ROAS is useful shorthand, but it can flatter a campaign that loses money and punish one that builds a business. Before you treat it as the target, decide what the business actually needs: profitable revenue, qualified pipeline, repeat purchase, or market entry.
Start with contribution margin, not the ratio
A 4x ROAS on a product with a 20 percent margin can be worse than a 2x ROAS on a product with a 60 percent margin. Build a simple margin model by product or category, then set a target that accounts for fulfilment, returns, and platform costs.
Separate acquisition from returning demand
- Report new-customer revenue separately from returning-customer revenue.
- Track cost per new customer against lifetime value, not just the first order.
- Use incrementality tests for brand campaigns and broad match, where attribution is most generous.
- Check whether lead quality holds after sales follow-up, not only in the platform dashboard.
Fix tracking before scaling
If conversions are double-counted, ROAS is fiction. Verify the primary conversion action, exclude internal traffic, use server-side tracking where appropriate, and reconcile platform numbers with your analytics and CRM. A smaller, trustworthy number is more useful than a large, ambiguous one.
Judge changes over a window, not a day
Bid changes, creative fatigue, and seasonality all move results. Give tests enough time and budget to produce a signal, and document what else changed during the same window. Scaling on one good day is how accounts become unstable.
If you want a second opinion on your account, bring the campaign structure and the last 90 days of data to a strategy call. We will tell you what we would fix first.