
Key takeaways
- Platform ROAS ignores product cost, shipping, payment fees, discounts and returns.
- Calculate contribution margin per order, then set an affordable cost per new customer.
- Report new and returning customer revenue separately, every week.
- Use realistic repeat-purchase data — not optimistic lifetime value — to justify higher acquisition costs.
Why platform ROAS flatters the numbers
Return on ad spend (ROAS) is the number most ad dashboards put front and centre: revenue attributed to ads divided by what you spent. It is easy to read, but it describes revenue — not profit.
Each platform also claims credit generously. A customer who clicked a Google ad, saw an Instagram ad and then opened your email may be counted as a conversion by all three. Add those reports together and you will often 'find' more revenue than actually reached your bank account.
Most importantly, ROAS ignores everything it costs to fulfil an order. A campaign at 4x ROAS on a low-margin product can lose money on every sale, while a 2x campaign on a high-margin range may be your most profitable spend.
How to calculate contribution margin
Contribution margin is what each order leaves behind after the variable costs of making and delivering it. It is the money available to pay for marketing, overheads and profit.
- Start with net revenue: order value after discounts and excluding sales tax.
- Subtract cost of goods sold: product cost, packaging and inbound freight.
- Subtract fulfilment: pick, pack, outbound shipping and any free-shipping subsidy.
- Subtract payment processing fees and marketplace or app transaction fees.
- Subtract your average cost of returns and refunds for that product range.
Turn margin into a target cost per new customer
Once you know contribution margin per first order, you know the most you can pay to acquire a customer and break even on that first purchase. If the first order leaves $40, a $40 cost per acquisition is break-even; $30 is profitable from day one.
Give that number to whoever manages your ads. Platforms can optimise toward a target cost per acquisition or a target ROAS — set those targets from your economics, not from industry averages or what last month happened to deliver.
Separate new customers from returning ones
Blended reporting hides whether you are winning new customers or simply paying to re-sell to people who would have bought anyway. Returning customers convert cheaply, so they make campaigns look far more efficient than they are.
Track new-customer revenue and cost per new customer as their own lines. Where platforms allow, exclude existing buyers from prospecting campaigns, and use email and SMS — which cost far less — to bring them back.
Use lifetime value carefully
If customers reliably buy again, you can afford to pay more than first-order break-even. The danger is optimistic projections. Use observed repeat-purchase data over 90 or 180 days, by acquisition channel if possible, and revisit the assumption every quarter.
A useful rule: only spend future margin you can see in your own cohort data today.
Build a simple weekly scorecard
You do not need complex attribution software to start. A weekly view that marketing and finance both trust will make budget decisions faster and calmer.
- Total marketing spend
- New customers acquired and cost per new customer
- Contribution margin after marketing
- Returning-customer revenue share
- Blended marketing efficiency (total revenue ÷ total marketing spend)
Frequently asked questions
What is a good ROAS?
There is no universal good ROAS. A break-even ROAS depends on your margins: a brand with 70% product margin can be profitable at 2x, while a 30% margin brand may need 4x or more. Work it out from your own contribution margin.
Is ROAS useless then?
No. It is a helpful in-platform signal for comparing campaigns and ads. It just should not be the number you use to decide whether marketing is profitable.
What is MER?
Marketing efficiency ratio is total revenue divided by total marketing spend across all channels. Because it does not rely on platform attribution, it is a useful sanity check alongside channel metrics.
This guide is general information, not tailored advice. Results vary by business, market and budget.

