If you are worried about robots writing the news that goes with your morning coffee, you can take a deep breath. Jeff Burkett, senior director of operations at The Washington Post, assured his audience at at the Big Data Media Conference, a joint venture of World Newsmedia Network (WNMN) and INMA, that they are not going to let that happen any time soon.
However, the Post has pioneered robot-like programmes that brilliantly measure the impact that human journalists have on human readers. These programmes were born out of the Post’s need to go beyond the unique visitor metrics — which Burkett said are not a valid measure of audience satisfaction — and look instead at time spent per article.
“How deep are people going?” is a more pressing question for Jeff Bezos and his team at The Washington Post.
Another change that the Amazon CEO made upon buying the Post three years ago: Zooming out to focus on the rest of the world.
“Before Jeff bought us, we were all about [Washington] D.C.,” Burkett said. “You have to do more than that.”
Here are some tools and methods that Burkett said have helped The Washington Post meet their ideology- and number-based goals.
Loxodo: The Post has its own dashboard that provides metrics based on each department’s needs — a customisation that is not offered by places like Google Analytics or Chartbeat. In addition, Besos has emphasised the need for the measures of right now, instead of the commonly acknowledged lag measures or measures of the past. Burkett has called for his team to focus on how they can change their right now, to impact lag measures of the future.
Websked: A tool that measures the freshness of content on the Post site. The digital world needs and thrives on that freshness of content. The tool also helps give reporters deadlines, of which they are regularly reminded by a Marty Baron-faced chatbot.
Break Fast: Fresh content also comes in the form of news alerts, which is what this tool measures. It ingests Post breaking news alerts, alongside those of competitors such as USAToday and The Wall Street Journal. The metric: Who is getting alerts out the fastest in terms of timeliness, coverage and spam-rate? Based on these, the Post makes technological and human resource improvements.
How interesting is content? The Post gathers their content and content from competitors, strips the brand names and asks samples of people which story interests them most.
Bandito: A “content testing multi variant,” or more simply, it is a live and self-updating tool that looks at headlines and images that drive the most engagement.
Clavis: Inspired by Amazon’s recommendation engine, this tool seeks to look at what people are reading to determine what they are interested in, thereby personalizing their experience. Being a local and global site, The Post has found location based recommendations to be key, Burkett said.
Native advertising content: It is kept separate from the newsroom, Burkett added. This, along with targeting, has led to an increased engagement in paid content — and more happy advertisers.
Another Bezos change: “Today we have one thing — and that’s to grow the audience.”