ROAS is reported as a ratio or multiple — 3.0 or 3:1 means three dollars of revenue per dollar spent — but the numerator hides a choice. Gross ROAS uses top-line attributed revenue; net ROAS subtracts returns, refunds, and discounts. The two diverge sharply in categories with high return rates or heavy promo dependence, and comparing across them is comparing different metrics.
The numerator is also a measurement choice, not a fixed truth: it is whatever the attribution model and window credit to ads. A 7-day click window credits fewer conversions than a 28-day click-or-view window, so identical underlying performance produces a lower measured ROAS under the shorter window. Post-iOS 14.5, platforms shortened windows and leaned on modeled conversions; measured ROAS fell across the industry even where ad efficiency was unchanged. The drop was a measurement artifact.
ROAS is a revenue ratio, not a profit ratio. Breakeven ROAS equals 1 / contribution margin — a 40%-margin product breaks even at 2.5 — so there is no universal target. A 3.0 on a 30%-margin SKU loses money after fulfillment; a 2.0 on a 70%-margin SKU prints. ROAS is also sensitive to AOV shifts at constant conversion rate. Read it alongside contribution margin and CAC before greenlighting spend.