Looking Back @ Stereomood
An outstanding music tech project
It’s 2012, there is no Spotify, Last FM is still very cool and Pandora is gone.
Very few buy vinyls like today. Very few pay to stream.
A good friend of mine is working on something interesting, he says. He is Claudio Gallo, owner of Bad Panda Record, an early bird net label that launched Dumbo Gets Mad and the mighty Indian Wells, and released fabulous music over 8 years.
I am finishing my degree in interaction design and running a music studio in Venice, when he approaches me asking if I’d be interested to work with the Stereomood team.
What is Stereomood in 2012?
It’s a music streaming service which database can be browsed by mood.
You can feel: Happy, Sad, Frantic, Bouncy, Calm, Peaceful..
You can be: Running, Studying, Jogging, Working Out, Eating Dinner
And many, many more.
The music catalog that counts tens of thousands independent tunes comes from the best music blogs active online at the time, and songs are assigned to moods by the team that manually tags each of them.
I jump on
It’s the end of 2012 when I start to work for Stereomood as interaction designer. The team is great and the job is really music oriented: I quickly learn how to deal with music players and start to understand what are the problems of having a huge database and giving it out for free.
The scalability problem emerges pretty soon, as the company is growing and has to step from simple player to community with a more complex revenue model.
We can’t keep on tagging the songs manually: we need to ask for help to the community.
The community makes a mess! We need to develop a technology that helps defining categories by mood.
Each mood has an album cover: we need to interpolate image services like instagram, imageshack and others to pull images as bulk into the database.
Let’s have a closer look to numbers:
There are over 1000 playlists.
Each playlist counts up to 900 songs.
The moods we can expose to users are around 20.
As you can see, a big part of our database is not easily accessible. This is what I have been working on for the next year, together with Eleonora Viviani and Daniele Novaga, we gave a shake to the interface so that people could jump to the opposite mood, type their own mood and get a match, get suggested mood through the existing database relationships between songs or between playlists.
We could do it for different reasons, one of which is that the IT team guided by Maurizio Pratici acquired a powerful analyst named Davide Totaro.
They developed an algorithm that automatically defines the mood of a song and place it in a 8 poles graph.
The engine checks speed, tone and scale of the tunes, and spits out results which are close to human opinion with 30% accuracy.
This is still one of the coolest things that happened to me, seeing the algorithm running on such a great database and distributing it over the graph, to show how much sadness, happiness, hecticness, melancholy or weirdness our service was offering.
Our design team saw a million possibilities popping up:
we had thousands of dots to link.
Being all of us music lovers, we started to define meaningful connections, fine tuning the opposites, putting some story telling behind some choices and pushing some moods over the others through marketing campaigns.
We started to cooperate with independent musicians to build curated playlists and release them as specials through the portal and through Mood’o’Clock, a mood radio alarm for Android that I designed.
To hire musicians and ask them for their music, we would go to parties in Rome and Milan. (I remember one of the legendary parties at Pescheria in Rome when Cosmo from Slow Motion was playing. At that time we were working with Jolly Mare for the very first Mood’o’Clock release).
Stereomood technology started to be the center of the work, while the community was getting smaller and too difficult to maintain. Spotify came in and it was a game changer. Music recommendation was growing bigger as a hype and Last FM was almost over. I spend my last months at Stereomood mainly working on the Spotify app as mood discovery engine. The company was sold as a technology and I moved to Berlin.
Stereomood experience was great, as it joined my passion for independent music with my will to design. The challenges were real and the market has been responsive for a while to us, so we could really measure the impact of our moves.
The people at Stereomood raised a great idea from scratch, and they did it in Italy, which is a tough country for music tech in 2013.