The Biggest Problem With Music Discovery Services

By Kyle Bylin of sidewinder.fm.

If you want the short answer: a majority of listeners don’t care. They have zero desire to be recommended fledgling artists or emerging songs. The biggest problem is that there isn’t a problem. Music discovery appeals to a niche audience. To gain a broader perspective on music discovery and the biggest problems that services have, sidewinder.fm asked several influential executives in music and technology to share with their thoughts on the matter.

Perfecting Music Discovery Algorithms Is A Huge Challenge

Liv Buli is the resident data journalist for music analytics company Next Big Sound.

Well if anyone had found the key to creating the perfect music discovery service, this discussion would be a very different one. There are issues on several fronts.

At this point, perfecting music discovery algorithms is a huge challenge. For one, tastes are often eclectic and can be hard to pinpoint. It can also be difficult to account for what exactly attracts a listener to a particular song. Is it the genre? The lyrics? The key? The rhythm? The drop?

Add to that the idea of intrinsic value, which can complicate matters further. I might listen to Mary Chapin Carpenter from time to time because it reminds me of summer car rides with my mom, but in general I am no particular fan of country music. Algorithms can account for unique instances, but such occurrences in aggregate are likely to throw off any algorithm that accounts for listening history.

What is more, I recently read a great quote from Marc Ruxin, the co-founder of TastemakerX, who noted that in the game of music discovery “algorithms lack the personal imperative,” something I believe will be difficult for these types of services to overcome. It is harder to get the listener invested in something new without someone trusted telling them: “This is awesome, I listen to it all the time, and you should too.“ (Unless they just came across a really stellar song that is.) This means that by relying solely on automated, algorithm-based music discovery it can be difficult to turn a passive listener into an active listener, who engages with and then retains new music on a consistent basis.

At this point, it seems that music discovery is heading in the direction of social. The Hype Machine is a great example of a music discovery service that tracks the most discussed tracks, albums and artists online. Also streaming giant Spotify has rolled out a series of features that attempt to address the issue by encompassing several approaches. By allowing users to track playlists from peers, artists, and influencers, promoting new releases from favorite artists as well as making recommendations based on a listener’s history, they are positioning themselves at the forefront of music discovery.

What Creates The Tension Around Music Discovery

Jed Carlson is the President and Co-Founder of the artist services company ReverbNation.

A utopian vision for music discovery would guard against the concept of the self-fulfilling prophecy — what some call ‘cumulative advantage’ or ‘the rich get richer.’ Instead, it would be a meritocracy where the best songs and most talented artists rise quickly to the top. But there are natural human tendencies that create a headwind for this utopian vision.

The psychology of repetition tells us that we remember and place value on things that we see/hear 7-10 times, almost regardless of the relative ‘quality’ of the content (though not entirely). This is why terrestrial radio has been such a critical channel, historically, for anyone looking to sell records or break out of obscurity — because repetition can literally train masses of people on what to like.

The best study I’ve ever seen on how music becomes popular is described here. It goes beyond repetition and delves into how we look to others to help us form our opinions about things. What it shows is that the perception that something is worthy (largely based on what others think) is more important than the opinion we would have formed — had we been able to live in a vacuum. This creates a tension around music discovery, as ideally we don’t want to be told what to like, yet humans crave for social cues in order to create context and make sense of the millions of songs out there. As a result, what one person calls a ‘social discovery engine’ another person might call ‘the new music gatekeeper.’ And therein lies the rub.

Lack of Follow-Up Problem For Music Discovery

Cortney Harding is the partnerships lead at the music discovery app company Soundrop.

The lack of follow-up, easily. I listen to tons and tons of music, all day long, and have for years and years. But I can count on one hand the number of new albums I listened to more than five times in the last year.

Part of this is the fact that while I still work in (or near, anyway) the music industry, it’s no longer my job to listen to albums and write about them. But I also find myself really preferring the playlist formula to the album format when it comes to daily listening. Most of my listening happens when I’m running, and I find it hard to listen to one artist for several miles — I like the variety. When I’m listening at home, I usually have music on as background — the Soundrop Chill room, for example.

I haven’t seen a discovery service (and full-disclosure, I certainly have not seen them all) that gives me easy next steps. I hear a song, I like a song — can I easily switch to hearing the record? I can do this in Spotify but the process is a little clunky. If I like the record, can I find out when the band is playing in Brooklyn and buy tickets then and there? Can I easily follow all their social channels with one click to keep myself updated?

Once these problems are solved, I think music discovery apps will be much better. If discovery can lead seamlessly to monetization, I think fans will spend more and artists will be better off.

Music Discovery Appeals To Niche Audience

Jay Frank is the owner and CEO of DigSin and the author of “Futurehit.DNA” and “Hack Your Hit.”

The biggest issue music discovery services have is the notion that people actually want this service. Services such as Facebook have succeeded because they have a mass appeal to wide swaths of people. Music discovery, on the other hand, really only appeals to a very targeted niche of consumers.

In addition, that narrow band of users are ones that have historically been loathe to pay for music to begin with. Since they view themselves as tastemakers, they don’t generally feel that they should be paying for their promotional value within their network. They also prefer to feel as if they actually discovered something by mining for it, even if that mining was biased by a curated light. So when they are pitched the idea of a music discovery website, that loses its appeal to them as it seeks to solve a problem they perceive that they do not have. If this type of service doesn’t appeal to a broad audience, and fails to convince the core to pay for it, it’s a difficult business to run.

To improve this, I would ditch the notion of music discovery and make entertainment the top focus with the discovery a by-product. Pandora is about listening to great radio, but then you can discover music in the process. Popular blogs start with entertaining riffs on artists the reader already knows about, but then segues into something they don’t know about subtly. The more a service can focus on entertaining the audience with music, the more likely the music that is discovered will have greater value. That discovery can purposefully be built into the system, but it mustn’t be obvious to the user.

_

Sidewinder.fm is founded and edited by Kyle Bylin of Live Nation Labs. If you would like to contribute a post to be featured on the site, please reach out.