Retention Metrics in a Case Interview: KISS
In one of my group coaching sessions this week, the question of retention metrics came up. A member of the group kept trying to come up with complex retention metrics in an effort to show off their ability to prioritize retention.
What they didn't realize is that retention metrics are complex in real life. Trying to show off your understanding of complex retention metrics in a short case interview is more likely to backfire than to succeed.
The interview solution: Keep it SIMPLE. You can simply measure retention with a supporting and counter metric pair like:
# active subscribers & # cancelled subscriptions
# DAU & # WAU (if DAU goes down but WAU goes up, it might indicate a coming problem)
In real life, we use retention cohorts to measure success as well as to dive into trends based on subscription start dates, location, user segmentation, etc. You need to show consideration of:
Distinct Groups
Acquisition Cohorts
Behavioral Patterns
Predictive analysis to forecast trends
Identify drop-off points
Equation 1 = # Active Users in X time period / total number of users in cohort
Equation 2 = N-day retention = # of users who first used the product on day X-N and returned on day X / The number of users who used your product on day X-N
Common Combos: Day 1, 3, 7, 14, and 30.
4 Key Retention Metrics All Require Math: CRR, RPR, Churn & LTV
Thus, there is no easy way to explain your command of retention cohorts in the typical product execution/analytics case question.
Don’t believe me? Try listing this as a metric in your next case interview and see if your partner follows. # Active Users in X time period / total number of users in cohort.
If you go for this approach, you open yourself up to poking and debate on:
how do you define active
which time period and why
cohort definition
This is not to say it can't be done, but I would wait for the interviewer to poke before adding such a complex metric to an already time-crunched question.
Additional Resources: