Can Big Data Help Music Labels Find That Perfect Backbeat?

Are we now witnessing the clearest signs ever that the music business could be turning into the next frontier for the kind of quantitative approaches previously seen on Wall Street and in Moneyball? Three deals at the intersection of data and music have been announced in the last few days:

1. Beats Music bought Topspin, a platform centered around data and analytics to connect artists with their fans.
2. Spotify acquired the Echo Nest, whose algorithms can allow users to find more music that’s similar to what they already like.
3. Shazam signed with Warner Music Group, giving the record label detailed statistics to help it find new artists.

With the growing volume of quantitative data about listeners’ tastes and habits, can smarter analysis lead to more predictable hits, higher sales, and a reduction in risk in an otherwise eternally unpredictable business? The industry may be headed in a direction similar to what we’ve seen in sports: numbers and data dictating strategy.

At the earliest stage in the music life cycle—before a song has been released—the most amazing use of technology might be a software program that can listen to a song and immediately predict whether it can be a hit. That’s exactly what Mike McCready’s Hit Song Science aimed to do a decade ago. By analyzing the “acoustic properties and underlying mathematical patterns in a song,” the product was designed to help label executives make decisions over which songs to release and promote.

At the time, the business model for Hit Song Science was not feasible, as too many music executives were sticking with an I-know-what-I-like approach. The music industry has been slow to change its ways. As McCready is fond of saying, “In the race to adopt new technology, the music industry finishes just ahead of the Amish.”

McCready now runs Music Xray, an online platform where artists submit their songs to industry professionals. Its technology, quant metrics, and predictive analytics help record executives and songwriters find each other. The tech tools allow good songs to bubble up to the top, as measured by what the industry execs tend to prefer.

McCready wants to change the dynamic from a “needle in a haystack” to a “needle stack,” where everything you find is a possible hit. The product can help only with the introduction; in the end, it’s the humans who have to agree to make a deal with each other. In 2013, McCready says, his platform helped 4,578 artists get selected for opportunities in the industry.

Music Xray is great for artists looking to sign a contract, but for the ones who are already on the radio, the benefit of Shazam comes in. Shazam’s app is known mostly for giving its users the ability to put their phones right up to the radio and click a button to instantly find out the name of the song, the artist behind it, and even the lyrics. That data add up: 17 million clicks a day, 500 million per month, from all over the world. Shazam has more data than you can imagine about where and when a song is blowing up.

Chief Executive Officer Rich Riley says the momentum of Shazam’s app clicks reveals which songs are going to be rocking the charts in the near future. For example, the service’s early data—just two days of preliminary app usage—predicted that Katy Perry’s Roar would be a bigger hit than Lady Gaga’s Applause. Riley says Shazam’s data give record label executives the ability to know which songs deserve increased investment in terms of marketing and promotion and which songs should be cut loose. With data specific to individual regions and localities, it might even help touring bands determine which cities to visit for concerts.

Nudging Our Habits
Anytime you hear about “platforms” that are “connecting artists with their fans,” that’s corporate-speak for using analytical data to understand which songs you like and to give you more of them. The music industry has several methods of affecting our tastes. One simple, traditional way is song rotation and placement: Playing new songs around established hits allows people to more easily get accustomed to unfamiliar tunes. This is the same reason new TV shows are scheduled with strong lead-ins, as a way of helping them succeed.

How much of music is actually data-driven, vs. just the whims of what people like? Academic research suggests that hit songs are often the product of what other people like: Song quality matters less than getting a small head start in our social networks or on the top 40 charts. Can music be considered another “stock market” type of environment, where bubbles are formed based on people chasing hot stocks?

If Jimmy Kimmel can get a bunch of concert-goers to agree to like bands that don’t actually exist, what does that mean for people expressing interest in songs that do exist? Are we all so malleable that anything goes? If our musical tastes can be encouraged or skewed or developed or tricked by games such as habit formation, numerical data analysis, and bubble-esque repetition, what does that mean for the value of what we individually actually like?

The Role of Humans
Not everybody is ready to give up to the computers. In the example of Hit Song Science, its algorithms were good at predicting what songs would never be a hit—but the software didn’t necessarily know which songs would dominate the charts. The software could be fooled, however, by certain “mathematically pleasing” songs that any human would agree would never be a hit—imagine, say, a song featuring an accordion.

McCready’s newer project, Music Xray, is back on the side of the humans. He wants real music executives to help cultivate our tastes, giving us a shared experience that we can all participate in, rather than the computer-driven playlists that want to separate every individual into their own niche.

As crazy as it seems now, even the Beatles needed a lot of help and luck through human circumstance to gain traction in the U.S. The band’s original singles went nowhere stateside and only took off months later.

Amy Doyle is a senior executive at MTV in charge of discovering and promoting artists on the full platform of the company’s networks. She says her team still relies on their eyes and ears, and in their efforts to find new superstars, they go with instinct and gut—it’s all done through human curation. Doyle says that if the industry comes to a place where “music is chosen by math and algorithms, you take the emotion and soul out of it,” and that she would leave the business if that ever happened.