What are the best resources to find songs that sound like each other?

DID YOU KNOW

Rather than basing the results on the characteristics of the song itself, tools that provide music recommendations analyze the listening habits of people with a similar taste in music. The algorithm creates a profile for an individual’s tastes in music, compares it with other users’ profiles, and fills in the gaps with songs that people with the same music preferences listen to but is relatively unknown to the individual. Spotify, a music streaming service with approximately 159 million users, has the advantage of 2 billion user-generated playlists to scour with their algorithm. As a result, the personalized ‘Discover Weekly’ playlists are a huge success for Spotify as it is able to accurately provide a list of songs that the user is likely to listen to. 

In 2016 alone, less than a year after it was released as a new feature on Spotify, ‘Discover Weekly’ playlists garnered 5 billion overall streams worldwide. Conversely, the program Pandora provides music recommendations based on the Music Genome Project. Thousands of musicologists were hired to describe songs using a list of 450 musical characteristics, such as the gender of the lead vocalist and the type of background vocals. The full list of attributes is kept a trade secret. Using their database of music genomes, Pandora uses a matching algorithm to provide song recommendations that are musically similar to each other and the reference song.