Recommending music is promising that you will make people like, feel or remember something when they’ll listen. What is pushing you to get adventurous and hit “play”?
3. Music is Personal
We associate music with
people, emotions, memories...
● Recommending music is
promising that you will make
people like, feel or remember
something when they’ll listen.
● Failing at recommending
something right is nothing less
than an insult or a
disappointment for users.
● And most of the time they’ll take
it personally.
● When Amazon recommends the
wrong hairdryer, it’s not so bad.
4. - -
Why would you want to discover music? Music
discovery is WORK.
Why?
5. Triggers
- Identity: Music is your identity,
listening to a genre makes you feel
like you belong to a community
- Social status: You value being the one
the others turn to when they want
something new
- Fear Of Missing Out: You don’t want
to be the last one finding out about
Major Lazer
- Boredom: You’re tired of listening the
same old songs. You need to feel
something and be alive.
- Fear of Loneliness/Distraction:
Hearing noise, especially human
voices, is comforting
What is pushing you to
get adventurous and hit
“play”?
7. Passive discovery
- Listening to radio while shopping
- Shazaming during a party
- Get recommendations from friends
when you meet them
- …
When music comes to you
• Slipping in your music bubble while
commuting
• Getting ready in the morning
• Take a break at work to escape for a minute
• ….
When you go after it
Active discovery
8. Is it really about me? Message Content in Social Media Streams
http://infolab.stanford.edu/~mor/research/naamanCSCW10.pdf
« On average, people spend 60% of conversations
talking about themselves – and this figure jumps to
80% when communicating via social media platforms
such as Twitter or Facebook. »
It’s about you. A personal process.
And we’re all at least a little self-involved.
10. /01
/02
/03
/04
Self directed expert
Curator
Curious wanderer
Guided listener
We are not all investing the same
amount of time in music though.
Inspired by: Understanding users of commercial music services through
personas; design implications
http://ismir2015.uma.es/articles/12_Paper.pdf
11. Preferred tools:
- Search
- Own playlist curation
Investment: +++
Guidance openness: ---
Trust in algorithms: ---
Self Directed Expert
Triggers / Drivers:
- Build identity
- Keep « trendsetter » social status
- Get recognized / go to person
- Fear of missing out
- Share/show off tastes
« When I listen to the radio, it’s KEXP,
and it’s usually a really short amount
of time in the morning. I know what I
want to listen to. »
« Pandora (…) they’re missing out on
something and I don’t know what it
would be called, like context, and how
the music makes me feel. »
« I do my own ways of [finding], and I
rely on my friends and people I write
with to recommend stuff. »
12. Preferred tools:
- Search / Advanced search
- Own playlist curation
- Similar Artists
- Channels
Investment: +++
Guidance openness: 0
Trust in algorithms: ++
Curator
Triggers/Drivers :
- Learn something new
- Learn something about me
- Understand how things work and how
they’re linked
- Share knowledge
« I would love to see the metadata
that goes into choosing each song…
I’d love to be able to pick and choose
those attributes, so I could say, ‘ok, I
do like those smooth jazz elements,
but I don’t like the saxophone solos.’ »
« I’m looking for linkages from music
to music. »
13. Preferred tools:
• Charts
• Weekly Recommendations
• Radios
• Curated playlists
• Similar Artists
• Channels
Investment: +
Guidance openness: +
Trust in systems: ++
Curious wanderer
Triggers / Drivers:
- Stay up to date
- Get entertained
- Daydream, escape real life
- Escape boredom
« .. When it recommends me things
that I never would have thought of, so
I think, ‘yeah, I’ll give it a shot’ »
« The serendipity of finding new music
is what I enjoy most. Generally if I’m
listening to new music it will be
because a friend recommended it or I
came across it on Youtube.. I listen to
pretty diverse things. »
14. Preferred tools:
- Radios
- Mood playlists
- Moment recommendations
- Charts
Investment: ----
Guidance openness: +++
Trust in systems: +++
Guided listener
Triggers / Drivers:
- Find ambiance/music for activities
(sports, dinner, ..) or mood (breakup,
party)
- Isolate to focus
- Relax
- Morning / commuting routine
- Drive out boredom
« I mean, I can get this thing booted
up and going within seconds, and then
I’m off doing dishes or whatever,
which contributes to my satisfaction. It’
s going to do what I want it do to
immediately. Boom. Off I go. »
16. - -
TRUSTING
ALGORITHMS?
When we don’t trust
algorithms, and when we do
« Algorithm avoidance »: people prefer
human judgment, and as a result often make
worse decisions
Mistakes are held against algorithms more
than against a human being, under the
(false) assumption that human judgment can
improve while an algorithm can’t.
To increase confidence in an algorithm,
people need to feel more that they (=
humans) are in control:
- Understand why an algorithm predicts a
result
- Tweak the results, give feedback
https://hbr.org/2015/06/when-your-boss-wears-metal-pants
17. Continuously adapting to
constant changes
THE RECOMMENDATION
CYCLE
Change of tastes, time, location, and
context in general lead to various needs:
1. Get to know your user: Keep
learning about his/her preferences
2. Build trust by showing your analytics
3. Advise
4. Learn from feedback: analyze why a
recommendation failed and learn
from that mistake
Get to
know you
Build trust
– how well
do I know
you?
Advise
you
Learn
from
feedback