Marketing & Sales
Observation 005
· 2 min read
A 4.0 ROAS that returns ninety cents.
Attribution asks who touched the sale. Incrementality asks who would not have bought otherwise. The difference can be most of the return.
One number
$0.90
gross profit per $1 of spend, from customers who were not already yours
One mechanism
Attributed return counts existing customers the ads reached. New-customer return, after margin, is the number that funds growth.
One decision
Ask for new-customer ROAS by cohort, and the test design that would prove incremental value.
Hypothetical. The ad platform reports a 4.0 ROAS. Then someone splits the purchases, and 55% came from people already in the CRM.
New-customer ROAS is 4.0 × 45% = 1.8. At a 50% gross margin, that is 0.90. Every dollar spent returns ninety cents of gross profit from people who were not already customers, before shipping, fees, returns or a single salary.
The dashboard is accurate. It is answering a different question. Some of those repeat purchases were caused by the ads; which ones is exactly what a holdout test is designed to reveal.
An audit alone cannot prove incrementality. A test can.
Evidence · Modeled · Hypothetical; revenue split assumed proportional