Did it work?
Causal inference and experiments. For a leading Latin American beverage bottler, I set the priorities of a causal evaluation engine that measures the sales impact of a commercial initiative every two weeks for 1.5M+ customers (700k in Mexico and 800k in Brazil), using Matrix Completion for Causal Panels with Causal Forest as the benchmark. I also measure the impact of cross-company sales initiatives with difference-in-differences, synthetic control, instrumental variables, and causal forests (results were escalated to the CEO), and I design experiments: treatment and control groups, sample sizes, and randomization.