The history of innovation is the history of failure. When asked about the difficulty of inventing the light bulb, Thomas Edison said, I have not failed. I’ve just found ten thousand ways that don’t work.
The growth of local radio has been fueled by programming innovation. Unfortunately, local radio today seems to be in the grip of fear. Whether it is consolidation, the recession, or new media threats, terrestrial radio seems to have entered a period of inordinate caution. There is little programming innovation going on.
Not so with Internet radio. Operators are looking for a hook, a way to differentiate their products from local terrestrial stations and other Internet stations.
Two services that are taking diametrically opposed approaches are Jelli using crowd sourcing, and Pandora using their Music Genome Project. In this post, we’ll take a look at Pandora. In the next, we’ll look at Jelli.
Pandora’s centerpiece is the Music Genome Project (MGP). According to Pandora, MGP:
Is the most sophisticated taxonomy of musical information ever collected....Each song in the Music Genome Project is analyzed using up to 400 distinct musical characteristics by a trained music analyst. These attributes capture not only the musical identity of a song, but also the many significant qualities that are relevant to understanding the musical preferences of listeners.
Their claim is that:
By utilizing the wealth of musicological information stored in the Music Genome Project, Pandora recognizes and responds to each individual's tastes. The result is a much more personalized radio experience - stations that play music you'll love - and nothing else.
Pandora believes that listeners prefer songs with similar objective characteristics. Once a listener has picked a few songs, the service can create a unique radio station that plays songs coded similarly in the 400 characteristics.
While the approach might seem novel and unique, Pandora is actually following in the footsteps of an earlier programming pioneer, Jim Schulke. Schulke believed that listeners prefer naturally flowing similarly textured music.
He loaded 14 inch magnetic tape reels with hours of songs sequenced with great care to make sure everything blended together well. He called his approach Match-Flow. His Beautiful Music format dominated the ratings for many years.
Another Pandora antecedent is Musak. To many, Musak is synonymous with elevator music, but it is much more than that. The company programs a wide range of formats, all designed to use music to manipulate mood and emotion. Here’s how Musak describes their products:
We call our creation Audio Imaging. It is the convergence of art and science, of methodology and intuition. Emotion by design.
Both Pandora and Musak assert their programming is based on science. Unfortunately, there is more pseudo-science than real science in their claims.
There is no scientific evidence that links the Music Genome Project’s 400 distinct musical characteristics to a song’s appeal. Furthermore, there is no research that proves that if a listener likes one song with a certain genome, that she will like others similarly coded.
The MGP helps convey the impression that Pandora has something special, something that other Internet services lack. In that regard, the MGP is a great marketing tool. It does not appear, however, to have given the service a programming advantage.
If programming by musical genomes created a better more personal experience, one would expect users to listen longer than users of less personal services. Not so with Pandora.
Rather than higher time spent listening (TSL), Pandora has below average TSL. According to Ando Media, Pandora has considerably lower TSL than the majority of services measured. AccuRadio, a more traditionally programmed Internet service has three times the TSL as Pandora.
It turns out that science can explain the problem. Research has shown that people when exposed to similarly sounding songs develop listening fatigue. Listeners need a change of pace from time to time. That’s why formats that sound too predictable generate less TSL than formats that throw in a surprise now and then.
Despite Pandora’s apparent confidence in the MGP, they appear to recognize the problem. Notice the screen capture from Pandora’s home page on the left. Pandora gives you the option to add more variety by clicking a button.
What does that say about the Music Genone Project? The MGP is supposed to create, "stations that play music you'll love-and nothing else." Why would you need a variety button?
Every new format is an experiment. Pandora is an interesting experiment, new in some respects, but with roots in Musak and Beautiful Music.
Next time we’ll examine crowd sourcing as a programming tool.