At SiriusXM and Pandora, we’ve always believed that the energy and passion that creators put into making music should be matched by our efforts to deliver their music to our listeners, whether they be long-time fans or recent converts. Beginning with our unique programs, stations, and playlists aimed at driving listener discovery, our mission has and always will be cutting through the noise to connect artists with their fans while entertaining and inspiring millions of listeners.
This commitment starts long before you hear a single note of a new song from an artist on our platform. Many assume that getting your music on a modern streaming platform is as easy as simply uploading any recording – and some platforms may allow that. But to do it well, for creators and listeners alike, the reality is more complex. And it’s that reality that is one of the most critical facets of the modern music industry – ensuring that artists and songwriters are properly matched with their songs.
From the start, our approach has been meticulous. Pandora’s Music Genome Project launched 20 years ago with human music analysis at the heart of our music discovery. Millions of songs, representing more than 95% of streams, have been through our music analysis. And while technology may have sped things up, that meticulous approach remains. Before new songs appear on a popular station or you discover it on your own, they have been subjected to extensive analysis to ensure that we play an artist’s preferred version of a song for fans, which in turn will help quickly and efficiently match creators with the royalties they deserve.
Taking this approach isn’t always easy. Many in the audio entertainment industry would prefer we quickly upload every song we receive to get it to fans immediately. But we’ve always believed in doing it the right way – taking the time to ensure songs were properly licensed, the metadata was correct, fans were getting the version of a song the creators wanted them to hear and made sense for the rightsholders.
With tens of millions of songs available instantly at the press of a button, the scale of today’s digital music catalog can feel overwhelming. But if you think that’s daunting, consider the challenge facing our music ingestion team, which faces the tens of thousands of new songs delivered each day to music services in the U.S. and deal with streaming music’s biggest challenges: inconsistent or missing metadata, duplicate songs, music spam, illegal covers and clickbait playlists meant to artificially boost streaming figures.
But this is where our investment in the right technology and commitment to a comprehensive, focused screening process allows us to deliver real music from real artists to their fans. Our human and machine curation process has helped create order amid the chaos, scrubbing the data to ensure it is complete and accurate, each version of a song is legitimate, and the content is properly licensed with organizations such as The Mechanical Licensing Collective (MLC). Our machine learning system helps identify the best version of a song by examining audio fingerprints, metadata, and human training data to tease out the different recordings of a given song and grouping them together. Once grouped, we examine a variety of factors, including the available rights, the album artist, and the album release type – Soundtrack, Deluxe, Remastered, Compilation, and so on – and rank the songs against one another to find the version we deliver to listeners.
This makes a difference. Recently, the MLC received unmatched royalties accrued by 20 digital service providers during the three-year transition period under the terms of the Music Modernization Act. Pandora’s total unmatched royalties were considerably less than many of the top streaming services, largely due to our approach that helps identify and compensate performers and songwriters more efficiently and effectively. Not that this was a surprise: our historic match rate using both machine and human analysis stands at a robust 96%. For context, the MLC’s initial machine matching rate for songs streamed in January was 80%.
This is part of our commitment and one we will continue to uphold. It ensures our millions of listeners are able to hear the songs they want to listen to, that creators and rightsholders receive the royalties they are legally entitled to, and fans and artists can continue to connect to the music they love…and the favorites they’ve yet to discover.