Digging Data to Grow Audience

Sitting in a statistics class at Stanford taught by Chris Wiggins, the chief data scientist for The New York Times. Attending an exclusive big data conference along with presidents and vice presidents of some of the largest media companies in the world.

Is this what journalism has come to?

Yes, if you’re a JSK Fellow researching how news organizations can use data differently to tell better stories and grow audience.

It’s a life that all journalists will soon know, but one I became intimately familiar with in the past year.

I started the fellowship focused on how publishers currently use data and technology, and how we should be using it. Wiggins was one of the first experts I reached out to. I’d been referred to him by Tony Haile, the former CEO of Chartbeat, with whom I talked about audience analytics over what passes for fried chicken and collard greens in California. The chicken was OK, it just didn’t taste like Southern fried chicken, if you know what I mean. Still, Tony said you have to talk with Chris. I said sure, and asked Tony if he had his direct number? He did and he gave it to me.

I called and Wiggins answered. We talked about how the Times is using predictive analytics to identify customers who are at risk of canceling and to optimize single copy sales. But he also told me about BlossomBot, an automated tool that the one member of his team embedded in the newsroom invented to help editors decide what content to post on social media. When Wiggins came to teach a class at Stanford, we talked about the most significant challenge facing audience growth in the journalism industry — newsroom culture. Hearing him say that confirmed every bit of what I’ve been doing, not only this past year at Stanford, but over my entire career.

We can’t all be Amazon. So what?

I started out on the business side of newspapers, working in circulation and advertising. I thought I’d be able to bring over some of that expertise to the newsroom. At the time, unfortunately, editors wanted little or nothing to do with the business side. At the Times, the division between church (the newsroom) and state (the business side) is still very strong. Wiggins’ team is located on the state side of the news organization. Fortunately, he reports directly to the CEO. That, and a recent change in newsroom management, has allowed for the data analytics team to empower more newsroom decisions.

While the Times is still navigating the culture issue, one news organization that is having much more success at changing culture is The Washington Post. I first truly recognized that something different was happening at the Post in October 2015 when I noticed how many millennials were clicking on — and discussing — a Washington Post story about Zola’s adventures in sex trafficking. (If you don’t know the story, then click and read more about it.) The Post went beyond the typical click bait that every other online news site published, and it paid off in spades both in terms of traffic, but also attention, the metric advertisers care about most.

Many of the users coming to the Zola story were considered “new” because they clicked on The Washington Post fewer than three times a month, or less. They are casual users of the Post and most any other mainstream news product. And they are part of a consumer group coveted by marketers. I wanted to know how we could use data and storytelling to get these users to return to the Post more often, and how we could marry data and storytelling to get them to register for the site, subscribe, or even click on a targeted ad. In short, this is really when my challenge became how to convert casual users into more loyal news consumers.

Again, thanks to being a JSK fellow, I had one of the most important experiences of the year when I met and talked with Jeff Burkett, senior director of product and operations for The Washington Post. Many already know that Jeff Bezos’ Amazonian influence is taking hold at the 141-year-old newspaper. I not only learned more about the software the Post is developing to help build user profiles of casual users, but got to hear about the culture change taking place there.

Among the new software the Post is using, there’s Bandito, which optimizes content for clicks; Riveting, which lets readers tell Post editors how likely they are to read another piece of content, whether that content was published by the Post or not; and Clavis, a state-of-the-art audience targeting technology that comes from Amazon’s recommendation engine. The Post is using these tools to also create its own user profiles. The Post has other proprietary software, too, that it plans to eventually sell to other media companies. The Post decided not to rent somebody else’s technology, but to build its own.

The Post is using data to put the focus on users, not clicks, Burkett told me. “It’s (data) expensive and costly. But the payoff happens when you do something with it.”

But building your own software tools isn’t the only way to reach new users and grow audience.

Confirmation & Culture

The problem inside many news organizations is that they are so hyper-focused on trying to figure out who their existing users are that they totally ignore potential users. Yet keeping existing users and grabbing new users is the only way to grow. In order to succeed, news organizations must do both of these.

Before buying a bunch of bling, or developing new technology, news organizations must first use the data they currently collect in different ways. Instead of counting clicks, find out who is clicking, what content they clicked on last, and what content they may click on next. Learn about what these users care about, their interests, not just their age and gender. The news industry already has the tools at hand to do all of this; publishers just need to change the way they’ve always done things.  

The payoff comes when you do something different with the data.

At the end of this fellowship I will produce a set of best practices and, hopefully, begin conducting a case study on using data to personalize the news.

From data scientists and analysts to leading economists who shared knowledge about advertising, monetization and platform competition in digital markets, my life as a JSK Fellow granted access to some of the top tier experts in audience analytics. From them I learned much, but these interactions also underscored something I’ve always known: It doesn’t matter how much data or technology news organizations have. If media content isn’t reflective or representative of users and their interests, those news organizations are doomed to fail.