Plan a New Image Search Product: Target Users, First Release and Success Metrics
Company: Google
Role: Product Manager
Category: Product Design & Strategy
Difficulty: medium
Interview Round: Other
You have been asked to build an image search engine. Explain how you would approach it as the product manager: who it should serve and for what, what the first version should include, how it would find and rank images at a high level, and how you would know it is working.
```hint Decide what kind of search it is
People use image search to find pictures that match some words, and also to start from a picture and learn what is in it or where it came from. Choose which of these jobs to serve first; the users, the technology and the metrics all follow from that choice.
```
### Clarifying Questions
- Is this a standalone product, or part of an existing web search engine that already has a crawl, users and an ads business?
- Should it handle text queries that return images, image queries (an upload or a camera photo) that return similar images or information about the picture, or both?
- Is the goal broad consumer search, or a specific need such as shopping, stock photography or searching a person's own photo library?
- Which platforms come first: desktop browsers, mobile web, or a mobile app with camera access?
### What a Strong Answer Covers
- A clear goal, a segmentation of image-search users and the jobs they bring, and a reasoned choice of whom to serve first.
- Pain points with today's ways of finding images, turned into prioritized features and a focused first version.
- A product-level explanation of how images are collected, understood, matched to queries and ranked, with the key quality and cost trade-offs.
- Success metrics that capture whether people found what they needed, with guardrails.
- Risks particular to images: copyright and creator traffic, explicit or harmful content, face search and privacy, and AI-generated or misleading pictures.
### Follow-up Questions
- With camera search, anyone could photograph a stranger and search for them. What would you allow, and what would you block?
- How would you judge result quality when a user clicks nothing because the thumbnails already answered the question?
- Photographers and publishers say your product shows their images without sending them visitors. How would you respond while keeping the product useful?
Overview: A product design question for product manager interviews about building an image search engine from the ground up. It tests choosing which users and search jobs to serve first, turning their pain points into a focused first version, explaining at a high level how images are found and ranked, and defining success metrics and image-specific risks.