I went with a top-down approach: global internet users, smartphone penetration, music streaming adoption, Spotify market share, then tried to figure out what fraction of users would be listening at a random point in the day.
Break the problem into a demand-side estimation: calculate Spotify's global daily active users (DAU), then estimate the fraction of DAU that are actively listening at any given moment by considering average daily listening time and peak concurrency patterns. Use a top-down approach with clear assumptions and sanity checks, and segment by user type (free vs. premium) and geography if helpful.
Pro tip: Anchor your estimate to a known metric like Spotify's reported MAU (~600M) and DAU/MAU ratio (~50%), then apply a concurrency factor derived from average daily listening minutes and peak-hour distribution. This shows you can leverage public data and think in terms of product metrics.
Confirm what 'actively listening' means (e.g., streaming a song, not paused) and whether it includes podcasts or only music. Assume global, all platforms, and include both free and premium users.
Start with Spotify's reported monthly active users (MAU) and apply a typical DAU/MAU ratio for music streaming services (e.g., 50%). Alternatively, estimate from device penetration and usage frequency.
Use known averages: Spotify users listen to about 2 hours per day on average, but this varies by segment (free vs. premium, heavy vs. light). Assume a reasonable average like 90-120 minutes.
Divide average daily listening time by 24 hours to get the fraction of the day an average user is listening. Adjust for peak hours (e.g., listening is concentrated in evenings, so concurrency during peak is higher than average).
Multiply DAU by the concurrency factor to get average concurrent listeners. Then adjust for peak concurrency (e.g., multiply by a peak-to-average ratio of 1.5-2x). Sanity-check against known data (e.g., Spotify's reported peak concurrent streams).
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.