Root cause a sudden decline in retention among podcast fans using Podcasts at global scale

Problem Statement Description

Product context: Spotify is an audio streaming company; its products include music, podcasts, audiobooks, playlists, recommendations, creator tools, subscriptions, and ads.

Spotify has observed a sudden decline in retention among podcast fans using Podcasts at global scale. You are the product manager asked to lead the root-cause investigation across product, data, engineering, content, recommendations, ads, and regional teams.

This is an RCA interview focused on how you structure ambiguity, not on naming one predetermined cause. Podcast fans may include users who regularly listen to podcast episodes, follow shows, receive podcast recommendations, download episodes, or use Spotify primarily for spoken-word content. The decline may vary by geography, platform, acquisition channel, subscription status, content category, creator segment, or listening behavior.

Your task is to frame the anomaly, identify the most useful cuts of data, validate whether the issue is real versus instrumentation-driven, generate hypotheses, prioritize evidence collection, and propose how Spotify should mitigate user impact while preventing recurrence.

The experience should consider:

- How “retention among podcast fans” should be defined, including time window, denominator, user eligibility, and whether retention means app retention, podcast listening retention, show-follow retention, or repeat episode consumption.

- How to segment the decline across regions, platforms, app versions, free versus premium users, new versus existing podcast listeners, heavy versus casual podcast fans, and content genres.

- Instrumentation and data-quality checks, including event logging changes, identity/session issues, podcast playback events, downloads, follows, recommendations, notifications, and attribution pipelines.

- Product-surface hypotheses across Home, Search, Library, recommendations, podcast pages, episode playback, downloads, queueing, notifications, and cross-promotion with music or audiobooks.

- Ecosystem factors such as content availability, creator publishing cadence, exclusivity changes, licensing, ad load, monetization experiments, and competitor or seasonal shifts.

- Evidence needed to distinguish correlation from causation, including cohort trends, funnel changes, experiment exposure, release timelines, outage logs, and customer-support signals.

- Mitigation options, escalation paths, and short-term monitoring without prematurely assuming the root cause.

- Prevention mechanisms such as alerting, metric ownership, release gates, experimentation guardrails, and post-incident learnings.

The goal is to demonstrate a rigorous, product-minded RCA approach that can operate at Spotify’s global scale, protect the podcast listener experience, and help the organization make confident decisions under uncertainty.