Expected value modeling optimizes marketing campaigns by assigning costs and benefits to each prediction outcome. Starting with a purchase likelihood model, you calculate the expected profit for targeting each customer by weighing purchase probability against campaign costs. For an e-commerce example with $50 profit per sale and $1 cost per click, the optimal strategy targets customers with >2% purchase probability. Profit curves extend this by showing expected returns across all targeting thresholds, helping balance budget constraints with ROI. Success requires deep business knowledge to accurately quantify costs, benefits, and interpret model outputs.
Table of contents
What is Expected Value Modeling?Start with a Purchase Likelihood ModelImplementing Expected Value ModelingTaking it one step further with Profit CurvesIt starts and ends with business knowledgeReferences96 Impressions