You work on a consumer financial product and your team launches a new feature intended to help more users complete a key funnel step. Early results suggest the treated group is moving faster through one part of the funnel, but it is unclear whether the feature is creating new value or simply cannibalizing behavior that would have happened later or elsewhere in the journey.
How would you determine whether the feature is truly incremental versus just shifting behavior from one part of the funnel to another? Walk through how you would design the experiment and evaluate the results.