11/01/2009

Informatics Tells You what to Like!

Here is the first weekend blog – where I explore the other side of eCommerce, web 2.0, and social media. Sure, it's all applicable to business – we just don't have to be too serious all the time ;-)

This article immediately caught my attention, mathematically analyzing music to determine the next blockbuster. If you've spent anytime listening to Pandora, and know the background to their Music Genome Project, then Music Intelligence Solution's concept of picking winners from historic data is no surprise. While it is always nice to hold an opinion, now you know you don't even have to put too much effort into that.

The same is happening with movies. While you can get recommendations at Netflix based on how your and others ratings of previously viewed movies, jinni takes it to the next level with their Movie Genome Project. I played around with it a bit, but my Netflix queue is long enough that I'm not desperate to know what I've missed. . . not yet. Here are two good blog reviews, one positive and one critical, if you want to read more. Another note of interest, Clerk Dogs believe the human aspect of picking out the right movies cannot be broken down to metadata and algorithms. They replaced the VHS rental experience by collecting movie clerk opinions and using those to suggest what you'll like – in a more human way? Here's a good blog for more on Clerk Dogs.

What's really exciting is where this can take us. . . can one engineer songs (or movies) that they know will be winners based on algorithmic analysis of historic data? Of course, one can suggest that engineering winners has been done in the past. I immediately think of cash-in phenomenon of boy bands or the ongoing success of Pixar animation and 007 movies. But, remove the experience and expertise that goes into developing those winners, find a small, highly intelligent team of data geeks in India, Nova Scotia, or Nowhere, OK, pick the genre you want to own, and start 'manufacturing'.

In terms of logging customer behavior and running algorithms to determine how best to serve them, this is not new. But there is a lot more to be done. If you're looking to get in on the ground floor with the next metadata/algorithm based recommendation site, you can always use rovi to start off at least. . .

Nov 1, 2009 4:44:43 PM | The Other Side

The comments to this entry are closed.

NEXT POST
Channel Cost Analysis: Savings Potential of the Community Channel This blog showcases a single, easy number to help folks understand the value of the new social media environment. Consider the cost-to-serve customers through an online support community. For service platform optimization, reducing cost-to-serve may be one objective, but certainly not the only objective. It is a good starting point considering the potential impact for improving the bottom line. This blog entry is the first of a series on developing a service platform optimization model to quantify the benefits. I'm starting with channel cost analysis and specifically with the community channel - a single, easy number that projects huge cost...
PREVIOUS POST
Using Facebook to Jumpstart the Community Channel An L.A. Times article on Charles Nelson, president of Sprinkles Cupcakes, inspired this blog. Charles and his wife run a viable, if not successful cupcake business in Beverly Hills. The business has been around for 5 years with an advertising budget of $0.00. That's right kids . . . the cupcakes aren't free, but Charles' ability to connect with his customers is. Now, I want to write about connecting with the customer, not marketing. I'm here to convince you of the need to develop a service community channel – and that it's not that hard! First, let's put all doubts...