New-customer ROAS (NC-ROAS)

Blended ROAS credits every order the platform can see, including orders from loyal customers who would have bought regardless. For an established brand that share can be large, and it flatters every account that spends on retargeting or brand search. New-customer ROAS asks a narrower and more useful question: how much first-order revenue did the spend buy?

A worked example, as arithmetic. A brand spends $50,000 and the platforms attribute $200,000 of revenue, a 4.0x ROAS. Matched against the customer list, $80,000 of that came from first-time buyers. New-customer ROAS is 1.6x. The account did not get worse when that number appeared. It was always 1.6x at acquiring customers, and the other $120,000 was the existing base doing what it does.

How NC-ROAS is computed matters as much as the ratio. A pixel can only call a buyer new if it has not seen them before, which makes pixel-based new-versus-returning structurally optimistic: anyone who last bought on another device, before tracking existed, or before a cookie was cleared reads as new. Matching purchases against the brand customer list, with identifiers hashed, is more conservative and closer to the truth. Say which method a figure uses whenever it is quoted.

Use NC-ROAS for acquisition decisions and blended ROAS or MER for whole-business health. They answer different questions, and a healthy brand can show a low NC-ROAS alongside a strong MER when first orders are deliberately bought near breakeven and the profit arrives on repeat purchase. That is a legitimate strategy only if repeat behavior has been measured, not assumed.