Thoora Shows How Publishers Can Use Real-Time Audience Data for Editorial Decisions

February 12, 2010
Category: Uncategorized

To the list of companies that say they measure audience sentiment to help publishers’ editorial judgment, add the name Thoora.

The Toronto-based startup promises to gauge how well individual news stories are doing by analyzing and calibrating real-time data from blogs, mainstream news sources and Twitter. Thoora’s software uses more than 100 attributes to determine not only the most popular content but also the highest quality, using measures such grammar and spelling and the authority of sites that link to the content.

The company said the data could be used to figure out, for example, where to position an article on a page (aiding internal data from Web logs and analytics), how to apportion resources to cover a developing story or even how to follow up on offshoots that you might not have considered. It could help a news organization determine where its individual story ranks against competitors covering the same thing.

CEO Mike Lee said this is the first time that a tool has approached audience sentiment for news at the story level rather than the topic level. For example, after Serena Williams lashed out at a line judge at the U.S. Open, Thoora spotted a tennis pro in upstate New York writing quality stories about tennis rules — stories that ranked in quality and authority with any of the larger news services and covered the topic in a way they hadn’t.

“We notice there are disparities between how specific stories are dealt with, interpreted and continue to drive reaction and conversation, where it may have dropped out of the news cycle itself,” Lee said. Thoora could see that after President Obama mentioned funding of high-speed rail in his State of the Union speech, conversation persisted awhile after traditional news coverage had trailed off.

Thoora, chosen last September as one of the TechCrunch 50, is part of a growing trend of attempts to measure audience sentiment, though Lee said his company is the only one that focuses strictly on news. Demand Media has gotten a lot of coverage for using algorithms to gauge popular search terms and matching them with evergreen stories such as “how to” articles. AOL is basing a large part of its editorial strategy on a demand algorithm for its dozens of sites. Aggregators like DayLife are pushing news stories to the fore based on measures of what’s most popular or important.

But, Lee said, many aggregators push to the top whichever stories are about the most popular and current topics, using recency and volume as a proxy for what is valuable and worthy of attention. “We hope to not just be a quantity aggregator but to actually drive quality to the surface,” Lee said.

At the recent OnMedia conference in New York, Thoora showed off Web-based charts and graphs representing its real-time data. Lee said Thoora uses subject-matter experts to vet and hone its computer selections, much as the site Pandora does with its Music Genome Project to group songs with similar characteristics.

Thoora does not yet have any clients, but Lee, a co-founder of the company, said it is in talks with a major Canadian news organization and a Canadian sports publisher. After the TechCrunch 50 presentation, Lee said, the firm got many publisher inquiries and decided to position itself as an enterprise solution, offering data to publishers rather than focusing primarily on the consumer news aggregation site at

The consumer site allows people to browse stories based on similar algorithms. Lee says the company may at some point release its APIs or a free version of its platform to enable people to input their stories and see how they match up against others.

While it’s impossible to say if Thoora, which is backed in part by the Canadian publisher Rogers, will succeed, I believe that editors and publishers need to be increasingly comfortable with using all the data at their disposal to make editorial and business decisions.

Journalists are finding sources and story ideas through Twitter and Facebook. Perhaps their editors can use data from those networks and others to help decide whether and how to position a story on a home page, even as Web analytics may indicate a different action based largely on page views.

To those in the industry who would complain that the ever-shifting sentiment of the crowd is replacing editorial judgment, I would say they should look at these as an adjunct to our judgment. The better we can learn to use the data, the more we enhance our input into decision-making.

News organizations that are in a fight for their lives are going to have to use every tool to attract and hold audiences that are enticed — via their own RSS readers, friends’ recommendations via social networks and guidance from aggregators like Google News — to rely on something other than editors’ picks.

No longer do geographic or distribution boundaries allow news organizations to offer the same story as everyone else and expect users to remain loyal. In another era, “people could not pick up a Denver Post so easily” if they weren’t in the Denver area, Lee said.

“Now you have to spend a lot of time thinking of how different you are,” he said, “and you have to communicate to your customers why they should come to you rather than XYZ news outlet or feed. But it’s impossible to do that unless you start to understand the underlying data of how you are measured and used relative to others.”

Or put it this way: Editors and publishers must figure out how to provide something for their products that major consumer brands have had to grapple with for years: differentiation. How are my automobiles, my sodas, my pairs of pants — or my individual news stories — any different, better and more satisfying than anyone else’s?