Discover new worlds of music using just a musical seed.
Gracenote Discover™ can recommend content using other content from catalogs matched to the Gracenote Media Database™, from MusicID®, or Mobile MusicID™. For example, you can use a track from a matched music catalog as a "seed" and that seed would return recommendations that you would probably like.
Gracenote Discover delivers highly effective recommendations for new artists, albums, and songs based on examples. Users can generate recommendations either while browsing an online music store catalog or looking through their own music collection.
The First "360-degree" Personalized Recommendations
Unlike other solutions which can only produce valid recommendations for music released in a particular geographic territory, Gracenote Discover can be deployed virtually anywhere on the planet to deliver recommendations incorporating local content that hit the mark every time.
Most recommendation solutions typically are only aware of a consumer's recently played or purchased music from a single provider's own service. Discover integrates with Gracenote MusicID to recognize and analyze all of the songs in a consumer's music library, not just a subset gathered from an individual music service.
Greater Ability to Scale
Most music recommendation systems use a single approach to generate results. These results often do not scale due to the limitations inherent in each technique and are not scalable with the rapidly growing digital media market. Gracenote Discover's proprietary system for generating recommendations combines three powerful approaches, while amplifying the strong points and compensating for the gaps in each technique.
- Editorial – Gracenote's international team of music experts is continually categorizing artists, albums, and songs into over 1,600 micro-genres, as well as assigning other descriptive attributes such as eras, artist types, and regions. This method enables the Discover system to consistently identify music which shares similar inherent subjective qualities, best identified by a human expert, across global content catalogs.
- Digital Signal Processing Analysis – Automated and scalable computer-based analysis of the audio waveforms of individual songs using Digital Signal Processing (DSP) techniques can objectively determine musical characteristics such as tempo, timbre, rhythm, instrumentation, harmony, melody and structure of individual songs. Gracenote can integrate third-party DSP data as well as provide even broader coverage through its own database of DSP-derived song descriptors.
- Music Community – Gracenote's community of millions of music fans, using popular Gracenote-enabled media players, provides insight into global music consumption patterns. Discover utilizes such exclusive data to create recommendations and can also integrate its customers' own sales ranking data or even third-party collaborative filtering results to augment the recommendations provided by the Editorial and DSP Analysis modules.
As combined and optimized in Discover, these three approaches complement each other to provide more consistently accurate recommendations across all possible situations than any of the techniques can produce individually.
High Reliability and High Performance
Gracenote Discover is deployed within the infrastructure of the music service provider or on-line store customer, minimizing any real-time reliance on an outside service. Additionally, Discover has been architected from the ground up to deliver results extremely efficiently, reducing the delays produced by excessive real-time calculations.
An Open Solution Puts Control in the Customer's Hands
Gracenote combines customer catalog data and other input to create tunable, targeted recommendations for users. The customer can integrate their own proprietary user data (for example, purchase history or play popularity) to help power recommendations.
Additional controls available to the customer include the ability to:
- Optimize recommendations to match regional catalogs and preferences - by utilizing Gracenote's regional popularity statistics
- Control which similarity criteria are given the most importance in recommendation calculations
- Prioritize specific content to surface catalog items for higher promotion, such as new releases, exclusive items, etc.
- Tailor the number, variety of recommendations presented to a consumer - providing the flexibility of either a limited or wide range of popularity and similarity levels.
- Integrate Gracenote Link to incorporate additional third-party content into the recommendation package to create a more compelling shopping experience.
Generate Recommendations Based on Any Album, Artist, or Song
Although Discover will only recommend merchandise available from the customer's available-for-sale catalog, Discover can use essentially any song, album or artist – regardless of its presence in the store catalog - as the starting point or "seed" for a recommendation. Additionally, any song, album or artist in the user's own personal digital music collection can be used as a recommendation starting point by leveraging the full Gracenote Media Database of more than 60 million tracks.
Enables Rapid Deployment of Recommendation Services
During initial set-up, the customer supplies Gracenote with their merchandise catalog data, along with parameters to establish global and regional sales priorities and other optional data. The depth of the Gracenote Media Database, and pre-linking to all industry standard identifiers, lets Gracenote quickly integrate and optimize the Discover service for each particular music store catalog.
Immediate Recommendations for New Releases
Because of its multi-technique approach, Gracenote Discover can incorporate and provide recommendations of new releases. There is no requirement to build up a sales history database or perform detailed track-level editorial analysis before good-quality recommendations can be delivered.
