Better Show and Movie Recommendations to Help Streaming Subscribers Find What to Watch
Company: Paramount Global
Role: Product Manager
Category: Product Design & Strategy
Difficulty: medium
Interview Round: Other
You are a product manager at Paramount Global working on one of the company's video streaming services. How would you improve the recommendation engine that suggests shows and movies to viewers? Decide what to improve and for whom, propose the changes, and explain how you would know they worked.
```hint Start from the moment of choice
Picture a viewer opening the app with no particular title in mind. Find the specific ways the current suggestions let that viewer down, and connect fixing them to why the service cares, before you talk about models.
```
### Clarifying Questions
- Which service is in scope: a paid subscription service such as Paramount+, or a free ad-supported one such as Pluto TV, where recommendations also shape ad revenue?
- Is the goal mainly keeping subscribers from cancelling, increasing viewing time, or giving the company's own originals more exposure?
- Which surfaces count as the recommendation engine: home-screen rows, the suggestion shown after an episode ends, search results, or notifications and emails?
- Is there a known problem behind this request, such as falling engagement with recommended rows, or is the question open-ended?
### What a Strong Answer Covers
- A business goal for recommendations tied to how the service makes money, and the success measure that goal implies.
- A segmentation of viewers or viewing moments, with one segment chosen and justified.
- Concrete pain points with today's recommendations mapped to product and model changes, prioritized by impact and effort.
- An experiment plan with a primary metric and guardrails that stop the team from optimizing clicks over real viewing satisfaction.
- Risks such as cold start for new viewers and new titles, filter bubbles, viewing-data privacy, and the tension between promoting the company's own content and serving the viewer.
### Follow-up Questions
- How would you balance recommending what a viewer is most likely to watch against what the content team wants to promote, such as a new original series?
- Several people share one profile on the account. How would you detect that, and what would you change?
- An A/B test shows more clicks on recommendations but no change in total viewing time. What do you conclude, and what do you do next?
Overview: A product improvement prompt for a product manager role at Paramount Global: improve how a video streaming service recommends shows and movies to its viewers. It tests setting a goal, choosing a viewer segment, turning recommendation pain points into product and model changes, and measuring the result with experiments and guardrails.