• by Ben Plomion , Op-Ed Contributor, September 9, 2013

After his surprise purchase of The Washington Post in August, Jeff Bezos hasn't yet said much about how he will change the paper. But there are a few things we can be fairly certain about. One is that Bezos won’t put too much energy into sustaining the print edition of the paper. Last year he told a German newspaper that there won’t even be printed newspapers in twenty years. The other thing we can be fairly certain about, based on Bezos’s record at Amazon, is that there will be lots of digital innovation at The Washington Post in the years to come.

So, what, specifically, will Bezos change at The Washington Post? For now, as usual, he isn’t saying very much. But, based on his record at Amazon, there’s every reason to think that Jeff Bezos will turn The Washington Post into a uniquely programmatic publication.

Programmatic, in this sense, means using data for online targeting. Programmatic marketing has revolutionized display advertising in recent years by making it possible to deliver ads to specific users based on everything the from the sites they have visited to the time of day that they’re browsing, to the searches they have performed on Google, Yahoo, Bing and other sites. Programmatic targeting is essentially the same technology that Amazon uses when it recommends other items you might like to purchase on Amazon.com. In other words, it’s the same technology that Bezos himself turned mainstream.

There’s no reason, of course, that programmatic targeting has to be limited to ads and product suggestions. Any piece of content can be targeted so long as you have the data. But at the moment, most news sites still rely on non-programmatic targeting. To get content that’s relevant to them, users first have to create an account and then tell a site the topics they want to read about. The closest that most publications today come to a programmatic approach is in the related content links, which are usually based only on keywords in the article the user is currently reading.

The problem with the non-programmatic approach -- beyond the fact that most of us never bother to create accounts, let alone set our preferences -- is that we’re not always entirely aware of what we really like to do online. With programmatic targeting, a content site can deliver content that we really enjoy by learning from our browsing habits across the Web. And there's no need to create an account or tell a site what you want to read.

Imagine, for example, someone who claims not to be interested in art and yet is twice as likely to click on an article if it has “art” in the title. A programmatic newspaper would notice a pattern like this and deliver more art news to the user. Or imagine that you stay on a page twice as long when a certain athlete is mentioned in an article. For a programmatic paper, that’s valuable data that can be used to serve you content about that athlete. Likewise, if you search a medical concept on Google, a programmatic paper can start to include articles on the topic on its home page or deliver them to you in your personalized emails.

And that's just the tip of the iceberg. A truly programmatic publication would be constantly adjusting content according to where you are, the time of day, and many other factors.

The limitation with content personalization, of course, is that not even Bezos can afford to have writers tailor future articles for each individual. But what Bezos can and will do is uncover new patterns in the way we interact with the news. And you can bet that The Washington Post will change the way it produces and delivers the news based on what this data reveals. That will mean more user engagement and and more revenue -- from programmatic ads, of course.