You launched a retention-facing change on a social product, and the week after launch your retention metric moved. The same period also overlaps with a known seasonal pattern, so you need to separate product impact from calendar effects.
How would you diagnose whether a launch caused a change in retention versus a seasonal effect?