Big Data and the Music Industry

Music Week, the trade paper for the music industry published an opinion piece from me on the relationship between big data and the industry.

The piece was originally written at the request of  Jonathan Robinson at  Music Tank. I will be speaking on the topic of Big Data at their forthcoming think tank debate – Moneyballing Music:  Big Data, Consumers & A&R,  on Tues March 10th, Central London.

 Further details and booking here – http://www.musictank.co.uk/events/big-data

JS on Big Data in Music WeekHere’s the text of the article:

The music industry has always had an interest in data. Ever since the top 40 became the most popular source of music played on radio, the Charts have held a place of almost religious centrality for the industry. Getting a number one record secured the artist’s reputation and the manager’s ongoing representation, ensured record company and retail commitment, and guaranteed media exposure regardless of the actual number of units sold. For a long period, it might be argued that managers and record companies spent almost as much money trying to make sure that their songs entered high in the charts as they spent on industry-bodies tasked with maintaining the accuracy of the charts.

What is meant by big data in music? Big data often refers to the very large and highly complex data-sets thrown off by global online consumer activity, particularly arising from social media activity such as on Twitter, Facebook, Youtube or from page-views of sites such as Wikipedia. Music big data can also include more traditional measures of information about plays of music audio and video, paid-for streams or downloads of music, or even illegal music sharing activity.

Combining data sets that cover all these areas of music consumer behavior provides a more comprehensive picture than a simple sales chart. It’s possible to build big data platforms and analysis systems these days because we finally have the high levels of computing power to do so. Because of the daily volumes, gigabytes of data, thrown off by all this activity, highly sophisticated and leading edge systems are required to process such vast quantities of data. It’s only in the last five or six years that it’s become cost-effective to build systems of this power and sophistication.

It is however one thing to build platforms and systems that are capable of asking interesting questions of data of this kind. It’s quite another thing to know what kinds of questions to ask – and then to do something useful with the answers. There are only a small number of companies engaged in this kind of activity with a dedicated focus on music. There are a larger number of more general consumer brand focused businesses. The big difference is that the dedicated music companies frequently gather data proactively about all the activity that is going on out there, rather than just what their clients’ commission them to analyse. Comprehensive, proactive analysis provides a much more useful environment for research, but is much more expensive to produce.

It is also important to add that much of the research activity that goes on around these kinds of large data sets in music is based on aggregated, anonymised data. In other words, the analysis is looking at overall trends rather than individual consumer / music fan activity. Large retailers such as Tesco in the UK and Target in the US have become leaders in the researching and personalized marketing derived from analysing individual consumer activities. While, such individual targeting is every marketeer’s holy grail, issues of trust, privacy and downright creepiness add a lot of sensitivity to these kinds of research.

For the most part, big data is being used to help understand what has happened in the recent past – by gender, demographic, geography and online platform or network. It also has the potential to be used to predict the future likelihood of success but this remains a controversial and less reliable field. Analysis can identify both trajectory of individual artists or tracks, but then also see overall trends in genres or different cities or countries or among different sectors of the audience.

These are very early days for big data in music. We are walking the mere foothills of what is possible. Whether it’s in talent spotting, A&R, sales marketing, catalogue revival or brand matching, big data has a big contribution to make. In the future the combination of computer analytics and social science will undoubtedly reveal even more powerful ways of targeting music to receptive fans. I suspect that a lot more big data will flow through the digital gateways before the industry fills the skills gap, which currently prevents it from realizing the real benefits data science can bring to the industry.

Advertisements

4 responses to “Big Data and the Music Industry

  1. Are there any books or papers out there exploring the possibilities or impact of big data in the music industry?

  2. Hi there, no I’m not aware of anything other than a paper on the music tank website here: http://www.musictank.co.uk/events/big-data

  3. Pingback: البيانات الضخمة وصناعة الموسيقى - هايبرستيج

  4. Pingback: البيانات الضخمة وصناعة الموسيقى - Hyper Stage هايبرستيج

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s