How should you renew or replace a show?
Company: Capital One
Role: Data Scientist
Category: Analytics & Experimentation
Difficulty: medium
Interview Round: Technical Screen
You are a data scientist at a streaming company similar to Netflix or Hulu. Your team helps decide whether existing series should be renewed or canceled, and whether new series should be launched instead.
Assume the following unless you explicitly state otherwise:
- Planning horizon: 2 years
- Each show releases 1 season per year
- "Viewers per season" means unique viewers for that season
- Revenue per viewer is the incremental revenue attributable to that viewer for that season
- Annual cost is incurred once per year
- Startup cost for a new show is a one-time upfront cost in year 1
- Ignore taxes and discounting
Case questions:
1. In general, how would you decide whether a show should be renewed? Provide a structured framework that goes beyond short-term accounting profit. Consider financial performance, subscriber acquisition and retention, strategic value, portfolio effects, and risk. Also explain what data you would want and what confounding or attribution issues could make the decision difficult.
2. Consider an existing show, **The Analyst**:
- Viewers per season: 5 million
- Revenue per viewer: $15
- Annual cost: $50 million
What is the 2-year profit for The Analyst?
3. Consider launching a new show, **Sharkbank**, under a 2-year contract.
If Sharkbank is successful:
- Viewers per season: 7 million
- Revenue per viewer: $15
- Annual cost: $60 million
- One-time startup cost: $20 million
If Sharkbank is unsuccessful:
- Viewers per season: 4 million
- Revenue per viewer: $15
- Annual cost: $60 million
- One-time startup cost: $20 million
(a) What is the 2-year profit for Sharkbank in the success scenario?
(b) What is the 2-year profit for Sharkbank in the failure scenario?
(c) If success and failure are each 50% likely, what is Sharkbank's expected 2-year profit?
4. Which would you choose, The Analyst or Sharkbank? Do not rely only on expected value. Discuss risk, volatility, portfolio strategy, cash-flow stability, and any strategic upside or missing information that could change the answer.
5. If you wanted to improve The Analyst's profit, what levers would you consider, and how would you measure whether those actions create truly incremental value rather than just shifting behavior that would have happened anyway?
Quick Answer: This question evaluates financial modeling, causal attribution, probabilistic decision-making, and strategic portfolio analysis skills for a Data Scientist in Analytics & Experimentation.
