Analyze homepage drop and feed ranking

You are interviewing for a product data science role at LinkedIn. Answer the following two product-sense questions. 1. **Diagnose a drop in Home Page -> Profile Page traffic** LinkedIn notices that weekly traffic from the Home Page to the Profile Page has declined materially. You are asked to determine whether this is a product problem, a logging issue, a traffic-mix shift, or an intentional product improvement. Describe how you would investigate this decline. Your answer should include: - how you would define the primary metric and its denominator - sanity checks for bugs, instrumentation changes, and release changes - what macro metrics you would review first - how you would analyze user paths after landing on the Home Page - how you would segment the analysis by platform, geography, member cohort, and traffic source - how you would reason about a case where overall session duration does **not** change, even though Home -> Profile visits fall You may assume LinkedIn recently launched a new inline profile preview feature on the Home Page: when a user hovers over a member's name or card, they can see key profile information without opening the full Profile Page. 2. **Measure the success of changing the feed default from "All" to "Relevant only"** LinkedIn plans to change the Home Feed default ranking from showing all content to showing only content deemed most relevant to each viewer. Design an experiment and success framework for this change. Your answer should include: - the core product hypothesis - the experiment design and unit of randomization - primary viewer-side success metrics - poster-side metrics and possible concerns from content creators - guardrail metrics - short-term versus long-term tradeoffs - how you would interpret a result where engagement rate improves for viewers, but some posters receive less reach Be explicit about metric tradeoffs, potential selection bias or traffic-mix confounding, and how you would decide whether the launch is successful.