Data. Data data data. Big data. The music industry is flooded with data: sales, streams, views, Likes, tweets, shares, pins… But how to make sense of all this data to make a difference to the careers (not to mention the incomes) of musicians?
This afternoon saw a panel debate on Getting International with Big Data, moderated by Ken Hertz, senior partner at Hertz Lichtenstein & Young, and principal at memBrain. The panel featured Ime Archibong, who handles strategic partnerships at Facebook; Scott Cohen, VP international at The Orchard; Gregory Mead, CEO at Musicmetric; and Christophe Maignier, director of strategy and resources at Sacem.
“I think data underpins our business and always has, it’s just becoming more and more relevant,” said Cohen. “It’s nice to know we had a stream on Spotify or a sale on iTunes, but that’s not where it stops. That’s just the beginning of it… It’s the analytics around the data, it’s making sense of all that, so we can figure out what we’re going to do next.”
Is data the greatest thing that ever happened to the music industry, or the worst thing that ever happened to the people that work in the music industry, wondered Hertz. “It’s either a hairball or it’s very important and very revealing. How do you bridge that gap?” he asked. Mead said Musicmetric’s business is about hiding the data, and helping people make decisions.
Cohen mischievously pointed out that at the music industry’s height, it had very little data beyond sales figures: it was all based on guesswork. But a serious question ensued: if the industry could grow based on gut feeling and talent-spotting, what is the impact of big data now – is it being used in an intelligent way by those traditional companies?
“When you think about all of this, you described the industry perfectly. Without data, we still had charts… and we relied on promotions. How do you get on the radio, how do you get press, and later how do you get a main page placement at iTunes. But the world has shifted now, and understanding data and algorithms is more important… Now you gotta understand how it works. That describes Spotify, Deezer, Rdio, Facebook.. Now you gotta understand what underpins all this, so that your stuff surfaces to the top at the right moment with the right person.”
How important is music to a company like Facebook? “Music continues to be a content type that remains remarkably important to the Facebook ecosystem,” said Archibong. “Music as a conversation starter inside the Facebook ecosystem has been around for a decade.”
Sacem’s Waignier talked about the challenges of coping with a huge quantity of new data, and distributing it to the collecting society’s members, but was challenged by Cohen. ”The opportunity is the data. Maybe the data is more valuable than the music,” said Cohen. “The data you collect from the usage of that music will be more valuable over time and generate more income than just collecting from the music.”
The conversation turned to music discovery, and big data informing that. But Cohen wasn’t impressed: “Discovery is not a problem. Recommendation is not a problem. I think it’s already been solved: nobody has a problem finding music, listening to music, anything… And they don’t want more! They have enough music. Are you gonna introduce a new song to them from an artist they’ve never heard once a day? Come off it… It’s too much… It’s too much noise. They want the things they want.”
Hertz asked who is taking all this data and giving back helpful information, or doing something intelligent with it? He noted that the music industry historically didn’t have a good track record on research. Archibong said he’s encouraged by services helping artists crowdsource their tours, playing where the most fans request their presence.
“Who’s doing it well? All the big companies you know. Oftentimes it’s the music industry that’s not using what’s available. Who’s checking their Google analytics every day? Their Facebook analytics? It’s all there.” Waignier said SACEM is using data to collect money and then distribute it. “The problem is quality of the data,” he said. “We are talking about money for our members, so it’s not just a question of big and volume, but it’s a question of good quality.”
Mead talked about the next phase of Musicmetric’s development: “Trying to automate what an analyst would do who has a phD in statistics,” he said. “Something that cuts out all the crap and looks at what artists who are successfully promoting themselves are doing, and puts it into context.”
How about fake fan buying, and people trying to game big data? Mead said Musicmetric is putting a lot of effort into detecting this and correcting it. Meanwhile, Archibong was asked if, when datasets become so enormous, they become ‘self-cleaning’ to avoid being tainted by people trying to game the system. ”There’s definitely some of that. We have an army of teams across the globe that have been spending the past 10 years thinking about this… We take pride in doing that at scale,” he said, about combating it.
Cohen finished off by talking about the coming benefits of big data: “The issue is we’re still in a Web 2.0 environment where we’re interacting with information. When we move to a Web 3.0 environment where instead of working for machines, machines start working for us, then we’ll start to see this,” he said.
And Mead talked about being “right at the start of recommendations and targeting. It’s the Alta-Vista of search engines. It’s very young. I still get targeted by things I’m not interested in, because they don’t have enough information on me. That’s something that needs to be worked on.”