Can IBM Help You Write a Better Blog Post?Mathew Ingram
If you have a personal blog, or even a corporate one that you help write, you've undoubtedly run into it: the Wall. Also known as "writer's block," it's the inability to come up with something to write about, or a lack of enthusiasm for doing so. Well, researchers at IBM think they may have come up with a way to get past the Wall, with what they're calling the Blog Muse—a kind of social recommendation system for blog posts in which users say what they want to read about, other users then vote on those suggestions, and the most popular topics get distributed to those most likely to want to write about them.
Casey Dugan and Werner Geyer of IBM Research started working on the problem a couple of years ago, after finding "blogger's block" or "blog fatigue" to be one of the leading problems for both internal social networks at IBM and some of the company's clients. "We asked what they had trouble getting people to contribute content to, and they said blogs," Dugan said. "They said that people would start them and then stop writing, or that there were a lot of blogs with not much on them." In fact, IBM research shows that about 80% of those who begin a corporate blog never post more than five entries.
Geyer said many companies see the value of blogs because they allow people to share information, but that "often people stop blogging because they don't get any attention with what they're writing, no one comments on their blogs, they don't know what to write, and so on." As the researchers describe in their paper:
"In order to inspire bloggers, our system suggests topics they can write about. The audience is given a voice by letting blog readers share topics they would like to read about with the blogging community. Our system then suggests these topics to potential blog writers, who can decide whether or not they would like to address the topic requested."
Tweaking With Widgets
The system consists of two simple widgets added to BlogCentral, the internal blog network at IBM, which was launched in 2003 and has since seen a total of more than 145,000 blog posts written on 16,000 different blogs by more than 14,000 users. One widget allows users to suggest topics they'd like to see written about, while the other allows them to vote on recommendations from others. In testing the system, using the profiles created by users in SocialBlue (IBM's version of Facebook), it found up to 50 users who might be interested in writing about those topics and sent them the recommendations. When a post was written, it sent the post to anyone who voted for the topic.
And what did the research show? According to Geyer, in a study of 1,000 users who tried Blog Muse, "blog posts created from our system got twice as many comments and got more views as well, and they got three times as many stars (or likes)." Interestingly enough, Dugan writes that "we didn't find an increase in the number of blog posts, so maybe there was some substitution going on there—maybe people didn't write more, but the ones they wrote got more readers." There was also some resistance from blog writers who wanted to follow their own muses, rather than playing to the crowd, As Dugan put it:
"Some described how they already have topics they write about, are without a shortage of ideas, and find blogging a 'personal' activity that suggestions might infringe upon. One went as far as saying, 'This would be similar to writing paid reviews for consumer products.'"
Among bloggers who didn't write as frequently, however, there was support for the system because it helped them come up with ideas. The researchers said in their report that their goal was "to inspire users to write more blog posts, and our approach is to involve readers by allowing them to share their topics of interest with the blogging community. Sharing and voting on topics adds a new communication channel to the blogging ecosystem." IBM says it's planning to roll out Blog Muse internally—and may look at commercializing it at some point in the future.
A number of blogosphere recommendation systems do something similar to Blog Muse—arguably, topic filters such as Techmeme and Tweetmeme perform the same kind of function by letting bloggers know what topics are getting the most attention from readers. But does this remove some of the serendipity that can make blogging so powerful? We don't always know in advance what we want to read about or what will move us. What do you think?
Also from the GigaOM network: