[Manic Monday] Empowering Music With Data

Several years ago, when I was still in University, I owned a USB thumb drive. It was still relatively new then and not many people owned one, to the point that I had to carry around the installer CD everywhere. That thumb drive was capable of storing a whopping 64 MB, much larger than the box of floppy disks that I carried everywhere, and was most certainly less prone to fungi. Today, a thumb drive with such capacity is probably not for sale anymore; even 1GB drives are usually given away as bonuses.

The revolution of storage technology supported the growth of various services and games taking advantage of the ever-increasing capacity, in step with ever-increasing internet speeds. More and more data was stored and transmitted - even a smartphone has an internal memory larger than my old thumb drive. The internet itself is not only an data delivery system, but the data is already processed according to our needs. The simplest example: a search engine will look for all web pages that are suitable to the search terms given.

Several music services have been empowered by the sheer amount of data available; even the amount of songs grows each year - a global music service like iTunes offers 26 million songs. With such a large number, a music service needs a strong and comprehensive search technology, that can respond to what the listener is actually looking for. Services like Pandora even only need a keyword to find suitable songs, and will play songs that match the keyword most. To determine keyword compatibility, Pandora uses more than 400 data attributes related to each song, to determine the next song played.

On the other hand, data for Indonesian music on services like Gracenote and Echo Nest, that complete music metadata for various global music services, still depend on music fans submitting the music data rather than the musicians or the record labels. This is in spite of how this data can be useful for recommending songs to other users who may have not heard the songs, hence increasing the potential amount of listeners.

With the power of data and its analysis, we can know which song is being heard the most, at what time, by whom, in what area, and how long it stays popular. How fast does a song spread in circles of friends, and how fast they stop listening to it. We can measure, which musician is currently in trend, from which songs, and is most often heard by whom. With the same power from data, we can recommend similar or related songs.

Empowering music with data needs a cooperation all through the music value chain, with a simple agreement: the music stored, tracked and analysed must use the same metadata standard. This is the smallest step that needs to be done by those who produce music. This data will be strengthened with market-driven data down the road. The more data obtained and analysed, the easier it is to deliver music to more suitable listeners, and the higher the chance that the listener will spend money on a recorded music product, or invest in another music product.

Ario is a co-founder of Ohd.io, an Indonesian music streaming service. He worked in the digital music industry in Indonesia from 2003 to 2010, and recently worked in the movie and TV industry in Vietnam. Keep up with him on Twitter at @barijoe or his blog at http://barijoe.wordpress.com.