
• email-assisted revenue

If brand search runs at 15x ROAS and nonbrand runs at 2.8x, the blended account number may look comfortable while the actual acquisition engine is barely above break-even. That does not mean nonbrand is failing. It means the business has to judge it against new-customer margin, repeat purchase rate, lifetime value, and inventory goals.

The key question is not “what is our account ROAS?” It is “which campaigns create orders we would not have won without paid media?“

3. Add Returns Before The Dashboard Looks Final

Returns are one of the fastest ways for ecommerce ROAS to lie by omission.

The order appears in the ad platform when the customer buys. The return shows up later in Shopify, the ERP, the warehouse, customer support, or the payment processor. If the ad account never receives that adjustment, it keeps optimizing toward a cleaner version of the business than the one actually operating.

This matters most in categories where sizing, fit, color expectations, shipping damage, gifting, or comparison shopping drive higher returns. Apparel, accessories, home goods, beauty, consumer electronics, and seasonal products can all look profitable at the click level while return-adjusted contribution tells a different story.

Build a return-adjusted reporting table every week:

• spend
• platform revenue
• estimated kept revenue after returns
• contribution after product cost and fulfillment
• new customers
• repeat customers
• return rate by campaign, SKU, and landing page

That table does not need to be fancy. It needs to be trusted. If the ad platform says a campaign produced $80,000 in revenue but the business kept only $52,000 after returns and discounts, the budget decision should use the second number.

4. Match Bidding Targets To Inventory Reality

Profit efficiency is not always the same as business health.

A strict ROAS target can push the algorithm toward the safest conversions: best sellers, brand-aware buyers, returning customers, discounted products, and lower-risk segments. That can protect short-term efficiency while leaving inventory sitting in the warehouse.

For ecommerce brands, inventory has a clock. Seasonal stock loses value. Products tied to trends, holidays, sizes, colors, or replenishment cycles may need different bidding goals at different moments. A lower ROAS can sometimes create better total contribution if it moves inventory before deeper markdowns are required.

Do not set one account-wide target and walk away. Assign a job to each product group:

• protect margin
• acquire new customers
• clear seasonal inventory
• grow a category
• defend brand demand
• support a launch

Those jobs require different bids, budgets, creative, and landing pages. A campaign built to liquidate slow-moving inventory should not be punished because it fails to hit the same target as a high-margin evergreen SKU. A campaign built to acquire new customers should not be compared with branded demand capture.

5. Decide What Gets Reported To Leadership

The report your team sends shapes the budget conversation.

If leadership only sees ROAS, the natural question becomes: why not spend more on the highest ROAS campaign? That can move budget toward brand demand, returning customers, and discounted products while starving the campaigns that create tomorrow’s buyers.

Replace the top-line ROAS slide with a tighter paid media scorecard:

• contribution margin by campaign type
• new-customer cost and payback period
• return-adjusted revenue
• inventory sell-through impact
• blended spend as a percentage of gross profit
• assisted revenue from email, organic search, and paid social
• campaign-level next action

For ecommerce brands, a strong paid media report should tell the operator what to do next. Raise budget, cap budget, change bids, adjust the offer, fix the landing page, split brand from nonbrand, clean up conversion values, or hold spend until inventory catches up.

That is also why broader paid media optimization should connect channel performance to merchandising and website conversion. The ad account is only one part of the revenue system.

What To Fix This Week

Do not wait for a bad month to audit ROAS.

Pick your five highest-spend campaigns and rebuild the numbers using kept revenue, contribution margin, return rate, new-customer share, and inventory status. Then decide whether each campaign deserves more budget, less budget, a different target, or a cleaner measurement setup.

High ROAS is useful when the inputs are honest. Without margin, returns, inventory, and incrementality, it can become a polished way to lose money faster.