Measure Relevant Feed Success
Company: LinkedIn
Role: Data Scientist
Category: Analytics & Experimentation
Difficulty: hard
Interview Round: Technical Screen
LinkedIn changes the Home Feed default experience from showing all content to showing only content predicted to be relevant to each viewer.
How would you evaluate whether this change is successful?
Your answer should cover both sides of the marketplace:
- **Viewer side:** engagement, satisfaction, content discovery, and retention.
- **Poster side:** reach, quality of engagement, incentives to keep posting, and fairness of distribution.
Please describe:
- The primary success metrics and guardrail metrics you would choose.
- How you would design an A/B test for this change.
- How you would define and measure "high-quality engagement" rather than raw engagement alone.
- What risks this change may create, such as filter bubbles, lower creator reach, reduced diversity, or concentration of attention on a small set of posters.
- How you would decide whether the launch is beneficial in the short term and sustainable in the long term.
Also answer this follow-up: if viewers show higher click and like rates under the new relevant-only feed, but posters worry that their content reaches fewer people, how would you analyze and respond to that concern?
Quick Answer: This question evaluates a data scientist's competency in product analytics and experimentation, specifically metric definition, A/B test design, interpreting engagement versus high-quality engagement, and balancing viewer-side and poster-side marketplace effects like reach, incentives, and fairness.