Yelp product design: recommendations that help diners discover restaurants worth trying

Quick Overview

A product design prompt set at Yelp: build a system that recommends restaurants a user should try next. It tests whether a PM can define what counts as new, choose a user segment, handle cold start and review credibility, scope a first version, and measure whether suggestions lead to real, satisfying visits.

Yelp product design: recommendations that help diners discover restaurants worth trying

Company: Google

Role: Product Manager

Category: Product Design & Strategy

Difficulty: medium

Interview Round: Other

You are a product manager at Yelp. Design a system that gives users good recommendations for new restaurants to try. Explain whom it serves, which problems it solves, how the experience works and how recommendations are chosen, what you would launch first, and how you would measure success. ```hint Pin down what "new" means "New" could mean a restaurant that opened recently or one this user has never visited, and each reading changes the data you have and the problem you solve. Choose deliberately, then ask why a user would trust a suggestion enough to spend a meal on it. ``` ### Clarifying Questions - Does "new" mean restaurants that opened recently, or any restaurant this particular user has not tried? - Does trying a restaurant include takeout and delivery, or only eating in? - Should this live inside the existing app, for example on the home screen or in search, or as a separate touchpoint such as notifications or a weekly digest? - Is the main goal more user engagement, more reviews, or more value for the restaurants that advertise on Yelp? ### What a Strong Answer Covers - A goal tied to Yelp's purpose and business model, with a clear definition of "new". - A chosen user segment, with reasons, and that segment's specific pain points when deciding where to eat. - A recommendation approach: the signals used, cold start for new users and for restaurants with few reviews, and how suggestions earn trust. - Concrete product surfaces with a prioritized first version. - Success metrics that capture real visits and satisfaction, plus guardrails such as advertiser influence and review manipulation. ### Follow-up Questions - A restaurant opened last week and has two reviews. How would your system decide whether to recommend it? - Restaurants that pay for advertising want to appear in these recommendations. How would you handle that without losing users' trust? - How would you know that a user actually went to a restaurant you recommended?

Overview: A product design prompt set at Yelp: build a system that recommends restaurants a user should try next. It tests whether a PM can define what counts as new, choose a user segment, handle cold start and review credibility, scope a first version, and measure whether suggestions lead to real, satisfying visits.

Google logo
Google
May 10, 2025
mediumProduct ManagerOtherProduct Design & Strategy
0
0

You are a product manager at Yelp. Design a system that gives users good recommendations for new restaurants to try. Explain whom it serves, which problems it solves, how the experience works and how recommendations are chosen, what you would launch first, and how you would measure success.

Clarifying Questions Guidance

  • Does "new" mean restaurants that opened recently, or any restaurant this particular user has not tried?
  • Does trying a restaurant include takeout and delivery, or only eating in?
  • Should this live inside the existing app, for example on the home screen or in search, or as a separate touchpoint such as notifications or a weekly digest?
  • Is the main goal more user engagement, more reviews, or more value for the restaurants that advertise on Yelp?

What a Strong Answer Covers Guidance

  • A goal tied to Yelp's purpose and business model, with a clear definition of "new".
  • A chosen user segment, with reasons, and that segment's specific pain points when deciding where to eat.
  • A recommendation approach: the signals used, cold start for new users and for restaurants with few reviews, and how suggestions earn trust.
  • Concrete product surfaces with a prioritized first version.
  • Success metrics that capture real visits and satisfaction, plus guardrails such as advertiser influence and review manipulation.

Follow-up Questions Guidance

  • A restaurant opened last week and has two reviews. How would your system decide whether to recommend it?
  • Restaurants that pay for advertising want to appear in these recommendations. How would you handle that without losing users' trust?
  • How would you know that a user actually went to a restaurant you recommended?