Spotify: Improve & Measure
In this article, I walk you through my thoughts on how to answer the Improve Spotify and Measure Success for Spotify case prompts.
Spotify is one of the most common products used by Product Manager interviewers. Why?
Commonly Used Product
They are always innovating
Marketplace Dynamics
I am preparing for a company that is know to ask improvement and tradeoff questions about Spotify. Think:
How would you improve Spotify?
How would you measure success for Spotify? (as CPO, head of consumer or head of creator)
What do you think is on Spotify’s 5 or 10 year roadmap?
Sample prompts that have been documented for DoorDash :
What do you like about Spotify? How would you improve it?
What problems does [insert your favorite music app] solve for?
How would you improve Spotify?
How would you increase engagement for Spotify?
How would you improve Spotify?
Strategic Overview
Spotify made its name with music discovery.
Leadership has doubled down on all things audio, adding podcasts and audio books in recent years.
I believe they overtook Apple on podcasts.
Most people don’t drastically change their music tastes after their early 20s.
Spotify is trying to help listeners and creators be happy with what they listen to.
Mission: to unlock the potential of human creativity—by giving a million creative artists the opportunity to live off their art and billions of fans the opportunity to enjoy and be inspired by it.
Goal: Revolutionize all of audio and become the world's number one global audio platform.
As such, Spotify is a Marketplace with two core stakeholders: listeners and creators (musicians, podcasters, writers, etc.)
Users:
The biggest thing shaping listeners is their listening style in terms of frequency.
High Variety - always looking for what is hot
Comfort + Variety - know what I like, get into ruts, but pleasantly surprised by new things
Comfort - know what I like and stick to it
This is stronger for music than podcasts, but it applies to some degree to all audio entertainment.
I would guess that the largest population is actually the comfort listener but given the current push for discovery, I think the Comfort + Variety listeners would be the best target for change. Anything that works for them would work for the high-variety personality.
Spotify is constantly delighting the user with new ways to discover either flat-out or subtlety. The most recent feature releases focus on increasing discovery. Therefore, what I want to do is put another filter on my user and look at the situational impact. Discovery is easier when you are at home and can take a moment, but most of the time we listen to music on the go. And I think for this mildly-interested-in-discovery user sees the greatest pain in discovery on the go. Think walking or driving. Roadtrips or Long-Walks.
Note: Another way of segmenting could have been to say, Spotify is known for discovery. When we think about discovery, I think about three situations:
Sitting down to plan a playlist (unmet need: M; pain level: H)
Listening on the go (unmet need: H; pain level H)
Hearing something randomly (unmet need: L; pain level: L)
When I prioritize these situations, I think about unmet need and level of pain. We could also use the terms frequency and severity but I tend to focus pain points on frequency and severity.
Pain Points:
Now thinking more specifically about this user’s pain points with discovery on the go.
Don’t have the ability to search (walking or driving, weak signal or expensive downloads) (freq: H; severity: H)
Hear something I like while phone is in pocket, I can’t save it easily to find it again. (freq: M; severity: H)
Mood changes: got bad news, need to calm down. (freq: M; severity: M)
Solutions:
Chat with my DJ
Trip Planner
Health & Safety
Chat with my DJ - The current Spotify DJ is just a bad version of Pandora’s Thumbprint radio. Add an audio version of ChatGPT to Spotify and describe the kind of mood and music you are looking for. Add a “request for discovery” feature.)
Trip Planner - When traveling (think subways, airplanes, and backroads), you can’t discover or explore because you can’t stream. Imagine I could tell Spotify (using API with permission or email a travel itinerary to Spotify), I am planning a trip and download some new X for my trip in case I get bored. There could be privacy issues, but for those who want to opt in, it could increase discovery when they are a captive audience.
Health & Safety - When traveling late at night (there are a lot of people if you read the road signs in Utah), staying awake is often a different playlist than normal. Imagine saying, I need something upbeat and happy. Or I want to chat with someone. Can you give me an AI Chat convo around X topic to keep me awake? (Not sure if Chat GPT is there yet, but we can’t be that far off.
Prioritization Framework:
Impact and Effort
DJ (Impact: H; Effort: H) - Trends, expectations and competition
Trip Planner (Impact M: Effort: H - privacy issues)
Health & Safety (Impact M - not a common use case; Effort H)
Metrics for Success
Mission: to unlock the potential of human creativity—by giving a million creative artists the opportunity to live off their art and billions of fans the opportunity to enjoy and be inspired by it.
Goal: Revolutionize all of audio and become the world's number one global audio platform.
Users (or Key Stakeholders):
Listeners
Artists/Creators
Other Stakeholders:
Advertisers
Publishers
Goals: Assuming the current state of the product
Business: Increase discovery across all channels (music, podcasts, books)
Listener: Make discovery of new content easier and painless
Creator: Be heard more often by more people to increase revenue
Metrics
# of unique audio tracks listened to (50% or more) per user per week
filter for unique recommendations
filtered by recommendation type
think: mixed in playlist, playlist packages, DJ, searches, voice queried
week because not everyone listens daily; many/most do, but not guaranteed, and podcasts might have a different cadence than music
# of unique tracks per listener per week
# of unique listens per creator per week
# of new uploads per creator per week
# of recommendations listened to per user per week
Note: I would hold this up as important, given all the search and discovery focus
Other metrics
Revenue
$ from ads & subscriptions
# of freemium upgrades per month
$ paid to publishers per month
Avg time on app listening per week
Avg time on app searching per week
% of recommendations acted on per user per week
Photo by Alexander Shatov on Unsplash
