Tag Archives: music

Listen Local: A Pandora for Live Music

27 Jul

I’ve always been a huge fan of music, but never a big concert-goer. Personally, I prefer listening to music to watching it, but when I do make the occasional trip out to support my friend’s band or am convinced to buy tickets to a show, I always end up having a good time. I would love to be able to check out more local bands, but where do you even begin?

In DC and every other major metropolitan area in the country – and around the world, really – there are numerous music venues hosting dozens of bands of various quality and success every week. At some point, as I’m sure psychology professor Barry Schwarz would agree, the seemingly unlimited options for live music entertainment become overwhelming to the point where making decisions becomes an anxiety-inducing endeavor since you’re always left wondering, “Did I make the best possible choice given all other existing options? This band is good, but what’s the opportunity cost of staying here versus trying to catch the 9:30 show at the venue down the street?” This might explain why you won’t see too many economists crowd-surfing at music clubs, but it doesn’t address the issue of choosing which band to check out the coming weekend.

Sure, there are myriad “music discovery services” such as TagWorld and Last.fm, not to mention the recommendations from friends’ Spotify and YouTube posts to Facebook, but these services are largely still designed for mainstream, studio-produced music. What about the local music scene? Gruvr helps users locate local concerts, and countless other websites profile and spotlight local musicians, but all of these options are largely aggregators of music. As in other digital media, I think it’s time to evolve from aggregation to recommendation.

What if there was a Pandora for live music?

Just as Amazon can leverage my purchasing and browsing behavior to make (often spot-on) recommendations, Pandora “learns” from users’ listening behavior and plays music that the user is likely to enjoy due to similar traits such as melody, harmony, form, composition and lyrics.

Pandora’s algorithm uses almost 400 attributes to classify each song in its library, and codifying each song required about 20-30 minutes of human analysis per four-minute song. Clearly, this level of granular analysis cannot be feasibly applied to a Pandora-like service for live music, but I think a combination of self-identification and crowdsourced voting would be sufficient to build a product that users, musicians and venues could all greatly benefit from.

A typical band interview always has a question along the lines of “So how do you define your sound?” Bands, tasked with the challenge of defining and labeling their art, usually respond with a reference to other bands that serve as reference points for the interviewee and readers. The Local/Live Pandora I’m imagining would allow bands (or their promoters and fans) to self-identify with genres, artists or songs catalogued in Pandora’s Music Genome Project. Users can weigh in regarding the degree of accuracy of bands’ self-identification, voting whether the sound is similar and making recommendations for a better match. Over time, enough data would be entered to produce a fairly accurate analysis.

Because the point of the service is to introduce users to bands to see live, the service would leverage smartphones’ geo-location features and the algorithm would only introduce users to bands that are playing within a certain radium and within a certain timeframe. You could also imagine being able to configure personal settings, so if you were visiting Chicago in a few months, you could get a sense of the local music scene and stumble upon a new band to check out.

Obviously the production quality would not be as high as it is on Pandora or similar services, but the point isn’t to use the service as a music player; instead, it’s a tool to identify bands to see live. This implies unique revenue streams that could be harnessed. For example, imagine you sign into your account and the service pushes you a band that you might be inclined to like, since the band or song has been identified as being similar to many other bands and songs that you have already liked (perhaps pulling data from users’ iTunes, Pandora and Spotify accounts). A notice would pop up, notifying you of an upcoming show. You could then click to purchase tickets, and the service would take a cut of the ticket sale; in effect, the service becomes outsourced marketing for local venues, who would be willing to trade a share of ticket sales for increased attendance (and therefore increased bar sales).

Is it just me or is this music to anyone else’s ears?