Content Decoded

Ideabox is a new feature on ContentDecoded. Every IdeaBox author is charged with coming up with an idea that solves for a current challenge in the business of content. Our first Ideabox column is authord by Rick Liebling.

Liebling is the Global Director of Client Management at  Taylor, a marketing communications agency. In this column, he talks about the possibility of a Pandora for Content. Read on!

I think one of the bigger trends of the last few years has been increasing ability for consumers to customize the content they receive. At first blush, this seems like a good thing. Being able to pull only the relevant content you want sounds great when there is an overwhelming amount of content out there.  This sort of filtering can work on both qualitative and quantitative levels. But what is often overlooked is what is lost with this sort of filtering, the serendipitous discovery of content.

When I flip through a magazine or newspaper invariably my eye will come across an article that piques my interest. Not something that I’m necessarily a huge fan of, but something close enough that my curiosity has now gotten the best of me. Let me give you an example. I remember seeing a review or interview or some sort of mention of the novel Netherland when it was first published. While critically acclaimed, this wasn’t the type of book that was going to generate Dan Brown-type hype. If I don’t flip by this story in the paper, I’m probably never going to know of its existence. I’ll come back to Netherland in a moment.

This idea of serendipity is important, especially to people working in creative industries. We need a wide base of knowledge and experiences, yet by subscribing to certain feeds, email alerts and newsletters we tend to shrink our knowledge base. But what if, rather than narrow our sources, technology could be used to widen them?

Pandora is a fantastic application for music lovers that helps them discover new music. Not completely alien music, but rather music that is somehow similar to music users already like. So, if say you are a fan of Belle & Sebastian, Pandora will recommend Kruder & Dorfmeister or perhaps Stereo MC’s. That’s cool, I like both of those groups too. But it will also give me Sofa Surfers, a group I’ve never heard of. But I feel comfortable checking them out because there’s a pretty good chance they’ll be in the same ballpark.

Now, let’s take this same concept and apply it to content, revisiting my Netherland example.

Imagine a web-based application that pulls in content from all over the Internet. Blogs, news sites, user-generated content, you name it. As long as it meets certain parameters (length, Fry Readability Formula, proper grammar, etc.) it goes into the system. Now, as you start to pick and choose content, the system starts to learn your preferences. Not only the subject matter, but also the type of content – interviews rather than mere profiles; hard news rather than opinion.

After a little while, the system learns that I like book reviews. It also knows that I read articles about Dutch culture team and that I have some interest in cricket. From there it starts making suggestions based on these preferences until I see a recommendation for a novel about a Dutch guy who plays cricket on Staten Island.

Sure, there will be some ‘false positives’ but even that won’t be a bad thing as the point is almost to create ‘false positives.’  Ideally you’d click on one of these ‘false positives’ and that would go into your database, ever so slightly broadening your content pool. There would be real benefits to content providers buying in to this system, especially those that produce unique, original content but have small user bases.

I would imagine there would be easy advertising opportunities for this as well. Again looking at the Pandora model where little modules slight from right to left, wedged in between the songs. Contextually relevant brands (or media properties) could have small, clickable ads worked directly into the interface.

I’m not a backend development guy, but this seems like a pretty doable application. I’d love to here opinions on both the UX viability as well as the development challenges.

This entry was posted on Wednesday, October 28th, 2009 at 3:13 pm and is filed under IdeaBox. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.