Finding a good song to listen to has never been easier…or is it?
The pool of digital content continues to swell on the Internet. Millions of songs are now available to us on-demand. For many, sorting through this vortex can be a daunting and intimidating exercise.
Developers have begun to recognize the need for better, more user-friendly tools to help us find that perfect song among the millions. Sites and applications like Pandora, iTunes Genius and even AllRovi.com have taken hold and are creating some big changes in how we find and discover new music.
These tools automate the music recommendation process, providing the consumer with personalized advice on what to listen to next. But are these recommendations as good as old-fashioned word-of-mouth? Are they better? We sat down with Michael Papish, Rovi product development director and expert in the field of digital music recommendations to get some insight.
Have we reached a golden age of music recommendations?
People have always been the primary vehicle by which music spread around the world – from the days of the wandering minstrels before music was recorded to the heyday of radio with DJs like Wolfman Jack and Vin Scelsa, and then on to hometown record stores and tight-knit circles of fans rallying together around photocopied ‘zines and basement venues.
However, over the past 10 years we’ve entered a new phase of music discovery – one powered by, like many facets of life these days, the wizardry of the internet. The internet has amplified the inherently social aspects of recommendation giving. For example, people can unite around the globe around any genre or musical interest with a few clicks of a button. This new type of musical exploration has no geographical boundaries the way your hometown music circle once did.
The internet has also allowed the birth of music recommendation engines and tools that help to automate the music recommendations process for users, making it that much easier to come in contact with new music. That said, we still have a way to go until this process and most importantly the consumer experience around internet-based recommendations is perfected. So while it may not be the golden age yet, it is certainly an exciting time for entertainment and how we share it with others.
What challenges still exist with internet-based recommendations?
It’s important to remember that not all listeners are looking for the same experience every time. Sometimes you might want to sit back, relax and let the music come to you, so a passive tool like Pandora would provide a good experience.
Other times you may be willing to put in extra effort to find that perfect song. In this instance you may turn to a tool like the Rovi Finder application on AllRovi.com, which requires more active input from the listener.
My point is, you can’t take a one-size-fits-all approach with building out music recommendations. Rovi is looking at ways to broaden the options for users, creating the right mix of experiences based on listener preference at any given time of the day and on different platforms from the TV, PC, smartphone or tablet. Each scenario should evoke a different experience for the user.
Are music recommendations an art or a science?
Good question.
Science is crucial, specifically because the space of music is so large (10M+ tracks). Well-determined methods are needed to figure out relationships between all that music. This can be considered the foundation for today’s recommendation engines. However, to make really good recommendations you must embrace the fact that recommendations are also very much an art.
I’ll argue that over the past 10 years, as an industry we’ve reached the stage where we understand how to eliminate the blatantly wrong recommendations. Our next challenge is to figure out how to make great recommendations and (as importantly) how to accurately measure our progress towards this goal. Critical to this argument is knowing that the goal of today’s music recommendations engines is not necessarily to make you click and buy more, but it’s to build trust between the listener and the system/tool/company making the recommendations – this is where artistry becomes critical.
Take the engine interface for example, the computer scientist responsible for building this interface must know how to code (science), but they also need to have a keen understanding of how people think or what motivates them to share information (art). They must then carefully blend this knowledge together to build a site or application that draws the listener in, engages them and builds that level of trust.
At Rovi, we are constantly looking for ways to perfect this delicate balance between the scientific and artistic elements that comprise an awesome recommendation – from the search algorithm down to the staff-written editorial reviews – each piece of the puzzle is essential.
Rovi’s Top SXSW Panel Picks for What’s Hot in Connected Entertainment
Rovi will be at SXSW 2012 and we’re looking forward to checking out the discussions and panels taking place about the future of connected entertainment. Since we happen to be sorting through the SXSW Panel Picker this week to pick out and vote for our favorite panels, we thought we’d share our top picks for the sessions we’re interested in seeing make the final cut at SXSW – including a few shameless plugs for our own Rovi panel proposals. If you’re in the business of entertainment technology, we think you’ll enjoy these sessions too.
Voting ends this Friday, Sept. 2, so get to it and get your vote on. See you at SXSW 2012!
(Speaker: Richard Bullwinkle, Rovi)
(Speaker: Michael Papish, Rovi)
(Speaker: Tom Erlewine, Rovi)
(Speaker: Jeremy Toeman, Dijit Media)
(Speaker: Oren Nauman, AnyClip)
(Speakers: Carlton Cuse, Carlton Cuse Productions; Brad Pelo, i.TV; Lisa Hsia, Bravo Digital Media; Alex Iskold, GetGlue; Bill Gannon, Entertainment Weekly)
(Speakers: Mark Rivis, TiVo; Billy Hogan, MLB; Matt Graves, Twitter; Nick Johnson, NBC; Mitch Oscar, MPG)
(Speakers: Michael Aragon, Sony Network Entertainment; Jason Spivak, Sony Pictures Home Entertainment)
(Speakers: Stephen Bradley, Farmingville Capital; Alan McGlade, MediaNet)
(Speakers: Jim Cady, Slacker)