Music Ally

Today’s interview with Duncan Freeman, founder and president of US firm Bandmetrics, marks the final part in our five-article series exploring metrics for digital music. Focused on the long tail of the music industry, Bandmetrics is currently helping artists from unknowns through to Tiesto and Death Cab For Cutie using the web to assess the effectiveness of their marketing activities.

Already we have spoken to Eric Garland of Big Champagne; Alan Ault of WaveMetrix; Marie-Alicia Chang and Gregory Mead of MusicMetric; and Brad Little of Nielsen Buzzmetrics.

Duncan’s comments can be found in an extensive feature on many of the companies operating in the music metric space; to view the full article log inor sign up for a free trial of the Music Ally Report.

Read on for the Q&A with Bandmetrics, after the jump…

How did Bandmetrics start?

I played in a couple of different bands in the late 80s / early 90s and after graduating college I went to work in the internet industry in ‘94 and have been in it since then. When I examined what was changing from the way companies and brands looked at themselves in the late 90s to how it unfolded in the mid 00s it was clear social media was going to have an impact. So I began to think about what this means for artists and bands.

My initial thought was there’s got to be a way to gauge comments within the blogosphere. The idea was to amalgamate quantitative and qualitative data from the topography of the net and provide easy but meaningful insights back to a band – that’s what we all hoped to do initially. But as technology improves, as more data becomes available via companies releasing their data via APIs, we’ll be able to improve the information we provide.

Are your targets just the low-level acts?

I wouldn’t call them low level – we refer to them as long tail of the industry. But we are not going after Big Champaigne’s model – if I understand their business correctly they’re looking to do some consultation where they charge $2,000 per band along those lines. The bands we cater to could no way afford that fee. Of course we do have some bigger bands – Death Cab For Cutie for one of them – and if the bigger acts want to use us we encourage it.

What are your sources of data?

We look at the usual suspects: Myspace, iMeem, Last.fm, LaLa, Facebook YouTube – that’s the initial version. Where we go beyond that is into the qualitative realm, the blogosphere, and there’s great content from Twitter – also a lot of great comments people write within various music blogs. We want to pull that data. The step after that is to pull in sales data. We already have a signed deal with the Orchard and we’re waiting on the Orchard to deliver their API. It’s our hope that when an artist signs up they’ll have a unique ID and pull sales data into Bandmetrics. Tunecore and CD Baby will eventually be partners – they’re waiting on us

What kinds of challenges do you face in collecting this data?

One of the Achilles heels for Bandmetrics is that we’ve struggled with development – we’ve made mistakes – we’re working through those mistakes and there are three core problems you try to solve when you’re talking about this kind of data – it doesn’t matter if it’s music, film or whatever vertical. First, the ability to collect data from disparate sources; second, to store it in a way that can scale; and thirdly to analyse the data. They’re all fairly significant problems to solve when you’re trying to create a project in a near real-time environment while allowing anyone to sign up, so if 60,000 bands sign up you can handle it.

How important is the notion of sentiment?

It’s all about sentiment. Everything we do is about sentiment – at the end of the day even an expression of quantitative number of times a song was downloaded is an expression of sentiment. It’s an expression of people being passionate about the song. More and more sites are allowing people to express how they feel about a site or artist – such as on Hype Machine where you could love a track Bebo does it – you can love a band and of course iLike, Facebook has thumbs up, these are all expressions of feeling. Sentiment in real time is the Mecca of popularity – that’s not going to happen in 2010.

What’s in the future for Bandmetrics?

One of the things I find compelling is the ability to look at sales data and make correlations on your social activity. So it’s our mission to do just that and we believe that there are correlations that will be quite intriguing to a band – they’ll be correlations that will untack over a longitudinal period of time. When you start mining this data month after month you can begin to tap into potential trends.

Where we want to go later is natural language processing. It’s going to be a difficult nut to crack and I don’t know if we’re going to get to it in 2010 – the English language has over a million words and each can be expressed differently. So if I say the Dave Matthews Band is great, that’s an expression with a certain weighted value. But what if someone says GREAT !!! (in bold capitals with exclamations) or if they say it’s the greatest band in all time….that one word in different contexts has so many different variables and then you have to take into SMS language, abbreviations like gr8. It’s complex stuff and it’s not as if you can simply download this software, plug it into your system and go – you have to programme the machine in this context.

Example videos of Bandmetrics in action:

http://screenr.com/seH

http://screenr.com/HOH

http://screenr.com/Is7

http://screenr.com/a28

Tags: bandmetrics, duncan freeman, marketing effectiveness, measurements, metrics

This entry was posted on Friday, January 15th, 2010 at 5:27 pm and is filed under Interviews. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.