What to expect, what they’re really testing, and what a strong answer looks like — scored.
Audio content discovery, personalization, creator tools, and podcast growth. Spotify PMs must think about the tension between user experience (better discovery) and content supply (keeping artists and podcasters happy), and reason about how recommendation systems interact with creator incentives.
The question below was asked by Spotify interviewers. The answer is graded on the five dimensions real PM interviewers use: structure, specificity, reasoning, decision quality, and delivery.
“How would you improve Spotify's Discover Weekly for users who only listen to one genre?”
Discover Weekly's core assumption is that users want to discover music similar to but different from what they already listen to. For single-genre listeners, this creates a recommendation loop — the algorithm finds artists in the same genre, reinforcing the existing behavior rather than expanding taste.
The question is whether we want to break the loop. Some single-genre listeners are perfectly satisfied — a jazz purist who only wants more jazz. Forcing pop on them is a bad product decision. Others are single-genre by default, not by preference — they've only ever been exposed to one genre and don't know what else they'd like.
I'd focus on the second group: people whose listening history suggests habit, not intention. Signals of habitual listening: same 3 artists in heavy rotation, low skip rates (they're not curating), no manual playlist creation. This user is a good candidate for gentle genre expansion.
Proposed improvement: introduce a 'Branch' row in Discover Weekly — two songs that are musically adjacent but in a different genre. Adjacent means: same tempo range, same emotional valence, similar production style. Example: a hip-hop listener who gravitates toward lo-fi beats might be surfaced lo-fi jazz. Not jazz in general — specifically the lo-fi jazz that shares sonic DNA with their existing listening.
Success metric: branch song save rate (user adds it to a playlist or liked songs). Secondary: whether users who save a branch song go on to listen to more music in the new genre within 30 days — a leading indicator of taste expansion.
Identifies two distinct user types, focuses on the one worth improving, defines the behavioral signal to identify them.
The lo-fi hip-hop to lo-fi jazz example is instantly concrete and demonstrates real music domain knowledge.
The 'habit vs. intention' distinction is a genuinely insightful reframe that drives the intervention design.
Commits to a specific feature design ('Branch' row) with a clear success metric.
Well-paced; the example does a lot of work efficiently.
The 'habit vs. intention' framing is the answer's best move — it avoids the mistake of trying to convert every single-genre listener, which would be a bad product decision. The 'musically adjacent' constraint on the Branch recommendation is the right design instinct. The one gap: the answer doesn't acknowledge how this competes with Spotify's existing radio feature, which already does adjacent discovery. An interviewer will ask.
Add one sentence differentiating 'Branch' from Spotify Radio — for example, that Branch is proactive (pushed into Discover Weekly) rather than reactive (user-initiated), which addresses a different discovery gap.
Paste any Spotify PM interview question and your answer. Get scored on the same five dimensions — instantly, free, no signup.
Grade my answer free →First grade is free. No account needed.