Design the event instrumentation for Audiobooks at 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 expanded beyond music and podcasts into audiobooks, creating new user journeys around discovery, sampling, purchase or entitlement access, listening progress, recommendations, creator/publisher reporting, and monetization. In this Technical PM interview, you are asked to design the event instrumentation needed to support Audiobooks at scale across Spotify’s consumer apps, backend services, analytics systems, and advertising or monetization workflows.

The focus is not to design the audiobook product itself, but to define how Spotify should reliably capture, structure, validate, and use product events so teams can understand listener behavior, measure audiobook engagement, support personalization, report business outcomes, and enable advertiser or partner use cases where appropriate. The instrumentation must work globally, across platforms, and under real-world constraints such as offline listening, cross-device playback, privacy requirements, experimentation, and high-volume streaming data.

You should assume multiple stakeholders depend on this data: product teams tracking discovery and conversion, engineering teams maintaining playback and entitlement systems, data science teams building recommendations, publishers or creators monitoring consumption, and advertising or monetization teams evaluating campaign or inventory performance. The design should balance completeness, accuracy, latency, cost, governance, and user trust.

The experience should consider:

- Core audiobook user workflows to instrument, including discovery, previewing, starting, pausing, resuming, completing, abandoning, saving, purchasing, borrowing, or using an entitlement.

- Event taxonomy and schema design, including event names, required properties, identifiers, timestamps, content metadata, user/account context, device context, and session linkage.

- Client-side versus server-side instrumentation trade-offs, especially for playback, entitlement checks, offline listening, retries, deduplication, and cross-device continuity.

- Data quality and reliability requirements, including ordering, idempotency, validation, versioning, late-arriving events, monitoring, and backfill strategies.

- Privacy, consent, security, and regional compliance considerations for listener behavior, advertising use cases, publisher reporting, and data minimization.

- APIs, pipelines, and downstream consumers, including analytics dashboards, experimentation platforms, recommendation models, royalty or partner reporting, and ad measurement systems.

- Rollout and observability plans, including phased deployment, instrumentation QA, anomaly detection, alerting, ownership, documentation, and rollback paths.

- Product trade-offs around event granularity, latency, engineering complexity, storage cost, and usefulness for decision-making at Spotify scale.

The goal is to define a scalable instrumentation approach that gives Spotify trustworthy, actionable audiobook data while supporting personalization, monetization, creator ecosystem needs, and operational reliability without over-collecting or creating brittle tracking systems.