How to Determine Causal Effects when A/B Tests are Infeasible through Adopter Analysis

Tags: #DataScience #CausalInference #DifferenceinDifference #Matching #MatchedDiD #PropensityScoreMatching

E-commerce businesses often introduce treatments (decisions) for their customers and want to measure the impact of those treatments on metrics of interest. For example, an e-commerce subscription business often faces the challenge of determining whether changes in member benefits result in increased spending or retention. The most reliable way to assess the causal impact of such changes is through randomized experiments via A/B tests. In some cases, conducting an A/B test is not …

analysis article blog global link medium tech tests topic walmart