Upcoming Idealab post by Andrew Grice


Hi folks. My name is Andrew Grice and I'll be blogging in here to help my good friend and colleague get some important things done. Benjamin Melançon is a workaholic workhorse. I’ve been amazed at all he’s accomplished over the past year; such as founding a successful new business, landing a grant to develop his “Related Content” module and guiding People Who Give a Damn through the maze of tax law to achieve 501(c)(3) nonprofit tax status. So when he said he needed time to concentrate on writing the actual code for his “Related Content” module, I told him I’d be glad to help. Considering the outstanding brilliance of so many here in the Idea Lab, I feel both honored and humbled to write here. I can only hope you’ll find some value in my posts.

So what is “Related Content?” Well, let me answer that first by telling you something about Benjamin. Benjamin Melançon is a young man who has a truly great idea with the promise to one day revolutionize journalism and help re-democratize political culture world wide. Related Content is not that idea. But that said, Related Content is a excellent idea in itself. Specifically it’s an idea to incrementally make the content management system Drupal an even better tool for online journalism sites.

Whether or not we’ve ever given any thought to it, we’ve all seen on journalism sites where an article is followed or otherwise displayed along with a short list of links to related articles or something similar. It’s a natural idea for keeping readers reading articles on that particular site. But how are these short lists of supposedly related links generated? Who makes those lists? How do they do it? And how much time and energy to they waste doing it? Soon the Related Content module will provide the best answers for those questions, but first let’s consider for a moment how things can be done right now.

Obviously one way to do it would be to pay people to do it all by hand. Sounds expensive, doesn’t it? Well thankfully there are already software solutions for easily adding a short list of links to any given article on a journalism site. Of course nobody using modern tools has to enter raw html line by line anymore. And there are already a number of software solutions to the coding part of displaying lists of related stories. That saves time and money.

But where do the links come from? Someone or some thing has to select the links. And the more appropriate those related links are, the more effect this feature will be at keeping readers reading more pages on any given journalism site. Hiring an individual to handpick great links can be too expensive. Tweaking a search algorithm can yield useful results. But any automated search algorithm will also bring up garbage along with the good links. The better the tweaked algorithm, the less garbage. But do we really want any garbage links in our related stories list?

Another solution is to use tags. For example any story about George W. Bush would get a Bush tag. One story that made headlines this week dealt with President Bush expressing something akin to regret over his cowboy rhetoric during the run up to the invasion of Iraq. That story gets a Bush tag and an Iraq tag. But just using those tags, what do I get for my related stories? Iraq tends to get at least a mention in passing in just about every article about Bush these days. Therefore the supposedly related stories could easily be any random article that mentions President Bush. One might be a discussion of Republican chances in the Presidential election. Another a story about the New York Giants football team visiting the White House for the traditional photo op with the superbowl winner. Another could be a puff piece on one one of the Bush daughters. Yet another might be a piece on Hilary Clinton’s latest plea for power. That’s not very impressive at all if those are the stories the tag system dishes out automatically as “related stories.” It’s certainly not the level of quality that would keep readers engaged on a newspaper’s website.

This is where the real beauty of Ben Melançon’s Related Content module comes in. Related Content will make it incredibly easy for anyone to intelligently add links to these related stories lists themselves. And once readers start viewing and following those lists of related stories they will be able to easily provide feedback on whether those links really are related or not. Newspapers would be able to start with an automatically generated list of links using whatever combination of tags and search engines they like. Those would be the initial related links. Soon after that, readers would quickly vote out garbage links and also add their own contributions from the newspaper’s archives which will surely prove to be vastly more on topic and on target than anything any automated system will generate.

In summation, this Related Content module will be good for journalism sites because it will make it easier to keep readers on the site reading more articles. And it’s going to be good for readers as well because it will make it easier for them to find articles that provide the contextual information they seek.


Syndicate content