First-Principles Journalism: A Blueprint

What journalism could like look if we rebuilt it from scratch for our times.

Our bodies have tails that never make it out of the womb, specialized ear muscles that allow for better predator-tracking, and the means of making the fur we lost over a million years ago stand on end. We’re full of these vestigial structures, all optimized for conditions that have long ceased to be relevant for us. If you were to design a body for the role humans play in the world today, you'd do a lot of things differently.

The same is true for journalism. Its default models were built for the world as it was in pre-digital times. And when the internet came along, journalism ported itself instead of reimagining itself. Rather than ask how it could better serve us with these prodigious new powers, it settled for faster horses and Twitter-dunking.

Today I want to look at the four main vestigial structures that are holding journalism back from accomplishing its highest ends, along with the replacement structures by which the next era of journalism startups will endrun the old:

1. The static article model (vs. managed wikis with update subscriptions)

2. High-friction feedback loops (vs. paying for feedback/information flow)

3. The faux-neutral voice (vs. taxing wrongness and unfairness)

4. Attention-hacking as marketing (vs. optimizing for the truly curious)

---

We reward corrections. See something wrong, misleading, or unfair? Say something. It helps. New readers can learn more about our mission here.

As ever, #notalljournalists (especially local reporters). I know a lot of you are following here now, and my goal is (almost) never to shame individuals. It’s to discuss system-level reform.

---

Context-Setting

One overarching philosophical question here: what job ought journalism to be doing?

Taken as a separate question from what it does do today, my anchoring first principle is that humanity moves upward on the backs of explanations. Thus journalism, being the primary source that we turn to for updates to our explanation libraries, ought to provide those updates with reliably high levels of accuracy and usefulness. And it ought to make it easy for those with less baseline context to get more.

Put another way, I think journalism does its job to the degree that it provides readers with the information they need to make choices that better their spheres of influence, without forcing them to look too hard or guess at who to trust.

While there’s certainly room for other theories, my sense is that they’re mostly just compromised visions that offer less by setting the bar where it’s easier to meet.

---

Vestigial Structure #1: Static Articles

When we turn to the news, it’s generally to learn one or more of:

1. What’s new

2. What the new thing means practically

3. The background and larger context of the thing

(I’m going to refer to these three things by number for the rest of this section.)

Now, imagine an era where your main tools to meet these interests are just a printing press and some newsgatherers. Each edition of your paper isn’t cheap to produce or distribute, so you need significant per-edition sales to make the economics work. And updates (1) are where the broadest demand is, so you do a lot of that.

But one thing about optimizing for 1 is that you have to work quickly. Your product decays in value rapidly. So while you may want to meet needs 2 and 3, you can only accommodate them so much before the extra time involved erodes your sales. So you try to hit some minimum of 2 while treating 3 like a luxury good.

This naturally leads to brisk inverted-pyramid articles where:

• The headline grabs attention

• The subheadline sets context

• The lede communicates the gist of what’s new (1)

• A few clarity/color paragraphs add some of 2 and 3

But consider the modern world. Where in the past we needed newspapers to collect newswires for us, we have Twitter now. While outlets can still break news in the form of investigative reporting (and its much lesser cousin scoops), we aren’t relying on journalism for general updates in the way we used to. Lots of news can, and increasingly does, break itself. Our real needs map more to 2 and 3. We need journalists to add clarity, to give us a sense of proportion, to help us locate the new in terms of the old, and to make sure we really understand the old. But while these things are essential, they don’t match well with the incentives of 1-based models!

• If a reporter comes by information by other means than their own intrepid legwork, it’s likely that others will come by it too. So they’ll often say “those behind me can care about 2 and 3, I have to get this out there”.

• But when those other outlets see the story, they’ll assign reporters to make sure they’re getting some of the attention window too, where that window closes by something like 10-15% per hour. And there’s virtually no chance that said reporters can add much that’s useful within those constraints (mostly just a mix of rehash, rushed authority quotes, and whatever they can glean from google or colleagues).

If your model is driven by 1, you’re never going to be very good at 2 or 3. Especially if your product is static. And it’s hard to get away from static articles when all your industry knows is “readers want fast”. If you invest into updating an article to make it better later, how do you drive new distribution? How do you make your money back?

But imagine a model that didn’t involve static articles at all, where:

• When important news breaks, a team of professional explainers quickly begin building a wiki to contextualize it.

• Once they’ve reached a certain minimum of information (run through a minimum fact-checking process), they share a link to the wiki on Twitter.

• They keep updating it forever using feedback loops and incentives that we’ll get to in coming sections.

This will end up looking a bit like Wikipedia, except much better on substance. You’re not relying on volunteers, you’re not diluting your editorial attention across all popular human knowledge, you’re able to add elements like explanatory comics for visual learners, and most importantly you’re able to have a distribution element.

As for that distribution, how would readers get informed of updates? Once the wiki-link is shared, a reader who sees it on their social feeds again won’t know if there’s new information or not. Which is why Explainapedia (the news org we’re building) has partnered with Newscard (partial demo here). Our readers will be able to subscribe to stories they want to stay in-the-loop on, and we’ll send out regular bullet-point digests covering meaningful changes/updates to whatever stories they’re following.

But there’s more!

(The few readers familiar with it will think “ah this is just WikiTribune 2.0”. And there’s a limited way in which that’s true. But the differences are not few, not small, and not accidental. We admire their thinking, but learned a lot from their failures.)

---

Vestigial Structure #2: High-Friction Feedback Loops

Let’s start with a simple premise. Which of these models would you bet on producing a more reliable product:

1. One that pays for useful inputs and never stops factoring both new information and helpful feedback about accuracy, clarity, context, fairness, etc.

2. One that has a lot of friction in the way of both information flow and general feedback, and limited ability to act on that feedback anyway.

The striking progress of the modern tech industry was built on the power of iteration. “Move fast and break things.” The team that iterates more and faster will nearly always win, especially if they’re acting on a reliable stream of high-quality feedback.

But journalism has largely ignored this dynamic because of their static-article model (and the culture built around it). Their content is somewhere between 98% and 100% locked at the time of initial publication. Maybe you’ll get them to change a wrong date or statistic, but otherwise they wrote what they wrote and the best you can usually hope for is to maybe tilt their future coverage in some small way.

(I’m not being hyperbolic here. I’ve exhaustively documented my struggles getting major news outlets to correct anything, and otherwise examined how they go about corrections more generally. While many outlets once had public editors that were responsible to be internal champions for the public’s concerns, now no national US paper outside the Associated Press does. The rationale for the cuts was mostly just “well Twitter criticizes us lots now, so that’s just as good right?” But they don’t listen to Twitter really, or at least not short of a mob. And even then the mob has to be their subscribers or social peers, or at least from their side of the culture wars.)

Anyway, when I tell people that I have a corrections policy, what I sometimes fail to mention is that paying readers to surface my mistakes is only partially about accountability (though that’s certainly important). It’s also to up the odds that what I write can become and remain the most informative thing on a topic. I got something wrong? People have a reason to tell me. I missed something? People have a reason to tell me. Something new came to light? I’m beginning to pay for that too.

This system is a lot more scalable than people think. If you create the right incentives for the feedback loop (and the right structure to process that feedback efficiently), it’s easy to lap the legacy competition. And the trust and utility that come from this can make it somewhat trivial to attract the money required to pay for it.

At Explainapedia, we're going to reward readers to the point where anyone with useful corrections and updates will have no better place to bring them. And this will let us get into the “what’s new” reporting side too. We’ll be a much better partner for sources (at least for sources motivated by wanting to make sure that the world understands what they have to say, where what they have to say is truthful and useful).

---

Vestigial Structure #3: The Faux-Neutral Voice

There's a certain virtue, or at least a virtuous-seeming veneer, in the idea of being conspicuously neutral — of saving one’s personal opinions for the opinion pages and otherwise keeping the news itself holy. But it’s an impossible standard, and forcing people to perform as neutral when they aren’t actually just hurts the reader.

What blogs and newsletters get right is that humans with histories and perspectives are relatable and interesting. Journalists sacrifice this for the idea of neutrality. But they’re not neutral. They have their own biases and definite ideas, and are constantly slanting their stories accordingly (whether consciously or unconsciously). As I've covered here and here, you can tell a lot about a piece by looking at the paragraph number of where the other side gets their hearing. The journalist might say "ah yes I am neutral and told both sides" while failing to acknowledge the obvious point that locating the defense fifteen or twenty paragraphs after the the accusation is so far from fairness that it’s closer to not printing both sides at all.

Journalists aren't objective because people aren't objective. The thing to do with our biases is to be honest about them — to put them in front of the reader as an interpretive key, followed by “oh and we pay readers for raising fairness and accuracy concerns” (which can be adjudicated by a panel of sorts where necessary).

We don’t need neutrality, which when it isn’t a lie can also be a form of cowardice. We need people to be interested in being right, in subjecting their rightness to deep and constant scrutiny, and in being judged on how fairly they deal with feedback.

The solutions here are scorecarding and taxation. I keep a public list of my corrections. Readers ought to have an idea of what sorts of things I’ve gotten wrong, why they happened, and what I did in response. While we all make good-faith mistakes, we can also make mistakes borne out of situational indifference to truth and fairness. Writers who continually get things wrong on that axis shouldn’t be trusted, regardless of whether they use a neutral tone. Either they care about doing right by the reader or they don’t. If they do, having a point of view on a personal level is just fine.

I believe things. I advocate for things. I think some parties are guilty and some innocent. I think those distinctions are important. While tact and discretion can be appropriate for some complexities, directness is a virtue, and not a small one.

A bet, as the saying goes, is a tax on bullshit. It ensures the kind of objectivity that actually matters. Our biases at Explainapedia will be declared. We won’t try to hide them. We’ll just make sure that our structure keeps us honest in a reliable way.

---

Vestigial Structure #4: Attention-Hacking

There are three ways in which a presentation can be interesting:

1. If it’s sexed up with color and sensationalism

2. If the presenter is unusually charismatic and/or passionate

3. If the presentation communicates interesting things to the already-curious

While I suppose you could label both 1 and 2 as attention-hacking (as both rely on the presenter exploiting our wiring to generate interest that otherwise wouldn’t be inspired by the subject matter itself), the second is obviously less problematic. That said it’s also only useful in a narrow-ish way. Lots of people can sit in rapture and walk away with their interior libraries totally unchanged. To the degree that 2 helps the world, it’s largely when it sparks a fire that moves the audience towards 3 after. (Though in some exceptional cases you get 3 delivered by a 2, which is just lovely.)

But so far as newspapers go, the obvious problem is that 1 is the easiest and most reliable way of driving sales. For example, I wrote on Tuesday about ProPublica leaking dodgily-obtained private tax returns for the sake of making their preposterous op-ed on taxation policy more superficially interesting and “newsworthy”. (My story has a new appendix by the way, including FAQ and a few minor corrections.)

If your goal is reaching the highest number of readers (which is largely true even for subscription-based newspapers, as sign-ups are a volume game for them), your incentives are towards the sensational. While 2 and 3 may still matter, they just don’t fit as well. The whole faux-neutrality thing means your individual journalists largely aren’t free to be very interesting, and limiting focus to the genuinely curious would mean leaving too much traffic on the table.

And when you get locked into this sort of dynamic, a few bad things happen:

• You begin selecting against the most curious journalists (as well as the most curious readers, who are likely to be your best feedback sources)

• Your product becomes the Cheesecake Factory, serving up sugar and volume to make up for a lack of dishes of real value to those with specific palates

Of course, all this is avoidable. When's the last time Wikipedia ran an ad or a juicy headline? They win because the curious come to them. This is why Explainapedia is going to be a mix of all worlds. Like Wikipedia, we’re going to cater to the curious, and we’ll never feel pressure to over-sell anything. But like newspapers, we’ll also have a Twitter presence and active distribution. And like newsletters, we’ll employ the top explainers, each well-compensated and with their creative freedom unleashed.

Whether we’re successful or not, I expect this to be the future of bundling.

---

Call to Action

First principles would not lead you towards the current structure of journalism (or that of books, a related topic for another day). They’re both anachronistic vehicles defined by forces and limitations of eras long past. We've just lacked the resolve and strength of vision required to create something fitting for our times.

It's time to build.

(Sorry for the gendered language in the image. I don’t endorse that part of it!)

As the next step of what we’re building, if you’re interested in serving on an appeals panel for corrections litigation (which is to say weighing and voting on whether something qualifies for a corrections bounty), please DM me on Twitter.