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 activation for Feed among job seekers. Feed activation should be treated as a meaningful early-use milestone: a member reaches the Feed, sees relevant professional content, and takes an action that indicates the Feed is useful enough to engage with, such as following, reacting, commenting, saving, clicking, or otherwise interacting with content.

Your task is to diagnose the drop as an RCA exercise. Focus on the job-seeker experience: members may be visiting LinkedIn to search for roles, update their profiles, respond to recruiters, learn new skills, or monitor professional updates. The Feed must balance relevance, trust, freshness, and professional intent while supporting LinkedIn’s broader ecosystem of identity, hiring, learning, and network effects.

Assume the anomaly is recent and material enough to require cross-functional investigation. You should frame how you would validate the metric movement, isolate affected cohorts and surfaces, form hypotheses, identify the data or experiments needed to confirm them, and recommend how the team should contain impact while the root cause is being investigated.

The experience should consider:

- How “Feed activation” is defined, including numerator, denominator, event timing, eligibility, and whether the definition changed.

- Whether the 20% drop is global or concentrated by geography, platform, app version, acquisition source, logged-in state, tenure, job-seeker intent, or traffic source.

- Key job-seeker workflows leading into Feed, such as home tab visits, job-search sessions, notification clicks, profile updates, recruiter messages, and content recommendations.

- Instrumentation checks, including event logging, tracking gaps, pipeline delays, bot or spam filtering, attribution changes, and dashboard correctness.

- Product and ranking hypotheses, such as relevance changes, content quality issues, notification changes, onboarding friction, latency, broken UI states, or reduced professional trust.

- External and ecosystem factors, including seasonality in hiring, macro job-market shifts, competitor activity, content supply changes, or policy/moderation changes.

- Evidence needed to distinguish correlation from causation, including time-series analysis, experiment readouts, funnel breakdowns, user feedback, and operational logs.

- Mitigation and prevention plans, including short-term containment, monitoring, owner alignment, communication, and safeguards against recurrence.

The goal is to demonstrate a structured RCA approach that can separate a real user-experience decline from measurement noise, identify the most likely root cause, and guide LinkedIn toward an evidence-based recovery plan for Feed activation among job seekers.