Next up is a panel session on knowing your fans, moderated by Hypebot’s Bruce Houghton, and featuring EMI’s Cory Ondrejka, Band Metrics’ Duncan Freeman and Million’s Neil Cartwright.Freeman is up first, who explains we have ubiquitous internet access nowadays, and that there’s been a sea-change (the buzzphrase of MidemNet so far) of social networking, with MySpace, Facebook, Twitter and so on. And also the decreased cost of developing a music site – you can have a Muxtape up and running in a few weeks.”What you’ve done is created this platform to launch data at an astounding rate,” he says. So what can you do with all the data that’s being produced around music – everything from how many times people have played x track on MySpace, to people’s comments on their social networking profile or Twitter about a band, song or gig. Which ties neatly into what Band Metrics is doing, of course.Over to Ondrejka, who agrees with all of the above. “Look at that from inside a major, the scale becomes much bigger. The amount of data that we generate even on a weekly basis runs into the terabytes,” he says. So EMI is trying to get a handle on that data. “Then the more interesting questions are how can people do their jobs better – how can artists find fans better?”EMI.com, the D2C site, is the “first piece of that” for EMI – “a first toehold”. So what’s the second toehold and beyond? He says EMI is building out a team to work on the back of what Web 2.0 firms and startups have done. “It’s looking at what we can fix in terms of how EMI can handle music – we do a lot from the time when we get the master all the way through distributing that – so making that process efficient is our first task, then we can get to more fun stuff.Now over to Cartwright, who says Million has been working on lots of advertising stuff recently. For example online advertising for live promoters – “the advertising business has been doing al this for years – behavioural targeting – and these networks cover upwards of 85% of the most trafficked-websites, tracking you from the moment you visit a site, and the sites that you visit subsequently. So if you were to visit one of our artist sites and subsequently visit NME or Tiscali or football – you will suddenly see an advert that is relevant to you” – because of a cookie attached when visiting the artist’s site.”Revolutionary!” whispers the man sitting next to me. In a sarcastic way, I fear.What else is Million doing? It’s trying to build a system that does a similar thing, so if users have read an article about an artist, they’ll see an advert saying ‘do you want to listen to their music?’ while also tying in Facebook Connect, so that people can then tell their friends about that song or artist. Million is also working with Google on ‘gadget ads’.Back to Freeman. “If we can reach this holy grail of 1 to 1 marketing…” – but he stresses while none of this is new in general internet terms, “it has not been vertically focused” (i.e on music).And now Ondrejka, who talks about the way data can be mined to see, for example, which cities are hotbeds of fans for a particular artist. “We’re trying to make things easier for artists and for fans,” he says.So how does advertising fit in? Ondrejka talks about the importance of AdSense and says it’s about “drive-by monetisation” (I’m not quite sure what it means, but it does sound impressive, eh?). So that’s one type. But beyond that, “if you look at people’s music experiential behaviour, you tend to see this diminished pattern [as they get older] of music discovery. So if we can help people not slip into that…”But he stresses that it’s not just about the labels – it’s about “all the smart people out there” using the data “better than we could have”. Does that mean EMI will be making efforts to release the data it holds to third-parties, I wonder.Back to Freeman – how important is the quality of the data, how do you identify what’s good data, and sort and filter it? He says there are several challenges – the retrieval challenge where “you go out and pull this data from all these disparate sites”, and then the storage challenge – how to keep it – and then how to process and analyse it. “Data is only as good as it’s being analysed and provided back to someone”.He says it’s a community effort to make sense of it – Band Metrics asks its artist customers what their needs are, then builds out its tools to deliver that.How about issues of privacy, and the interrelated issue of who owns the data. “The last thing we want to do is figure out who they are and then piss them off,” says Houghton. “We’re quite capable of pissing them off without that” shoots back Ondrejka, smiling.”When you think about data that allows people to do their job better, in a lot of ways your goal is to get that data out. The stricter you try to hold on and own that data, you may end up defeating your goals. If your goal is helping fans discover music, or helping artists discover fans, hiding that data is probably not the best strategy.”Houghton points out that artists can do a lot of stuff on MySpace or Facebook, but they don’t control that data. Is that an issue? Cartwright says it is. “We point that out to artists on a regular basis,” he says. “Bear in mind that MySpace can occasionally break, you don’t own the data – you do need to get the email addresses. Facebook is a bit better in terms of being able to add friends, but you can’t add as many.”He says a lot of websites let you download the CSV files, but that you have to still respect fans’ privacy – “just because they’re fans doesn’t mean they want to be sent a message, particularly on mobile”. For example, when they buy gig tickets they’ll often leave their mobile number, but that DOESN’T mean they want to receive texts from the artist. “It’s extracting the ones who do from the ones who get a bit angry, even if you text them saying thankyou for coming to the show,” he says.So how do you get around that? “It goes back to knowing who your fans are,” says Cartwright. “The ones who buy the deluxe edition and go to see you play constantly. They’ll the ones who’ll use the artist. What MySpace and Facebook do is allow you to communicate with the millions of fans who don’t like you that much. They like you, but they just want to buy the odd track, they might go and see you live and buy the album, but they don’t want to be emailed.”It’s about ‘levels of fans’, in other words.Can we use all this data to identify patterns and predict the future – or as Freeman puts it, “predict the next Vampire Weekend”? He thinks we’ll get there as we grab more data and analyse it over time.”It’s interesting,” replies Ondrejka. “As a newcomer to this space, it’s interesting that there’s not historically more introspection. We don’t know the what’s of what’s happening, and often don’t know the whos. But the why is easily the most interesting question and the most fun to dig into. But shorter term, I want to understand the what and get a picture of what’s happening.”EMI looked at Katy Perry’s website for example, “several months before anything broke”, although EMI wasn’t monitoring it at the time, and couldn’t figure out why it happened. “Going forward, we want to know – if there’s a bump on an artist’s website, something caused it. So we want to understand the why, but shorter term we have a lot of work to do to understand the what.”Got that?Freeman is now talking baseball stats, and how the Oakland A’s looked at data differently – this is that sports book that every football manager in the world is currently reading – “What was most important was the percentage that a player could actually get on base,” says Freeman. “The takeaway is what is underneath the surface – the data that hasn’t surfaced that can tell us something much larger”.Ondrejka agrees, and points out that the research that led to the Oakland A’s came from amateurs – rotisserie players (I’m English, I may be misspelling baseball jargon) – stattos basically. But his point is how can the music industry enable those kinds of people to make these discoveries for music.Back to Cartwright. “The more you can understand where to place an artist – which blogs are talking about an artist, and maybe as a result of those blogs you get an uplift in visits to your site. Does Twitter matter? We’re analysing Twitter posts and how many people are talking about you, for example.”Houghton says that the books and references from this session will be posted on the MidemNet blog by next week, so you can go out and read them.Question from the crowd – what about smaller artists that DON’T have the same level of data as a Katy Perry? Is there a threshold that becomes necessary to mine all this data and create new revenue streams? Cartwright says Million works with a lot of developing artists, where the advantage is establishing stronger relationships with fans from the start.Ondrejka talks about community management – “the joy of having 20 fans is you know all of them,” he says. “It’s much cheaper to be generous with them. And out of those 20, who are the connectors – who are the five of them that if you send them something, they’ll go out and tell everyone how cool you are.”But he stresses that this has to change as you build a larger audience.A question for Ondrejka, about whether artists can do this themselves without the need for a label. Radiohead, for example? He replies pointing out that while some artists have grown up with this tech, others haven’t, and need the label to help them. And that’s it for this session.