Article originally posted here by Doordash, with permission. A/B tests and multivariate experiments provide a principled way of analyzing whether a product change improves business metrics. However, such experiments often yield flat results where we see no statistically significant difference between the treatment and the control.... Read more