Diagnose a 20 percent drop in activation for Feed

Problem Statement Description

Product context: LinkedIn is Microsoft's professional network; its products include profiles, feed, jobs, recruiting, LinkedIn Learning, sales tools, messaging, and ads.

LinkedIn has observed a 20% drop in Feed activation among job seekers. In this context, Feed activation refers to a new or returning job-seeker segment reaching a meaningful first-use milestone in the Feed experience, such as viewing relevant professional updates, engaging with posts, following companies, or discovering job-related content after entering the Feed.

Your task is to diagnose the drop as a product RCA. Focus on how job seekers arrive at Feed, what they expect from it, and where the experience may have changed across onboarding, notifications, search, Jobs surfaces, profile activity, content ranking, or Feed engagement flows.

The investigation should separate whether this is a real user-behavior decline, a measurement or instrumentation issue, a segment-specific shift, or the result of recent product, ranking, traffic, seasonality, or ecosystem changes. The answer should show how you would structure the analysis, prioritize hypotheses, and move from signal to evidence.

The experience should consider:

- How activation is defined, including numerator, denominator, time window, and whether the metric changed recently

- Which job-seeker cohorts are affected, such as new members, recently active job applicants, laid-off users, students, premium users, or specific geographies

- Entry points into Feed, including home tab, notifications, email, Jobs, profile views, search, and mobile push

- Funnel steps before activation, such as landing, content load, relevance, scroll depth, first engagement, follow actions, or job-related content discovery

- Instrumentation checks, including event logging, client versions, platform differences, tracking delays, bot/spam filtering, and data pipeline health

- Product and ranking changes that could affect professional relevance, trust, content freshness, creator distribution, or job-related recommendations

- External or contextual factors such as hiring seasonality, macro job-market changes, competitor behavior, notification deliverability, or content supply shifts

- Mitigation and prevention, including how to size impact, identify owners, monitor recovery, and avoid repeat incidents

The goal is to demonstrate a clear RCA approach: frame the anomaly, validate the data, segment the impact, generate plausible hypotheses, identify the evidence needed to confirm or reject them, and propose how the team should respond without jumping prematurely to a product fix.