Spotify, We Need To Talk About Your Data…
An Open Source Product Strategy for Hipsters and Data Scientists
It was the twenty-something-th time that I found myself at another not-so-sold-out show for my favorite band at a tiny club in SF. After the set, the lead singer, Eric Frederic, who I admired like a god but who had probably got to know me as the kid too young to be at his show, handed me a burned CD with the word “Wallpaper.” sharpied onto it.
That was 2006. Since then, Eric Frederick, known today as Ricky Reed, has taken Wallpaper. to impressive heights and produced Jason Derulo’s two recent top-5 mega hits “Wiggle” and “Talk Dirty To Me.” This guy is going to be a super star if he isn’t one already.
I like to tell this story, admittedly, because I feel like a genius for recognizing Eric’s musical talent before the rest of the world did, especially my friend Nick. Indeed, we all have told this story:
“I liked them before they got big!”
But alas, we don’t have the data to back it up. But Spotify, you do…
As one of the most popular music streaming services in the world with over 40 million active users spending hours a day on your platform, you generate massive amounts of data. Listens, likes, follows and favorites for all users are collected, categorized, time stamped, and then — as far as the I can tell — locked away forever. You have flirted with giving us users a taste of our data, but still favor using it to feed your increasingly expensive recommendation engine that still thinks Skrillex would be just perfect on my Michael Buble radio station — a collaboration between the two would be cool though!
Why, Spotify? All we want is to better understand how we listen to music and maybe prove to a certain friend of ours that we really did find Wallpaper. before he did — screw you Nick.
You should want it too. It could really help your business.
A 2013 study found that “music subscription services are all losing money, and that is going to remain the case until they find a way to monetize a worldwide user base.” $10 a month for Spotify Premium membership from a mere 20% of your total users isn’t going to cut it anymore, especially when competition is stiff from the likes of iTunes and Pandora. The study also suggests that “sustainability of an ad-supported revenue model is a big question mark.”
Maybe I’m oversimplifying by saying that your monetization woes can be resolved if you build data visualizations into your platform… but they can. At least a little.
Let me explain.
Introducing Spotify Analytics, For Everyone
There are two ways I think you can add visualization to your platform. The first is simply adding a stats button to every user and artist’s profile, showing relevant visualizations. Don’t give this away for free. Bundle it with Premium and incentivize the crap out of users to want it (more on that later).
The second is a brand new product — I’d call it “Spotify Insights,” but you might call it, “Wow, This Is Way Cooler Than The Next Big Sound.” It would essentially be a market intelligence tool for record labels to use to source and track musicians. You could charge a nice subscription fee for this.
Let’s look at some specific visualizations you might include in each product.
Start Out Easy
Whether I’m looking at my activity data from Fitbit or sales data from Salesforce, at the most fundamental level, I’m going to need a simple dashboard to understand three things about it: what, when, and where.
For music, that’s what am I listening to, when am I listening to it, and where am I listening to it.
A simple visualization of this data could yield some interesting insights. I might discover that on my most productive days at work, I listen to 90s pop until lunch then alternate between dubstep and baroque until I leave… who knows. Ok, maybe its not a groundbreaking discovery, but it sure as hell would make Spotify Premium a little more enticing to me and the 75% of other Spotify users still reluctant to shell out $120/year so we don’t have to listen to ads we already zone out for.
For record labels, this information is a little more interesting; mostly because there is money involved. Say you’re a record label exec, would you kick off a Southern US tour for an artist if the number of listeners in the South has stagnated for the last 2 years despite a Texas tour and a few other scattered shows? What if they started blowing up in the North East in the last few days? Hopefully, you have the tools to figure this out today, but big red marks on a map would probably make things a little easier.
Thinking A Little Harder
You can uncover even more exciting (monetizable) insights for users and record labels after answering a few key questions with some basic math:
- Who are the early adopters for a particular artist? Take the top quartile of users sorted by the date they first listened to the artist. For Wallpaper, I’m definitely up there.
- Who are the biggest fans of a particular artist? Take the top quartile of unique users sorted by the total number of listens for the artist. Perhaps average total listens over time to differentiate that artist a user has been obsessed with for the last month and the artist they’ve seen 11 times in the last 5 years.
- Who are the experts in a particular genre? Add the two metrics above for each user across multiple artists in the genre. Chances are that if you listen to Mayer Hawethorne’s full discography every day and you were into Aloe Blac before he made it, you know your Pop Soul.
There are an infinite number of uses for this information. I’ve got 3 here, you cover the rest.
What would happen if you rewarded users for being an early adopter for an artist or becoming an expert in a particular genre?Even if you gave us stupid badges, something great would happen on your platform.
Especially for the hipsters among us, the competitive pressures of being known as a music expert would drive us to more actively and exclusively use the platform. No more Rdio because Spotify won’t give us credit for the 10 hours of obscure indie music we listened to there that one time. And no more passively listening to Today’s Top Hits but telling our friends we were the first to like Wallpaper like Nick does — we’ve got mathematical proof that isn’t true man.
Psychological pressures aside, just being able to know which of my friends would be best to talk about post-progressive-dubstep-bluegrass-hau5 music with would, again, be further reason for me to sign up for Spotify Premium.
Record labels could use genre experts as early growth indicators for unknown artists. If multiple Pop Soul experts with a validated history of discovering artists in the genre before they get big are all listening to a particular artist, chances are we’ve got a winner on our hands. This could be revolutionary for poor A&R interns who currently source new artists by going to a lot of shows and trying to convince their bosses to have the same ridiculous taste in music they do.
These growth indicators could also be used by labels negotiating music licensing contracts for existing artists to build more accurate pricing structures — just like the stock market.
What it comes down to is that we don’t want to pay for music anymore. We will, however pay for insight and understanding, especially if our jobs or reputations depend on it. Spotify, you have the most expansive music listening data in the world, its up to you to bring it to life for us.
But until that happens, you’ll have to just trust me — Ricky Reed and I go way back.
See?
Follow me on on Twitter and Medium if you enjoyed this post!
I’m also a musician myself. Have a listen on Soundcloud.
