LONDON (Hypebot) – As they continue their serach for a new CEO, eMusic has promised to stay competitive with a series of upgrades. It unveiled two recently: a home page redesign and a new recommendation engine that combines human analysis with computer algorithms designed by MediaUnbound.
The new design attempts to solve what has always been a vexing problem for me (and I suspect others) on eMusic, finding how to best use my monthly download allotment. When I visit Amazon or iTunes, it's to purchase something I've discovered elsewhere and I know with 99% certainty that I'll find it. But that strategy often leads to frustration on eMusic due to the vast but spotty nature of it 4.5M track catalog that includes no major label product.
The goal of the new eMusic site is to replicate a trip to the corner record store where the staff gets to know customers' tastes quickly and suggest music they might like. A hybrid engine, created by MediaUnbound combines computer algorithms and human analysis to help users discover new releases, classics and "hidden gems". "The recommendations get better and better with every action the customer takes, encouraging true music discovery and trust in the service," says Deirdre Stone, eMusic VP of Product Development. "(And) it doesn't require you to upload your entire music library like competing digital music services."
eMusic's recommendation engine builds a profile for each subscriber based on their individual actions (searches, downloads, ratings, etc.), specific musical tastes and overall usage. To develop music recommendations, it applies multiple algorithms on the profile along with global data culled from the internet (P2P, blogs, web radio, etc.) to find correlations and patterns between artists, tracks, albums, eMusic editorial content and other eMusic subscribers. Then MediaUnbound adds data, tuning and input from its team of human music analysts.
