The Future of News Filtering: By Credibility?

In May, my former colleague Corvida predicted that filters are the logical next step for social media. Social media sites, such as Digg and Reddit and Delicious and Techmeme, are really filters themselves — filtering news and information based on what your readers and other writers find valuable and interesting. But social media sites have themselves become part of the problem of information overload, sending so much information our way, that as Corvida predicts, we’ll soon need filtering for our filtering.

But one thing is clear, at least in my mind: filtering is the future of news. Y Combinator founder Paul Graham recently called news filtering sites like Digg and Reddit “just the beginning” of the next evolution of news delivery web sites. Graham challenged web innovators to come up with a new type of news site that will push news dissemination forward. “News will morph significantly in the more competitive environment of the web,” he wrote.

Two sites think that the next wave of news filtering will be based less on popularity, and more on measuring credibility. The problem many users have with social media sites like Digg and Delicious is that they reward popularity rather than necessarily highlighting the best content. Two startups NewsCred, which is supposed to launch tomorrow, and NewsTrust, which is actually a few years old, aim to sort news based on the credibility and reliability of the source rather than its popularity.

Both sites work essentially the same way — users rate news based whether or not they think that news source is credible and reliable. This is different from other social media sites, where news is voted up or down based on whatever criteria people use personally — it could be credibility or it could be “coolness” or because the voter’s friend wrote the story.

“We firmly believe that a transparent and credible media industry is vital to the health of the world democracy,” says the NewsCred about page. “We’ve also done our research, and there are hundreds of surveys and papers that show that the public continues to fault news organizations (both traditional and online) for a number of perceived failures, with solid majorities criticizing them for political bias, inaccuracy and failing to acknowledge mistakes.”

Each site works basically the same, but NewsTrust has a much more comprehensive voting mechanism. While NewsCred opts for the simple approach — every article fed into the site from pre-approved top news and blog sources can be either ‘credited’ or ‘discredited’ and the site then figures out a ‘credibility’ score for that article — NewsTrust takes a more extensive approach. On NewsTrust, stories are rated on one to five star scale in 11 categories, including fairness, balance, context, importance, style, and trust. The individual ratings are combined and averaged for an overall rating on the story that, in theory, reflects all of those factors.

On both sites users are allowed to comment on stories and add additional information and sources. NewsCred also lets users customize which sources are given priority when creating the main page, giving users an extra set of filters to cut down on the flow of information.

Of the two sites, NewsCred is certainly the slickest designed with the lowest barrier to entry for users — the simple credit/discredit voting system is easy to pick up, especially for converts from other social media sites which have similar voting systems (Digg, Reddit, Mixx, etc.), and the site’s design mimics mainstream news outlets. But NewsTrust probably has the better system. NewsTrust specifically asks users to weigh in on articles according to specific criteria, so they’re much more likely to get accurate results than is NewsCred — whose system could be treated by users as no different than Reddit’s up/down voting system. Even though NewsCred asks people to consider things like credibility, accuracy, bias, and quality when rating stories, it never defines those terms. What I may think is quality, you may think is rubbish.

That same problem of personal definitions might affect NewsTrust as well, but probably to a lesser extent because the site clearly defines the things that it wants people to consider when rating stories. And its more granular rating system is also superior — I can see that a story might be, for example, considered by most readers to be well-written and interesting but not particularly fair. That might mean it is worth reading, but on NewsCred, the story’s tendency toward bias might sink it.

In general, NewsCred and NewsTrust represent an interesting new approach to social media sharing and could yield a better way to cut the noise from the signal than traditional up/down news voting sites.