Pramp Engagement
You’re a data scientist at Pramp and the product team wants to improve user engagement. What metrics would you choose and how would you tackle this?
Make sure to read our Interviewer Guide. While you’re at it, you should read our Interviewee and Interview Structure guides to increase you own chances of acing your interview.
Make sure your peer identifies the goals of Pramp’s product team before determining what metrics to evaluate. For example, this team might be concerned with growth.
Then have your peer define a metric that is in line with their goal of growth; for instance, one metric might be monthly active users.
Afterward, encourage your peer to think about the variables they think would matter to move that metric. For example, specific user characteristics (age, location, number of FB friends, etc.) and/or user behavior characteristics (online behavior, device, etc.).
Then ask your peer to pick what model they’d use, for instance, a tree-based classifier like Random Forest. Make sure your peer gives a justification for their model and feel free to ask them a follow-up question on why they chose one model/algorithm over another.
Next, prompt them to come up with a few possible model output scenarios. See the interviewee/interviewer guides for an example.
Finally, make sure your peer concludes with next steps about what actions they’d suggest the team should take to improve engagement, given the company’s overall goal and the possible model output scenarios.
[Pra]ctice [m]akes [P]erfect!