The first step in the process for our technologies is to analyze a representative sample of music (up to date we have analized more than 1 Million tracks). In conjunction with different music partners we have access to the audio content of more than 250,000 CDs.

The analysis process measures a number of physical characteristics of the audio, including such parameters as brightness and tempo and how these parameters change over time. The characteristics we measure have been identified in user testing to be the ones that produced the strongest reaction in testers. Often the characteristics are detected unconsciously by the listener, and the mix of parameters is more important than any individual parameter.

Our HSS (Hit Song Science) technology takes the analyzed data and overlays extra parameters relating to the commercial success of the music. These parameters are data such as total sales, highest chart position, date of release and others.

Using this extra dimension, new releases, potential releases and even unsigned acts can be compared with the database to allow a record label to see how well it fits into the current market and to identify potential hits. As the market changes, the system reflects this by finding new patterns in the hit clusters and applying these to the process.

The system allows for trends to be identified as they develop over time, meaning that a song that contains strong characteristics that are becoming more prevalent in new music and less of the characteristics that are diminishing can be identified as having high potential. Simply put this means that a song that sounds uncommercial to a human listening to it right now may just be ahead of its time and in fact contains the right ingredients to appeal to the CD-buying public a few months from now.

Once we have a large selection of music analyzed, there are two key ways for music recommendation to an individual user. One simply links a song or, more interestingly, a group of songs to a selection of music that has a similar profile. This technique takes the individual profile of the song or songs and matches it to the whole catalogue of music in the database. So given a list of songs, each can have a "more like this" link to similar music.

The second way to capture a user's own personal taste profile is to allow them to take a "music taste test". In this process, the user will be presented with a number of binary choices between two short audio clips and will choose the clip they prefer. After a series of these questions, it is possible to generate a profile for that user; the profile is analagous to a song's own profile, as measured in the analysis phase. So in this way songs from the database that share commonalities to this profile can be identified and presented to the user to preview.

In a retail environment, both the "more like this" and the "music taste test" can be efficiently presented on in-store terminals, or on a retail website. The same technique could be applied to many other situations, such as automatically recommending songs from a personal collection as a playlist, or anywhere that commonalities between pieces of music can be useful.

In both cases, results can further be customized as the user can select to receive music matches across genres, from a given epoch, new releases or other customizable variables.