Challenge

From a near infinite set of interactions and events that a user can perform on an app / site, what set has the maximum pay offs in terms of the Liftetime Value of the User? Identifying this involves a lot of moving pieces, including data consolidation, engineering, finance metrics allocations, and user cohorts. People signing up in different months, across events/non-events, weekends/weekdays, across geographies, across different point-of-access, across different starting-points adds complexity to normalize the comparison.

Approach

User Actions data were first consolidated from the multiple places they were polling into one location. They were then integrated with the Visit-level table prepared.Three kinds of analysis were performed using the new data:A. What happens to the value of people who perform a certain action vs those who don’t. B. What is the incremental change in behavior in those who have performed a certain action. C. What is the inflection point or the critical volume at which the impact tapers off.

Outcome

Marketing department was allowed to allocate an extra five percent budget to plan activities towards nudging key User Actions that yield higher user actions.Assessing impact of user actions has become a norm for planning campaigns, marketing and CRM activities.