ROAS is the most dangerous ecommerce metric
Return on ad spend measures efficiency within a channel. It tells you nothing about whether the channel is contributing to the business.
A brand reports a 6x ROAS on Meta. The team celebrates. The CFO asks why cash is tight.
Both things are true. They describe different systems.
ROAS measures the ratio of attributed revenue to ad spend within a channel. It does not account for returns, fulfilment costs, cannibalisation of organic, or the marginal cost of the next order.
In practice, most ecommerce ROAS calculations are inflated by three mechanisms.
First, attribution windows capture purchases that would have happened anyway. A customer who searched your brand name, saw a retargeting ad, and then purchased gets attributed to paid. That isn't incrementality. It's accounting.
Second, high ROAS often correlates with retargeting, which by definition targets existing demand. The metric looks efficient precisely because the hard work — creating demand — was done elsewhere.
Third, ROAS doesn't decay. A 6x ROAS on £10,000 spend might become 2.8x at £50,000. But the dashboard still shows the blended number. Marginal returns are invisible until the budget has already been committed.
The teams that navigate this well don't optimise for ROAS. They optimise for contribution margin per incremental order — which requires connecting ad platform data to warehouse-level economics.
That's a systems problem, not a marketing one. And it's why the metric most teams use to justify spend is often the same metric that obscures the real cost of growth.