Optimize the Predictive Power of Your CX Measures
Composite CX measures made up of multiple metrics are often found to be more predictive than single metrics (NPS, CES, or CSAT) of key business outcomes, such as retention or lifetime value of a customer.
That said, building the right predictive composite measure requires testing of how each individual metric relates to the desired outcome, and taking time to revise the composite measure to improve the outcome prediction.
Using regression analysis, we look to see what contributes to a composite model in terms of precision, recall, and other attributes. For example, each single metric that is part of the composite should contribute independently to the prediction of the outcome. The individual metric should provide dimensionality to the overall composite measure and not be co-variates for each other.
This session will detail the necessary best practices of building an optimal composite measure that will predict critical business performance over time.