Personalizing Yelp's restaurant picks after users say they feel generic

Quick Overview

A product sense exercise on responding to feedback that Yelp's restaurant recommendations feel generic. It tests how a PM confirms and sizes the complaint, picks a user segment, diagnoses why suggestions miss, proposes personalization changes, and sets success metrics with guardrails.

Personalizing Yelp's restaurant picks after users say they feel generic

Company: Google

Role: Product Manager

Category: Product Design & Strategy

Difficulty: medium

Interview Round: Other

Yelp keeps hearing from users that its restaurant recommendations do not feel personal: the suggestions look like what anyone nearby would see. As the product manager, what would you do about it? ```hint Pin down the miss "Not personal enough" can mean several different failures. Work out which surface and which users the complaint comes from, and what is going wrong for them, before you change the recommendations. ``` ### Clarifying Questions - Where does the feedback come from (app-store reviews, in-app surveys, support tickets), and roughly how many users does it represent? - Which surface do users mean: the suggestions on the home screen, results for broad searches such as dinner nearby, or recommendation emails and notifications? - Is the complaint concentrated among new users with little history, or among long-time users who expect Yelp to know their taste? - Is success more diners acting on a suggestion, such as booking, calling or getting directions, or more engagement inside the app? ### What a Strong Answer Covers - Validating and sizing the feedback, and narrowing it to a specific surface and user segment - A diagnosis that separates the possible causes of generic suggestions and checks them against data - Concrete, prioritized changes across data capture, ranking and user controls - A success metric tied to diners acting on suggestions, with guardrails for diversity, businesses and privacy - Risks such as filter bubbles, creepiness and newly opened restaurants losing exposure ### Follow-up Questions - How would you personalize for someone who has never rated, saved or booked anything? - If taps on suggestions rise but bookings and calls stay flat, what would you conclude? - How would you keep personalization from burying newly opened restaurants with few reviews?

Overview: A product sense exercise on responding to feedback that Yelp's restaurant recommendations feel generic. It tests how a PM confirms and sizes the complaint, picks a user segment, diagnoses why suggestions miss, proposes personalization changes, and sets success metrics with guardrails.

Google logo
Google
Sep 8, 2024
mediumProduct ManagerOtherProduct Design & Strategy
0
0

Yelp keeps hearing from users that its restaurant recommendations do not feel personal: the suggestions look like what anyone nearby would see. As the product manager, what would you do about it?

Clarifying Questions Guidance

  • Where does the feedback come from (app-store reviews, in-app surveys, support tickets), and roughly how many users does it represent?
  • Which surface do users mean: the suggestions on the home screen, results for broad searches such as dinner nearby, or recommendation emails and notifications?
  • Is the complaint concentrated among new users with little history, or among long-time users who expect Yelp to know their taste?
  • Is success more diners acting on a suggestion, such as booking, calling or getting directions, or more engagement inside the app?

What a Strong Answer Covers Guidance

  • Validating and sizing the feedback, and narrowing it to a specific surface and user segment
  • A diagnosis that separates the possible causes of generic suggestions and checks them against data
  • Concrete, prioritized changes across data capture, ranking and user controls
  • A success metric tied to diners acting on suggestions, with guardrails for diversity, businesses and privacy
  • Risks such as filter bubbles, creepiness and newly opened restaurants losing exposure

Follow-up Questions Guidance

  • How would you personalize for someone who has never rated, saved or booked anything?
  • If taps on suggestions rise but bookings and calls stay flat, what would you conclude?
  • How would you keep personalization from burying newly opened restaurants with few reviews?