You launched a retention-facing change on a Meta surface such as Instagram Reels or Facebook Groups, and retention moved right after launch. The same period overlaps with known seasonal traffic patterns, so a simple before-after read is not enough.
How would you diagnose whether a launch caused a change in retention versus a seasonal effect?
Primary metric is retention, often read alongside AARRR funnel movementUseful covariates include pre-period retention, prior Reels engagement, and Instagram Save behaviorIf this came from an experiment, check Sample Ratio Mismatch before trusting treatment-control comparisonsWatch for novelty effect if the post-launch lift decays quickly