Business Intelligence in the Music Industry

The use of social networking and digital music technologies generate a large amount of data exploitable by machine learning, and by looking at possible patterns and developments in this information, tools can help music industry experts to gain insight into the performance of the industry. Information on listening figures, global sales, popularity levels and audience responses to advertising campaigns, can all enable the industry to make informed decisions about the impact of the digitization on the music business. This can be achieved through the use of Business Intelligence assisted with machine learning.

Machine Learning is a branch of artificial intelligence, which gives computers the ability to implement learning behaviour and change their behavioural pattern, when exposed to varying situations, without the use of explicit instructions. Machine learning applications recognise patterns as they emerge, and adjust themselves in response, to improve their functionality.

The use of real-time data plays an important role in effective Business Intelligence, which can be derived from all aspects of business activities, such as production levels, sales and customer feedback. The data can be presented to business analysts via a dashboard, a visual interface which draws data from different information-gathering applications, in real time. Having access to this information almost immediately after events have occurred, means that businesses can react immediately to changing situations, by identifying potential problems before they have a chance to develop. By being able to regularly access this information, organisations are able to monitor activities closely, providing immediate input on changes such as stock levels, sales figures and promotional activities, allowing them to make informed decisions and respond promptly.

Using Business Intelligence to monitor P2P file sharing can provide a detailed insight into both the volume and geographical distribution of illegal downloading, as well as giving the music industry with some vital insight into the actual listening habits of the music audience. By analysing patterns in data on downloads, music professionals can identify recurring trends and respond to them accordingly, for example, by providing competitive services – streaming services like Spotify are now driving traffic away from P2P filesharing, towards more monetizable routes.

Social networks can provide invaluable insight to the music industry, by giving direct input on fans’ feedback and opinions. Automated sentiment analysis is a useful method of gaining insight into these unofficial opinions, as well as gauging which blogs and networks exert the most influence over readers. Data mined from social networks is analysed using a machine learning based application, which is trained to detect keywords, labelled as positive or negative. It is necessary to ensure that the technology can adapt and evolve to changing patterns in language usage, while requiring the least amount of supervision and human intervention. The volume of data would make manual monitoring an impossible task, so machine learning is therefore ideally suited. The use of transfer learning, for example, can enable a system trained in one domain to be used in another untrained domain, allowing it to keep up when there is an overlap or change in the expression of positive and negative emotion.

After the available data is narrowed using machine learning based applications, music industry professionals can be provided with information regarding artist popularity, consumer behaviour, fan interactions and opinions. This information can then be used to make their marketing campaigns more targeted and efficient, helping in the discovery of emerging artists and trends, minimise damage from piracy and help to identify the influential “superfans” in various online communities.



Source by Oswald Bousseau