Incrementality and Counterfactual Measurement: The Question Behind Real Marketing ROI
A business-friendly introduction to treatment, control, counterfactuals, incremental revenue and why observed conversions are not the same as caused conversions.
Incrementality is a difference, not a dashboard column
Incrementality is the change in an outcome caused by an intervention compared with what would have happened without it. If a campaign reports 1,000 purchases, the incremental question is not whether those purchases followed marketing exposure. It is how many more purchases occurred because the marketing existed.
The counterfactual is the core difficulty
The problem is that the no-marketing outcome for an exposed customer is unobservable at the same moment. Measurement therefore needs a credible substitute: a randomized control group, a geographic control, a time-based counterfactual model or another design whose assumptions are explicit.
Incremental revenue is not attributed revenue
Attributed revenue can include customers who were already going to buy. Incremental revenue attempts to isolate the extra revenue created by the intervention. Incremental return on ad spend then compares that causal lift with the causal change in spend. This is a different quantity from platform return on ad spend.
Good experiments start with the estimand
Before choosing a method, define exactly what you want to estimate: incremental purchases, incremental qualified leads, incremental gross profit, search lift or something else. Define the population, period and treatment. A perfectly executed experiment with a vague target can still produce a decision nobody knows how to use.
Measurement maturity is a portfolio
Not every decision deserves the same method. Use platform reporting for daily operations, controlled experiments for high-value causal questions, and broader models for budget allocation across channels and time. The strongest system connects the methods instead of forcing one tool to answer everything.