The only email you need to read to become a better product manager

Every week, I spend 50+ hours researching an important product management topic.

I summarize the learning into a practical guide.

I share it in this newsletter.

    No spam. Unsubscribe any time.

    The Most Loved Spotify Features, and How I Would Measure Success
    7 min read

    The Most Loved Spotify Features, and How I Would Measure Success

    The Most Loved Spotify Features, and How I Would Measure Success

    Today's post is slightly different from the usual. I talk about an app that I love using - Spotify.

    In this post, I share a long list of features that make the app easy to use, comprehensive, and superior to its competitors. For each of the top features, I am also sharing a list of product metrics that I would use to measure success if I were the PM.

    Let us start with the top features.

    A large and diverse collection of content.

    I love listening to Hindi, English, and Punjabi music. I also have a few Spanish, Bengali, and German songs saved in my playlists.

    Despite my varied interests, finding favourites or the newest songs in each genre is super simple and very quick.

    I like to find my favourites by merely searching for them. I prefer the playlists on the search screen for discovering new songs. The "Charts" section in the search screen is ideal for discovery, as it highlights the most heard songs across the globe.

    Left: Search screen with multiple options. Right: snapshot of the "Charts" selection

    I also listen to Podcasts - not as often as music though - and Spotify's collection is large enough to find my preferred Podcasts.

    How I would measure success:

    1. Click Through Rate (CTR) for top 10 search results - helps me understand if the users are able to find the content they're searching for.
    2. Number of searches before the first click OR time between first search and the first listen  - helps me understand how long it takes the users to find the first relevant content.
    3. Time spent on the songs found via search results - validate relevance of the search result. This can be used to create a relevance score, which can be used as feedback data in the search algorithms.

    In addition to the above metrics, I will also*:

    1. Measure the impact on overarching product metrics like retention, engagement, 7/14/30 day retention, and CLV. This allows me to ensure that these features are contributing to larger product goals.
    2. Analyse these metrics across various dimensions - paid vs free users, geography, device type, number of devices. Looking at data across multiple dimensions brings out interesting trends that are tough to discover otherwise.

    *This section is common to all success measures mentioned in the article.

    Easy and accurate discovery of content

    Finding the music that I already like and discovering new music that I will like are the two most important aspects of choosing a music app.

    Spotify does exceptionally well on both these parameters. The features that enable accurate discovery are:

    Playlist first experience.

    Spotify has chosen to make its experience playlist focused. The Home, Search, and "Your Library" screens display multiple playlists when you land on them.

    I think the playlist first approach works well, as most users relate to playlists more than browsing lists of unalike songs. I feel playlists humanize the experience. Every time I'm browsing a playlist, I think of the playlist creator as another user who likes the same song as me. The probability of me enjoying other songs that she likes is very high. And that is why I prefer discovering music via playlists and not song lists.

    Search results allowing you to browse playlists.

    How I would measure success:

    1. Playlist CTR - helps me validate if the users are aware of the feature.
    2. Listen-time of each playlist - signals on the relevance of the playlists that users found via search.

    I would measure both metrics for each of the entry points like home page, search page, etc.


    Radio is a feature that lets Spotify continuously serve you with tracks based on any Artist or Playlist you choose. I specifically like the Playlist Radio, as it is the best way to find similar music to your playlist.

    Click on settings on any playlist, and get the option to play Playlist Radio.

    How I would measure success:

    I would use the same metrics as I did for the previous feature.

    Song recommendations for playlists

    I find this feature better than the Playlist Radio because it does the same job, of finding relevant additions to the playlist, in lesser time. The placement is too apt - right after the last song in the playlist. This section displays a list of songs similar to songs in your playlist and will make excellent additions to your playlist. I have used this feature extensively, and cannot recommend it enough.

    List of recommended songs at the end of every playlist.

    How I would measure success:

    1. CTR on the recommended songs - validates feature awareness and signals user's interest in the recommendations
    2. Avg. no. of songs added per user - signals user's liking of the recommendations

    All of the above features are very powerful and accurate at helping you discover songs that Spotify knows, with a high degree of certainty, you will like. In this case, accuracy is an outcome of the multiple recommendation models running in the background.

    Sophia, now an engineer at Spotify, shared in her 2017 article that Spotify uses three different models to power the recommendations:

    1. Collaborative Filtering models, which analyze both your behaviour and others' behaviours.
    2. Natural Language Processing (NLP) models, which analyze text.
    3. Audio models, which analyze the raw audio tracks themselves.

    The best and the most used application of the recommendations are the personalized playlists that Spotify creates for each user.

    My favourites: the yearly "Your 2020 Wrapped" and the weekly "Made for Siddharth Arora" mixes.

    Spotify's playlists for me (on the home page, second fold)
    Users who love Spotify Discover Weekly
    Users who love Spotify Discover Weekly

    Seamless experience across devices

    I currently use many devices: Android phone, iPad, a Mac Air, a Macbook Pro (for work), Samsung Smart TV, PS4, multiple Bluetooth speakers, and an Amazon Fire Stick.

    I use Spotify on each of these devices numerous times a day. I have never felt any friction in switching from one device to another.

    I listen to music offline while commuting; and as long as I have downloaded the desired songs or playlists beforehand, the offline experience is as seamless as the online.

    If you're using multiple devices at the same time, Spotify tells you the device that is active right now. For example, I am listening to Spotify via my Fire Stick, and I have the web player open on the laptop that I am using to write this article. As you can see below - the web player highlights that the Fire Stick is the active device. I can control the music on the Firestick app from my laptop. And then switch back to the Firestick remote when I am done working on the laptop.

    There is no extra work that I am doing to make this multi-device magic happen.

    Spotify Web Player displaying the name of the active device at the bottom of the screen.

    In the rare chance I log in to Spotify from a public laptop/desktop, I use the Web Player. And the web player doesn't disappoint either.

    Before I move on to the next feature, I would like to summarize this section by saying - there hasn't been a single situation when I wasn't able to listen to Spotify exactly the way I wanted to.

    This alone makes me a loyalist, and gives me enough reason to stay away from competitors.

    How I would measure success:

    1. Percentage of users using multiple devices - helps me validate if this is a big enough user segment using more than one device.
    2. Distribution of the devices used - helps me understand the most used devices, so I can optimise for it.

    Ability to Crossfade songs for up to 12 seconds.

    Crossfade, for me, is one of the delight-the-user features. Crossfade makes Spotify play the next song before the current song is over. You can switch it on, select the overlap duration, and switch it off as per your convenience. I dislike the few seconds of gap between two songs, and this feature lets me get rid of that gap.

    Crossfade option is available in the main Settings section of the app

    How I would measure success:

    1. Percentage of users who use Crossfade feature - signals feature awareness and user interest
    2. Percentage of users who have switched it off after switching it on - helps understand the number of users who know about the feature, tried it, but then decided not to use it. This signals interest, and dissatisfaction with the feature.

    If you still don't believe in the awesomeness of Spotify, this additional list of features, in no specific order, might change your mind.

    1. Keyboard shortcuts: Spotify makes it easy to navigate the app without using the mouse.
    2. Recover deleted playlists: Accidentally deleted one of your playlists? Spotify has you covered. It backs up your deleted lists so that you can recover them later.
    3. Information about storage space: if you're running out of space on your device, you can free up space by clearing the cache or changing the downloaded files' location.
    4. Spotify car view reduces distractions: if you're using Spotify on your phone while driving, the app enters into a simplified view, which increases the size of all controls and removes unnecessary CTAs from the screen.
    5. Data saver mode: If Data Saver is on, Spotify caches the recently played songs on your device. Playing cached songs can reduce data usage by as much as 75%.
    6. Collaborative playlist: You can collaborate with friends to create playlists together. Friends can add or remove songs, as long as they have access.
    7. Advanced search: Spotify search modifiers allow you to make your search extra-specific by adding parameters like artist, album, genre, year, etc. in the search keyword.

    Like what you read? Subscribe to our newsletter and get one practical guide every week.