hack text

In which I don't bother to asssign a subtitle.

A recent presentation on the semantic web API called Amplify really put into focus the issue of targeted advertising online and how to solve the recent decreases in display ads for newspaper websites.

I went to a Social Networking and Semantic Web meetup on May 13. The event, which was the Roy Rosenzweig Forum on Technology & the Humanities, was run in concert with George Mason University’s Center for History and New Media.

There were a number of great products and ideas put forward. The group of presenters covered how to make Recovery.gov more “citizen-friendly,” Zotero, and Glue. The one presentation that really got me thinking, however, was Amplify.

In the demo at the Social Networking and Semantic Web meetup, Amplify’s CIO Mike Petit showcased a quick example with a paragraph written by someone who hates baseball. Amplify analyzed the paragraph and ranked topics by importance and than took those topics and analyzed their sentiment. As a result, it could tell that baseball, despite many occurrences, was a negative topic.

This sort of technology could be the key to saving newspapers online.

One of the big issues that all newspapers are facing, from the college level all the way to The New York Times, is how to make their websites profitable. The big stumbling block on the path to a self-sustaining website is the reluctance of advertisers to buy online banner and button ads from media companies.

At the end of 2008, Gawker Media Ring’s Nick Denton noticed the trend. Predicting (accurately) that making a profit off of being an online news source was about to become much harder. His very first point in his 2009 internet media business plan was “Get out of categories such as politics to which advertisers are averse.”

This is missing the point. It wasn’t that advertisers are adverse to divisive politics or product reviews (Wonkette and Consumerist, two of the blogs that Gawker shed, covered these categories), it’s that covering sensitive issues means that an advertiser’s ad is more likely to show up next to something critical of the advertiser.

The Consumerist is a great example. Wal-Mart may want to reach that audience, but it can’t do so if its ad is sitting next to an article on how Wal-Mart might be the worst company in America. It’s the same for Wonkette.

The problem is that newspapers are well known for their criticisms and advertisers don’t want to take the chance of having an ad show up next to a bad review. As a result, with the economy down and companies looking to decrease risk, no one wants to chance a Kayne and Lynch incident. This is a special concern in hyperlocal ventures, where your advertisers may often be the people whom you are covering.

That’s where Amplify comes in.  In the example shown, most semantic ad services would see the numerous occurrences of the word “baseball” and shunt an ad about baseball on to the page, exactly the opposite of what you’d want to do. Amplify, however, understood that a baseball ad was the opposite of what you’d want.

This sort of brand protection may be exactly the thing that newspapers need to bring confidence back to their advertisers and survive online. If we’re looking for a way to save journalism, Amplify and similar technology may be the key.

In case you couldn’t tell, their presentation was very impressive.