Optimizing Journalism for Trust

Jay Rosen
The Correspondent
Published in
6 min readApr 14, 2018


In a book called Politics and Vision, the philosopher Sheldon Wolin said that when there is vision, “things appear in their corrected fullness.” This helps explain what I mean by optimizing for trust. It is a vision or direction in which we can move.

For trust can no longer be assumed. Its continuous production has to be designed in. Nor does trust any longer follow from good practice, which is what American journalists used to mean by the term credibility. Once upon a time, you “had” credibility if you followed the rules of good practice. That doesn’t work anymore.

We have to design the modern news organization so that it is easier for people to trust it. (Which of course doesn’t guarantee that they will.) We might even say that trust has to become more agile.

And we have to make it easier for people to form a relationship with the sites whose work they value. Otherwise Facebook and Google and maybe Apple News will own that relationship.

It was Aron Pilhofer of Temple University (formerly of The Guardian and the New York Times) who got me thinking in this direction. In 2016 he asked: What would a news organization look like if were optimized, not for clicks… or for scoops… or for time-on-site… but for trust?

I thought this was a good question. So I started talking about it on social media.

Emily Bell did not agree with me. She said the question was badly framed. She said that trust was a “poor metric” for quality journalism. “Breitbart optimizes for trust,” she said. “So does the Daily Mail.”

At first, I didn’t understand this objection. Certainly I didn’t trust Breitbart. Nor did I think it was built for trust. And yet I had to admit: its core supporters did trust it. Breitbart was optimized for them. So in a way it was optimized for trust.

In wrestling with Emily’s dismissal of trust as a “poor metric,” I realized that she was right about the question being badly framed. It’s easy to get some people to trust you, if you present as news only those things that support their existing beliefs. Or if you demonize the people they resent, as Breitbart tends to do. And it works even better if the beliefs you are always confirming with your headlines have been labeled “extreme” by elites, or dismissed as unfit for public discussion. A kind of trust can be built that way: on resentment news, on transgression, on confirmation bias.

So I realized that my image of a newsroom optimized for trust was incomplete. The problem is not how to generate trust by publishing news. Trump does that with his Twitter feed, in a way. And there’s data to indicate that he is more trusted as a source of information than the news media is — by Republicans. In this sense his campaign to discredit the American press is working. In one poll, Trump as an information source was more trusted by Republicans than even Fox News.

If trust by itself is a poor metric, our design problem becomes how to combine high standards of verification with optimizing the news organization for trust.

Here’s another way to say it: In publishing news, the hard part is not to stay in business. The kids in Macedonia, using Facebook and producing a simulation of news, can do that. The hard part is not to stay in business, but to stay in journalism. Which means to accept its constraints, the most important of which are: “wait, did that actually happen?” And: “does the public need to know?”

Did that really happen? Have we given a fair description of it? Breitbart doesn’t care enough about these things to “stay in journalism.” Trump cares even less, which is why he’s a terrible source for news, and it’s terrifying that some rely on him for that. For the kids in Macedonia, and other bad actors like them, it’s probably a defect if the crazy thing they’re reporting as news actually happened.

Let me repeat: The problem is not how to generate trust by publishing news that some people like and share. In a world of open platforms, that’s pretty easy to do. The problem is how to combine high standards of verification with a bold and creative vision that optimizes for trust, and finds support among readers, viewers, listeners. Again, it’s not how to stay in business, it’s how to stay in journalism. (That phrase is not mine. It’s from the British publisher Harold Evans.)

Time to unpack “optimizing for trust”

Now I want to unpack this suitcase I have been carrying around: the phrase “optimizing for trust.” Here are some of the items it holds:

  • When I can easily understand not only the news story I read when I clicked through, but the data policy I bought into when I signed up… that’s optimizing for trust.
  • When I know that you’ll report it when it’s nailed down, and that you’ll correct it when it comes apart… that’s optimizing for trust.
  • When I can click on your reporter’s name and find not only her bio and archive, but where she’s coming from, and what motivates her journalism… that’s optimizing for trust.
  • When I can feel you getting better at listening to the internet, even as you publish on the internet…
  • When I can add my knowledge to yours to make for a better product….
  • When my attention is not grabbed but given…
  • When you as a reporter not only know your stuff, but show your work…
  • When responding to criticism, and sorting the valid criticism from the invalid, is considered a vital newsroom skill…
  • When educating people with your journalism is joined to educating them about journalism, and how it’s done…
  • When reporters share their learning curve even as readers share their expertise…
  • When the people who value the work elect to support it financially, and want it to spread to the public it was made for…
  • When you not only ask supporters for money but explain how you use their money…
  • When radical transparency combines with genuine diversity to make something better than objectivity…
  • When all those things start happening together, and form their own newsroom culture, then we’re beginning to optimize for trust.

For the past year I have been working with the Dutch site, De Correspondent, as it expands into English language publishing. They have a clear sense of how to optimize for trust using the membership model. When I look at their design, the things I have been talking about appear “in their corrected fullness.” That’s why I am helping De Correspondent break into the American market. They have vision, in Sheldon Wolin’s sense.

To conclude: everything I have talked about here arises from a shift we ought to recognize and embrace. The users of journalism — the readers, the listeners, the viewers, the subscribers, the members — have more power now. In part because they have more choice, in part because they are paying more of the costs, as the advertising subsidy declines.

Because the users of the product have more power, the makers of the product have to listen to them more. Increasingly the quality of your journalism will depend on the strength of your relationship with the people who use and value your work. I want to repeat that because it is so important… the quality of your journalism will increasingly depend on the strength of your relationship with the people who recognize, and value your work.

Optimizing for trust is a name for the shift in imagination required if news organizations are to recognize this new balance of power.


This post was originally a keynote talk at the International Journalism Festival in Perugia, Italy, April 13, 2018. I am an ambassador for De Correspondent’s expansion to the English language.

Want to be the first to know when we launch our English-language site? Sign up for updates at thecorrespondent.com.



Jay Rosen
The Correspondent

I teach journalism at NYU, and direct the Membership Puzzle Project. Advisor to De Correspondent as they expand to English-language publishing.