My apologies in advance for my design skills — I’m first and foremost a Product Manager @ Skookum. Design is a hobby muscle I’m slowly flexing.
As Apple and Spotify compete for my attention in the music distribution space, it’s clear that their areas of focus are the visual components of an interface and the content catalogue that can be packaged along with it. None of the big players are truly building a service around customization, personalization and empathy.
That all changed last month, as Spotify took ground by releasing their ‘Discover Weekly’ feature tucked away within the ‘browse’ menu item. Discover Weekly, for those of you that don’t know, is a weekly recommended playlist generated by one of Spotify’s spiffy algorithms.
Spotify’s newly released Discover Weekly feature (with my profile picture on the cover!)
Say what you want about their interface or their spacial sense — this tiny feature blew me away for its simplicity and the empathy it deployed with relative ease. Before this feature, I had never come across a service willing to curate a weekly playlist for me based on my music taste. Sure, Pandora had smart radio stations…but Spotify went along and put my picture on the cover of something!!! I FEEL SO IMPORTANT.
After the fuzzy feelings from an inanimate application subsided, this got me thinking…What was it about this feature that had me be so excited? What is it about the music discovery process that has me continue to find new music and (ultimately) share it with my friends? Is it possible that it’s the recurring social component that drives my active engagement?
Designing a no-ui, personal music assistant
Invisible or no-UI interfaces are slowly coming into vogue and have been on my mind a lot lately. I thought I would use this opportunity to create a product around a virtual assistant that mimics the interactions that music sharing engenders between two friends.
Some questions I had while developing this concept:
- How do friends interact with each other as they share music?
- How do product managers and UX designers build, design and conceptualize product features for a Siri or primitive Her type interface? Perhaps we have to start thinking of product definition in terms of the creation of behaviors and personalities?
- Could this interaction be leveraged to develop a new kind of interface that increases engagement, likability and empathy?
Mapping out the product
To start, I created a journey map for a conceptual app called TwoFriends, an AI product that interacts with users as two friends would as they share new music with one another.
While exploring each step of the music sharing/discovery process, I attempted to answer the following questions:
Discovery Ritual How do I find new music? What sort of real world actions do I take?
Additional Exposure How does my continued listening in various contexts shape my attraction to a particular song?
Share with Friend What rituals do I perform with my friend? What sort of things do I say and/or respond to?
Incorporate Learnings How do I consciously/unconsciously take in my friend’s feedback? Does our interaction reinforce my music preferences or make me question them?
Relationship is reinforced Through our interaction, are my choices validated or invalidated?
For the purpose of this exercise, I intend on focusing on the last three stages of this process to answer the question: What if we were able to treat our music services more like friends and less like services?
Creating the Personality
Major assumption: In cases where invisible UI’s or digital assistants dominant the interface, personality ties directly to ‘product features’. By mapping out the personality of the person/persona/identity of the application, we can start to generate how they might respond to certain requests.
I remember doing a brand exercise a long time ago where I was prompted to write out a description of my brand as if it were a person. What were its qualities? How did they handle certain situations? I decided to use the exercise for the creation of two different, distinct personalities for TwoFriends: James and Justin, where each personality represents a set of features that could be developed for the application.
By distinguishing different personality traits, one can begin to map out how each ‘person’ (too soon?) reacts to certain situations in the form of behaviors. These behaviors will inform the corresponding product features and interface decisions.
Next, I took a possible behavior and attempted to tie a product feature with said behavior. For example, James enjoys ‘Passively creating playlists for you in the background after asking what you enjoy’. This could translate into a product feature where James is contextually aware of your activity and informs you that he’s collected music for you:
Or a feature where James suggest a set of artists based on two genres you regularly listen to:
A Mockup of a No-UI called TwoFriends
By connecting design elements to personality traits, it was now possible to try and add more realistic elements to make the application seem, to a certain degree, ‘human’.
Adding Character, Personality and Behavior to the Interface
What was really going to make this ‘real’ and ‘conversational’ for me was all the ways that James/Justin could interact with a user in situations where they didn’t need to i.e. exhibiting personality traits and characteristics. This gets at the idea of a ‘fake ui’ where interface designers create additional visual (and in this case auditory) elements of a system in order to conform to a user’s mental model of how a system, process, or in this case, human, should behave.
For example, in one situation I have Justin acting somewhat subversive to the user as he bashes yesterday’s decision to listen to Taylor Swift. In this situation, Justin asks if he can listen to something for his own enjoyment in order to ‘cleanse his ears’.
When was the last time Spotify asked you to play something for it?
Now, logically this makes less than no sense. Why would the application ask you to put something on for IT?? This behavior challenges the notion of our (current) relationship to software and more importantly, creates an added dimension for applications in the form of a personality. When was the last time that Spotify asked you to play something for its own enjoyment? In terms of practicality, it’s useless. HOWEVER, what it looses in practicality it gains in empathy and connection. It’s something that I would ask of a friend and it’s realistic. Because Justin has been defined as someone who might ‘critique your music’, it makes sense that he could pull something like this (it’s also pretty funny). What’s more, it’s a clever way to engage in a recommendation engine.
Another way you could play with time and space outside of the context of animation is to suggest that James or Justin have been actively engaged in activities in the background. James could tell you that he found this great song ‘5 minutes ago’ and that you have to listen to it. This creates a sense of individuality, time and space, as if your relationship and the applications, locality exist beyond a screen.
As product managers, designers and startup entrepreneurs, we tend to think of branding, visual identity (and now) animation as key components in the quest to build an application’s personality and connect with users at an emotional level. This exercise was an attempt to take product development and experience design one step further into uncharted territory.
If nothing else, I see this as a possible addition to the user experience design process as products and services take on a more human, predictive, conversational tone.