Introduction to the Case Study Interview
A data science interview loop may include a product or case study interview. Questions could range from “guesstimation” questions, such as, “How many tennis balls can you fit in a plane?” to business-relevant case studies, such as, “What would you do if Youtube noticed a decline in video views?”
The plane “guesstimation” question isn’t common anymore, because of how frequently it was asked. However, it’s representative of a range of “guesstimation” style questions. In these types of questions, you’re expected to make estimates for hard-to-guess quantities by reasoning through the problem and breaking it down into smaller components.
What interviewers are looking for
When asking case study interview questions, interviewers are assessing how you:
- Approach unstructured problems
- Share your thought process with others
- Make decisions with data
How to ace the case interview
To ace the interview, you should aim to:
- Engage the interviewer
- Break the problem down into a clear structure that makes sense
- Leverage that structure to “solve” the problem.
“Solving” the problem
The case interview often abstracts away the messiness of the real-world, so “solving” the problem doesn’t actually involve being “exactly” right. It’s usually more about being “directionally” right and identifying the most pertinent “sub-factors” or “drivers” of the problem. This can also often involve making smart assumptions and confirming them with your interviewer.
For example, in the question about how many tennis balls would fill a plane, you don’t need to correctly “guess” the real answer to do well. Rather, you should focus on demonstrating your ability to reason with two quantities: the volume of a tennis ball and the volume of a standard plane.
In How to Answer Case Study Interview Questions, we’ll describe how to approach these questions through a step-by-step framework.