On writing digital reports

What if “digital-first report” is an oxymoron?

In the very early days of television, networks frequently broadcast filmed versions of radio scripts. It makes some sense. After all, radio scripts already have characters and dialogue and stories, all just waiting to be mined.

But filming a radio script doesn’t make for particularly good TV. Think about the things that go into television.

1. Visuals. Radio scripts rely on dialogue to carry everything. Either the actors need to describe what they are seeing or a narrator needs to set the stage. But in television, a few establishing shots can cover a lot of ground. For that matter, you can convey a lot without words at all. For example, episode 7 of the 2021 show Only Murders in the Building—“The Boy from 6B”—features no audible dialogue at all. (Some characters speak in ASL, which was subtitled for those who do not speak it.)
2. Proximity. A good television program can put you right in a character’s face, allowing viewers to glimpse subtle — even fleeting emotions. That’s a fundamentally different sort of experience from radio or even theater, where distance from the stage limits actors to broad, over-exaggerated facial expressions.
3. Multiple angles. Panning or zooming can create its own sort of storytelling. Consider the now almost-cliched use of a scene that pans across some scene of devastation, only to conclude by zooming in on the face of a protagonist whose eyes reflect the horror. Conversely, a long tracking shot can build tension, as in “The Long Bright Dark" from season one of True Detective or a sense of exhausted determination, as in the 11-minute, single-take fight scene from “Blindsided" during season 3 of Daredevil.
4. Timing. If you’re still producing shows for ad-supported television, then you’ll need a sense of rhythm, too. That is, you need to create a set of mini-climaxes at set intervals in order to keep viewers around through two-minute commercial breaks. And you’ll need an even bigger ending to keep viewers coming back from week to week. (If you’ve ever watched a British show on American TV, you’ll get this. Commercial breaks are in weird spots — sometimes even mid-scene. Shows not written with ad breaks in mind lack the proper internal rhythms needed for natural ad breaks.)

Back to reports

So what does this amateurish analysis of television have to do with writing reports?

My worry is that “digital-first report” is the think tank equivalent of filming radio serials for television.

We’ve written the same way for such a long time that we maybe tend to forget that what we consider to be a “normal” way of writing research outputs is in fact extremely media-specific. Consider a few examples.

Boxes

These are meant to display content that is slightly tangential to the main body of content and/or the main flow of the argument. In print, you need some way to signify that a piece of content is:

1. Sort of related to the the information in the passages surrounding it, but
2. Sits outside the flow of text.

Hence, the box.

The web has very different conventions for tangental but related content. You put that on a different URL and you add a hyperlink.

Footnotes

This one particularly bugs me. Footnotes are a terrible user experience anywhere. We use them in print because they are the least terrible way of identifying references and/or including material that doesn’t quite merit a text box.

The print convention is to add a little number at the relevant point of the main text, then add the supplemental information (either the reference or the tangental text) at the bottom of the page. Getting back and forth between them requires a context shift.

Of course, in HTML, extra information—whether tangental text or a reference to another text—is literally only a click away.

Clicking a footnote link that then provides the information you would need to go locate a physical copy of a thing in a library or bookshop adds a totally unnecessary layer of complexity. The written-out citation is a print holdover that makes no sense on the web. Just link directly to the thing!

Linear writing

It’s tempting to think of the linear essay as just The Way Writing Is Done.

But the ordering of items isn’t an essential or intrinsic property of research findings or of the presentation of information. It’s a product of the technology we have traditionally used to communicate information.

Linear essays are the result of adding markings to a physical surface.

One need only look to oral traditions to realize that storytelling is not inherently linear. Anyone who has ever listened to a Southerner spin up a good yarn knows that oral storytelling is filled with digressions, flashbacks, flashforwards, brief forays into other stories.

Translating on oral tradition into print is an act of imposing an order upon the disordered — an order that exists not because it’s somehow inherent to the “real” story, but because it’s the way that print technology works.

All sorts of “natural” orderings are like this.

Take alphabetizing. What could be more natural than that? Only, it’s actually a fairly recent way of ordering information. Medieval information architecture was entirely topic-based. The more closely two things were related in subject, the closer they were physically located.

Indeed, alphabetization is a surprisingly modern innovation. As late as the start of the 17th century, alphabetized dictionaries came with explanatory text for using an alphabetized reference. The idea didn’t really take off until the French Encyclopedists in the middle of the 18th century.

That’s largely a matter of technology. The printing press made books cheap enough that libraries could amass thousands of items. Once a library reaches a certain size, subject affinity is an inefficient grouping mechanism, as it requires that a librarian be intimately acquainted with every book in the collection.

Just like alphabetization is a product of a technology that makes multiple copies of things cheap, writing in a linear fashion is a product of technologies (scrolls, codices) that require putting some things literally in front of others.

Hypertext isn’t like that. Ideas need not be spatially arranged at all. They can exist on different screens and be related via a series of internal links, which in turn can be followed in any order.

Hypertext lets us return to a world in which content can be organized by relationships precisely because—unlike the medieval monk—Google actually can be intimately acquainted with every text in the collection.

The upshot

The report isn’t some sort of Platonic ideal of the research output.

If there‘s a Platonic ideal of a report, it’s this one, designed for New America by Soapbox’s Anja Pölk.

It’s a radio script—an output that was optimized for a specific medium and whose entire guiding ethos is constructed around the needs of printed pages.

Translating a report into a digital format is like filming a radio script. Yes, it might get the basic ideas across. But it’s a pretty limited output. No matter what you do with a report, it can no more embrace the capabilities of hypertext than a radio script can convey the information needed to produce good television.

We shouldn’t be holding up the digital-first report as an ideal.

It is, at best, an interim solution—a stopgap that is good enough for the short term, but whose main function is as placeholder until we figure out how to write things that are built for the new medium.

Unfortunately, most of what I see in WonkComms land isn’t focused on replacing the report. It’s around moving the report into HTML.

Don’t get me wrong. A report in HTML is an improvement over attaching another PDF to your website. And page-to-PDF solutions can accommodate the people in your audience who still want to read on paper. (And yes, those folks do still exist, for all that their numbers are dwindling.)

But page-to-PDF solutions aren’t digital products. They are radio scripts that just happen to be filmed for television before they are broadcast on the radio.