Estimate infrastructure or operational capacity needed for Discover Weekly 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’s Discover Weekly is a personalized playlist experience delivered to premium subscribers globally on a recurring weekly cadence. In this guesstimate question, you are asked to estimate the infrastructure or operational capacity required to generate, store, deliver, and monitor Discover Weekly at global scale.

Frame the problem as a capacity-estimation exercise for a large-scale personalization system. Consider the end-to-end workflow: user eligibility, recommendation generation, playlist assembly, metadata retrieval, storage, push/update timing, client access, streaming demand influenced by the playlist, and ongoing quality monitoring. Your estimate should make clear what part of “infrastructure or operational capacity” you are sizing and what units you are using.

This is not a request to design the recommendation algorithm or propose product improvements. The focus is on scoping the system, making reasonable assumptions, estimating orders of magnitude, and identifying which variables most affect the final capacity requirement.

The experience should consider:

- Scope: which users, markets, devices, and Discover Weekly workflows are included or excluded.

- Unit of capacity: compute jobs, storage, network bandwidth, database reads/writes, playlist records, ML inference volume, monitoring load, or operational staffing.

- Population assumptions: global premium subscriber base, eligible active users, and weekly engagement with Discover Weekly.

- Frequency assumptions: playlist refresh cadence, generation windows, retries, peak versus average access, and listening behavior.

- Data and content assumptions: tracks per playlist, metadata size, recommendation candidates, personalization features, and historical listening data.

- Infrastructure sensitivity: how capacity changes with user growth, engagement, model complexity, cache hit rates, and regional distribution.

- Sanity checks: compare weekly batch load versus daily/hourly load, peak traffic versus baseline streaming traffic, and storage growth over time.

- Constraints: reliability expectations, latency needs, regional availability, privacy considerations, and the importance of avoiding failed or delayed playlist delivery.

The goal is to produce a clear, defensible estimate with transparent assumptions, logical breakdowns, and an understanding of the major drivers of capacity at Spotify scale.