InboundWriter: Content App Uses Data to Predict What People Want to Read

Usually writers and editors depend on their wealth of knowledge and instincts when it comes to picking what to write about. InboundWriter, however, believes that the fine art of story selection can be made more accurate with its special algorithms that mine social media data and other Web traffic analytics.InboundWriter's paid platform, which launched this month after 18 months of development, allows journalists to input topics. Then its propriety formulas figure out which topics would drive the most traffic, assigning each term a green, yellow or red score depending on how successful it thinks it would be in getting people to click on the story. Theoretically, this would cut down the time writers spend on stories no one cares about, allowing them to focus on topics that people want to read. "What we find is that most content doesn't perform," InboundWriter CEO Skip Besthoff explained. "For every 10 pieces of content that [are] published online, maybe one or two will break out. That means eight out of 10 [are] going to disappoint."Publishers like Time Warner Cable, The Enthusiast Network (which includes Motor Trend and Hot Rod), Skilled Up, TheUnlockr.com and GottaBeMobile jumped on board during the beta testing period. According to InboundWriter's trial with 20 pilot programs, its predictive formulas were able to garner between two to four times more traffic for stories it denoted with a green label than an overall comparison group. One case study based on a medium-sized blog that publishes 300 articles a month found that green topics performed two times better than InboundWriter's control group, and they were upward of three times more successful than topics in the yellow and red groups."We've seeing a significant increase in traffic, and it's helping us develop the headline and the supporting keywords that go into that," GottaBeMobile managing editor Josh Smith confirmed.Admittedly, it's not a perfect system. Besthoff pointed out that the algorithms reflect traffic derived from online searches more than from social media, although the company hopes to fine-tune its platform to predict those factors. In addition, the company is working on a way to anticipate the best-phrased headline for maximum traffic.Also, while the algorithms use real-time data, they're only customized for "evergreen" subjects, or topics that don't revolve around breaking or timely news. For example, InboundWriter would be able to predict whether a story about student loan costs for a personal finance blog would do well, but it wouldn't be able to tell whether or not it's too late to write about today's interest rate change.And, the service doesn't account for the journalist's own instinct."There are still times where we run with our gut," Smith admitted. "I'm still going to go with those [instincts]."