Spotify

Problem:
People want Spotify to function more easily when being used during their favorite activities
Solution:
The new Activity Mode mobile feature allows for automated play when connected to your favorite listening devices, a pared down interface, and customizable function and playlist assignment
My Role
UX Research
Prototyping
Visual Design
Duration
2 Months
Team
Self
Design Mentor
Tools
Figma
Illustrator
Marvel
Miro
AirTable
Photoshop
Background
Spotify was looking for a way to improve engagement and retention in the app by expanding on their social capabilities. My initial research was focused in that direction, but based on what I learned, I was led toward a different mobile feature that would help the Users reach their goals, while making Spotify an even more seamless companion to their daily lives.
Research

Tools Used: Market Research, Competitive Analysis, Provisional Personas, User Interviews

Before beginning to design, we must first learn what we can from existing sources of information, as well as directly from potential users of our product. This will ensure that our resources are being used to their maximum.
Before deciding where to aim my resources, I outlined our Research Goals, Methodologies, Participants, Timeline, Assumptions, and the questions we sought to answer.

My Research Goals were as follows:

- Learn about current trends in the streaming music space
- Discover what social implementations currently exist among competitors
- Understand current user demographics, habits, needs, and motivations

My Research Questions were:

- Who are the players in the current marketplace, direct/indirect?
- What are the priorities of Spotify Users?
- How does Spotify currently meet the needs of its Users?
- Are there gaps in the current satisfaction of Spotify Users?
- Are Spotify Users interested in further connecting with others?
- How do Spotify Users discover new music/artists?

Once I understood what the goals were for my research, I moved forward with collecting the primary and secondary research that would be needed to answer them.

Secondary Research: Market Research
Before generating my own data, I sought to learn more about the streaming music space, current trends, the competitors to their product, who the target client is, and what information exists on their expectations, priorities, and preferences.
Market trends and statistics:
  • The current percentage of subscribers out of total MAUs (Monthly Active Users) stands at 45% – slightly lower than the 46% it has been hovering around since mid-2018
  • Spotify subscribers as a percentage of MAU in Q4 2015 stood at 31%, in Q4 2016 at 39%, by Q4 2018 at 46%
  • Spotify claims a 35% market share, leading Apple Music (19%) and Amazon Music (15%). Tencent Music logs an impressive 11%, despite its heavy domestic focus (where it is split over a suite of four apps), while YouTube clocks 6%
  • Spotify quarterly revenue in Q1 2020 stood at €1.85 billion ($2 billion), €1.7 billion ($1.84 billion) of which came from Spotify Premium subscribers. (92% of Spotify's revenue came from Premium subscribers)
  • Gross profit in the same quarter was €472 million ($511 million), with operating loss at €17 million ($18 million)
  • Spotify annual revenue in 2019 at €6.7 billion ($7.3 billion), with gross profit of €1.7 billion ($1.8 billion)
  • There were 286 million monthly active users of Spotify according to the company’s Q1 2020 report. Of these, 130 million were paid Spotify Premium subscribers
  • This is up from 271 million monthly active Spotify users and 124 million Spotify Premium subscribers in Q4 2019. Year-on-year, it’s a 69 million increase in users (32%), and 30 million increase in subscribers (30%) – a growth rate that has been fairly consistent in recent years
    Demographics + User Statistics:
    • Average users listen to 41 unique artists per week
    • Spotify-created playlists attract roughly three-quarters of all the followers of the top 1,000 tracks on the platform
    • 55% users have registered on Spotify via their Facebook accounts
    • 48% of Spotify Users are from the U.S.
    • As of March 2018, Spotify’s user base was dominated by Millennials, with 29 percent of its users aged 25 to 34 and 26 percent aged between 18 and 24 years old. According to a 2018 survey, Spotify reaches almost half of 16 to 24 year olds in the United States each week
    • Spotify has the most youthful User base with over half of users aged 34 or under, compared with 40% of Apple and 39% of Pandora users
    • 56% of U.S. Spotify Users are male, while 56% of Apple Music Users are female
    History with Social Features:
    Secondary Research: Competitive Analysis
    I sought to learn from our competition in regards to what they are doing well, and where there’s opportunities to move the segment forward. I performed some high level initial research on some key direct and indirect competitors, and continued to benchmark against them throughout the design process.
    Direct Competitors
    Strengths
    • 68+ Million paid subscribers
    • Leads Spotify in paid US subscribers
    • Largest library of music 60 Million songs
    • Integrates best with all Apple devices/services
    • Focus on exclusives with artists to feature music earlier on their platform
    • Beats 1 live “radio station” channel
    Weaknesses
    • Not as seamless with non Apple devices/services
    • No free (ad supported) version
    • Not poweful enough in the marketplace to keep exclusives very long
    Strengths
    • 16+ Million paid subscribers
    • Large library of music 40 Million songs
    • Integrates with Google Home, and Android devices
    Weaknesses
    • Being phased out by YouTube Music
    • Not as seamless with non Google devices/services
    Strengths
    • 20 Million paid subscribers
    • Large library of music 30 Million songs
    • Integrates with Google Home, and has become the default music app on Android devices
    • Will have Google Play users crossing over
    • YouTube has access to 2 billion users, perhaps leading to advantage in emerging markets
    • Access to music videos in addition to audio
    Weaknesses
    • The marketing for YouTube Music/Google Play hasn’t been clear, and is cloudy with the merge
    • Early days for their generated playlists, but has access to Google cloud computing platform (same that Spotify uses) to crunch user data to generate them.
    Strengths
    • 55 million users, but unclear how many are paid
    • Large library of music 50 Million songs
    • Integrates with Alexa
    • Less expensive @ $7.99 than others for Prime members
    • Offers “HD” version for $12.99 lossless FLAC audio (For less than Tidal or Deezer)
    • Karaoke like lyric scrolling
    Weaknesses
    • No frills UI experience on desktop
    • Not yet focused on great music algorithms
    • No Podcasts
    Other Competitors
    Weaknesses
    • Not as popular as other alternatives, so your friends likely use a different service
    • Algorithims for discovering music tastes not as developed
    Strengths
    • 7 Million paid subscribers, 14 million overall
    • Large 53 million song library (More than Spotify)
    • Focus on audio quality with 28 million FLAC uncompressed audiophile quality songs
    • Similarly priced and featured to Spotify
    Strengths
    • Large library of music 60 Million songs (More than Spotify)
    • Owned primarily by artists, and is seen to be the most beneficial to artists
    • Has an option for “HiFi” subscription which offers lossless audiophile quality audio
    • Said to have one of the best looking interfaces
    Weaknesses
    • 4 million paid subscribers, but number hasn’t been made public in years
    • Hifi subscription is expensive
    • Don’t offer a free, ad supported version
    Strengths
    • 60 Million users, but used to have 83 million. Steady decrease qtr by qtr
    • Large library of music 40 Million songs
    • Best “radio” focused algorithm
    • Only $5 for ad free “radio” style service
    Weaknesses
    • Has only 6 million paying subscribers
    • Declining usership
    Strengths
    • New service, but will have access to all Sonos device customers
    • Will be default service installed on Sonos
    • “Radio” focused content, partnering with iHeartRadio, Tunein, and soon Radio.com. Will also have their own live station.
    • Artist curated stations
    Weaknesses
    • Only available to Sonos owners, so will be limited in size
    • New in a competitive market
    Provisional Personas
    Based on the studies that were referenced during the market research, I put together two provisional personas. These are the most likely users for our service, and will be used to narrow down subjects for the next step: User Interviews
    Goals:
    • To discover new artists and podcasts
    • To conveniently play music she enjoys without too much time invested
    • To be entertained on her commute
    Pains:
    • No time to explore for new music/podcasts
    • Feels overwhelmed with options
    • Not alway satisfied with the offered playlists she finds
    Goals:
    • To discover new music early
    • To connect with other enthusiasts and their favorite artists
    • To have recognition for their curated playlists
    • Listen with hi fidelity quality
    Pains:
    • Difficult to find tastemakers to follow
    • Not easy to connect with enthusiasts he’s not already connected with on another social platform
    • Wants the same satisfaction and dialogue on his latest playlist that he gets when he receives “likes” or comments on Instagram
    Primary Research: Heuristic Evaluation
    In order to greater understand Spotify's current interface, I completed a heuristic evaluation. I spent time going over the interface to see where issues might exist, and where Spotify either adhered or strayed from norms/best practices.

    Overall, I had a few main impressions:

    The evaluation in its entirety may be downloaded below:
    DOWNLOAD PNG
    Primary Research: User Interviews
    It's important to learn directly from users of our product to understand their perspectives, pain points, and propensities in order to make better design decisions for their needs. I recruited a group of current Spotify users and asked a series of open ended questions to better understand how they are currently using Spotify, how/if they share music currently, their history with other streaming music platforms, and how Spotify could better suit their needs while providing Spotify with the greater engagement and loyalty they seek.

    The statistics for the test participants are as follows:

    -5 interviews were conducted 
    -3 male and 2 female 
    -Ages ranged between 29-47
    -3 married, 2 single
    -3 self employed

    Research Synthesis

    Tools Used: Empathy Map, User Persona

    Empathy Map
    The tool I chose to use to synthesize the data I generated is an Empathy Map. This helped me to organize the data into information that we can use to discover common themes, patterns, and similarities that appear most often and define the problems we're looking to solve. Below you'll find an image of the results:

    The Empathy Map revealed a number of points stemming from our raw interview data. Outlined below you'll find the Insights, Goals, Pains, and Needs that were made clear during the research phase:
    Insights:
    • Most Spotify Users (4/5) are using Facebook and Instagram
    • Most Users use the Spotify algorithm “radio” feature to discover new music
    • When Users share music with friends, most will text spotify links
    • The heaviest users reported that they often share music with a group of friends that have similar taste
    • Users aren’t aware of new socially focused features. (Spotify Codes and Group Session)
    • All Users reported that the times they are listening to Spotify, they are doing another activity. The top activities were, #1 exercising (5/5), #2 (4/5) driving, #3 (3/5) working
    • Most Users (4/5) mentioned the automated algorithmic recommendations as their favorite Spotify features
    • Spotify Users haven’t found value in following or being followed by individuals on Spotify
    • No Users stated positive feelings towards having followers having access to what they were listening to
    • 3/5 Users reported visiting either Soundcloud or BandCamp to consume music from smaller artists not available on Spotify
    Pains:
    • Sharing music can be embarrassing if done automatically
    • New Spotify features aren’t easy to find
    • Following or having followers isn’t currently adding to their experience
    Needs:
    • Discover music that fuels their activities
    • Have a more valuable interaction using social features
    • Have personalized music recommendations from trusted sources
    Goals
    • I want to easily find new music to listen to that I know I’ll like
    • I like to share music with my friends who have similar taste
    • I want to have music that motivates my workouts and workday
    User Persona
    Meet Barry Carver. After synthesizing my research, I generated a fictional representation of a target user guided by the findings. Putting a face and name to our user reminds us who we’re designing for.

    Define + Ideate

    Tools Used: POV Statements + HMW Questions,  Solo Brainstorming, Group Brainstorming

    Defining the Problem
    To further clarify the problems that we will design to solve, I translated the insights and needs from my empathy map and persona into POV "Point Of View" Statements, and HMW "How Might We" questions.

    Insights #1
    Spotify Users haven’t found value in following or being followed by individuals on Spotify
    Needs #1
    Have a more valuable interaction using social features
    POV #1
    Barry needs to be able have a more valuable interaction using social features because Spotify Users haven’t found value in following or being followed by individuals on Spotify
    HMW #1
    How might we provide Barry with more value from his follower relationships on Spotify?
    Insights #2
    The heaviest users reported that they often share music with a group of friends that have similar taste
    Needs #2
    Have personalized music recommend-ations from trusted sources
    POV #2
    Barry needs to have personal-ized music recommend-ations from trusted sources because the heaviest users reported that they often share music with a group of friends that have similar taste
    HMW #2
    How might we help Barry share music recommend-ations with people who share his taste?
    Insights #3
    All Users reported that the times they are listening to Spotify, they are doing another activity
    Needs #3
    Discover music that fuels their activities
    POV #3
    Barry needs to discover music that fuels his activities because all Users reported that the times they are listening to Spotify, they are doing another activity
    HMW #3
    How might we give Barry the ability to discover music that fuels his activities?
    Brainstorming - Solo
    Once I was armed with my HMW statements, I needed to brainstorm ideas on how to solve them. I chose to do mind maps for each of the 3 HMW statements to generate as many solutions as I could within a given timeframe.

    Brainstorming - Group
    Getting more minds together that aren't as immersed in the subject can often lead to some unexpected and useful ideas. I recruited a few people to help solve for those same 3 HMW statements. A time lapse of the session can be viewed below:

    The final results from the group brainstorm session were sorted into themes in order to group like ideas. This can be viewed below:

    Strategy

    Tools Used: Project Goals

    Project Goals
    Assembling the business goals from the design brief, and the user goals from my research, aids in determining overlap and competing priorities in developing site features. The easiest decisions to make are where both business goals and users goals overlap. Those move to the forefront when moving to the next step, which is planning a Product Roadmap.

    Information Architecture

    Tools Used: Product Roadmap, Sitemap

    Product Roadmap
    With the shared goals in mind that I discovered in the previous step, I looked back to the brainstorming step in the process to see which ideas addressed those priorities. The features that align with the shared goals move forward first, followed by other features that will separately address User and Business needs.

    Features are further organized by which HMW question they address, how difficult they would be to implement, and my confidence in the feature directly addressing a stated need.

    It was at this point that the project took a bit of a turn. I had started with the intention of creating a social feature for Spotify. Throughout the process, it appeared to me that the people I had interviewed and gathered for a brainstorm didn't necessarily seem to want a more social version of Spotify. They weren't aware of or using the new social features within Spotify (Group Session and Spotify Codes). Many of the social ideas that were generated during the group brainstorm were actually possible within Spotify already. Also, most of the participants wanted to see what other people were listening to, but didn't like the idea of people seeing what they listen to.

    Since it appeared that the social feature wasn't what would most aid the Users I worked with, the feature that made it to the head of the product roadmap was "Auto Detect Activities", later coined "Activity Modes". Most of the Users stated most of their listening takes place while doing another activity. Activity Modes allows Users to get to the business of listening to specific music paired to their activities in a more automated, minimal fashion.

    App Map
    Moving forward with Activity Modes, I laid out the App Map that outlines the structural changes that would take place within Spotify to accommodate the new feature.

    Interaction Design

    Tools Used: Task Flow, User Flow, UI Requirements, Low Fidelity Wireframes, High Fidelity Wireframes

    Task Flow
    Informed by the Sitemap, I built a Task Flow in order to test some specific tasks that a User is likely to perform.

    User Flow
    I then created a User Flow to ensure that all of the user decision points are thought through, that the flow is an efficient as possible, and that all of the necessary pages and features are accounted for to accomplish specific scenarios.

    UI Requirements
    Referring to the documents just created, I mapped out all of the pages and high level requirements that I will need to design in order to complete the set tasks/scenarios. You can access that file at the button below:

    DOWNLOAD .PNG
    Low Fidelity Wireframes
    One of the challenges with this project is that I need to work the new feature into the current Spotify design using as many existing elements and patterns as possible. Advancing from User Flow and UI requirements, I know what is needed to implement the feature, so I can proceed with some initial sketches to understand how the feature fits into the existing design before spending too much time rendering the hi fidelity versions. Changes take minutes instead of hours at this stage.
    High Fidelity Wireframes
    Once the design was verified through the rough sketches I took the next step to design the high fidelity wireframes. I recreated the Spotify interface completely in vector (except for album artwork), created the Activity Modes icons, and then executed the changes to incorporate Activity Modes.
    Prototyping
    It was at this point I needed to learn whether or not the feature would work as intended, so I needed to build a prototype to as closely as possible approximate the experience of using Activity Modes. The prototype can be interacted with below:

    Usability Testing

    Tools Used: Testing Plan and Script, Affinity Map

    Usability Test
    Before testing I laid out the objectives, methodologies, script, and scenarios for the test.

    I conducted a think aloud, moderated, usability test. All of tests were conducted with the Marvel user testing features. They took place in person when possible and remote via Zoom when necessary. They went through the following three scenarios:

    1. You just put your AirPods in and want to go for a run. You want to try out Activity Mode: Run for the first time.
    2. You just got into a rental car that has a phone connector, but no bluetooth. You want to activate Activity Mode: Drive manually.
    3. You want to add the "Phantogram Radio" playlist to your #1 spot on your Activity Mode: Drive list.

    I completed my user testing with 5 individuals, aged 41-47, 4 females and 1 male.

    I've created a timelapse of the video and screen taps recorded through Marvel that can be viewed below:
    Following testing I completed a detailed document with user feedback and actions recorded. It can be downloaded below:
    DOWNLOAD.PDF
    Affinity Map
    In order to discover the most common issues users might have had during the testing I created an Affinity Map. The patterns created pointed out the items that need attention.
    User Patterns:
    1. Using the Activity Modes Run+ Icon wasn't universally recognized. This should be moved to the ellipsis menu, and either all 4 icons should be represented, or there should be text accompanying the icon to clarify its purpose.
    2. There should be a link to manage Activity Modes Playlists in the Settings -Activity Modes menu.
    Recommendations:
    1. During Task #3, 2/5 subjects attempted to use the ellipsis button to add the playlist to their Activity Modes. One was unable to complete due to this.
    2. During Task #3, 4/5 test subjects visited “Settings” first to add a playlist to their Activity Modes.
    Priority Revisions
    I made the recommended changes to the Activity Modes features, changes to the created Icons to make them more legible at small sizes, along with some small detail fixes. This iteration of the screen designs can also be viewed here:
    High Fidelity Prototype - Revised
    The most recent iteration of the prototype is available for interaction below:
    User Interface Design

    Tools Used: Mood Board, Style Tile, Figma Prototypes, High-Fidelity Wireframes, High Fidelity Responsive, UI Kit

    Mood Board
    Since Spotify already has a defined brand and look, this was more of an exercise in the process. I used Pinterest to classify some of the categories that make up Spotify's look and feel, and sought out images that suited them.

    Branding: Style Tile
    Using Spotify's own branding guidelines for their logo and some reverse engineering to determine the fonts, I was able to put together a style tile for Spotify.
    UI Kit
    I created a UI Kit for the items altered or created during the design of Activity Modes so any future reference/development will have a consistent library of components to work from.
    Reflection + Next Steps
    Reflection
    This project was a valuable lesson in keeping an open mind and allowing flexibility when approaching adding features that better serve the needs of a User rather than allowing the original project brief lead the way. Many of the questions I asked and directions I headed in the beginning were focused on the brief of adding a social feature to Spotify. I wasn't feeling a lot of enthusiasm from the Users that I interviewed for the social features that existed already, and didn't get a strong sense that they'd be interested in new ones either without a larger rethink of Spotify in general. I did notice the general trend earlier on in the process that what they were most enthusiastic about were the automated features that Spotify offers, and that they typically were using Spotify while doing another activity. It became clear to me that looking towards those User preferences was going to be the key to creating a feature that would be used and ultimately led to Activity Modes.
    Next Steps
    - User Testing + Iteration: I'd like to see another round of user testing in order to confirm the revisions were a solution to the issues I encountered during the last round of testing.

    - Design Handoff: Once the user testing and iteration cycle is satisfactorily completed, design should be handed off to development with our documents and communication successfully bridging that gap.