Mike McCready

Gartner Inc., a media analysis company, predicted in a report last year that by 2010, 25 percent of online music retail transactions will be driven by applications that allow fans to compare their tastes and by recommendation engines tracking their preferences. IDC projects the mobile music market to be turning over $49 Billion at that time. Due to the fact that mobile devices are more difficult to navigate than your typical laptop or desktop computer it is logical to assume that mobile consumers will demand that they be able to get more done in fewer clicks than they have demanded from their less-portable devices.

An essential part of getting what you want in fewer clicks is having an intelligent piece of technology running in the background that learns and knows your likes and dislikes and is able to use that information to bring to the surface the products and services you are most likely going to want to buy so that you can enjoy them sooner, with less time spent searching and digging.

In as far as music goes, I’ve often said that now that we’ve cleared the hurdle of actually getting millions of songs online and legally available the next challenge is finding which of those millions of songs any one person is actually going to want to hear and buy. That is the problem music intelligence solves and that is why music intelligence is such a hot space right now.

When I first got involved in this space, back in 2001, there was only one other company trying to do something similar. That company was Savage Beast (now Pandora Media). There are now at least 15 companies competing in this space today and each one has a slightly different angle on how to create the best music recommendation solution; but we all see the tremendous market opportunity and the wealth that will be created for those few companies destined to win this game.

In the future, recommendation engines will be a seamlessly integrated part of all our lives. On one hand we won’t even realize they are there. On the other, brand names like Platinum Blue will be as well-known as Dolby Sound is today. They will recommend products and services that are familiar enough to us to induce the feeling that they know us and “get” us. On the other hand they will also make recommendations that are new and just different enough from what we seek out on our own to make us feel like they are pushing our boundaries and exposing us to new experiences. They will grow with us and know the difference between John Q. Consumer at age 23 and John Q. Consumer at age 46. They will also know what others with tastes similar to our own have recently tried and enjoyed. This “social aspect” of recommendation will keep us sufficiently in tune with the rest of society so that we don’t isolate ourselves and so we don’t completely diverge from whatever demographic we most identify with. Social scientists say that a feeling of belonging is essential to human psychological well-being. Recommendation engines can’t ignore that.

No one company is yet doing all of these things. There are companies out there who base the recommendations they generate purely on social behavior i.e. those who like some of the music you like also like some of this other music that you might like as well if you give it a listen (I don’t think this is the best initial approach and believe that those companies will soon find themselves at a disadvantage. I’ll blog about that another time). Then there are companies who don’t use any social data to compile recommendations, like us at Platinum Blue. We use mathematical relationships to match songs someone already likes to other songs they are likely to also enjoy.

For the moment, both approaches are useful and valuable and a lot of different market factors will help companies using either one gain traction in the market. However, in the end it will take a combination of these methods to really provide the ultimate solution. Until then, we’re working to become the very best.

Trackback URI | Comments RSS