Platforms and bundles

Is bundling the future of research communications?

Digital publishing stands at a crossroads. With ad-supported content looking more-and-more like a dead-end (at least for everyone who isn’t Google or Facebook), digital publishers will need a new business model. The two most promising alternatives thus far:

1. Move to subscriptions.
2. Bundle content.

Ev Williams (of Medium and Twitter fame) recently outlined the two options ($). For “outlets that are big in terms of content volume/frequency or that have superfans,” subscriptions are an appealing option. That is, if you’ve the New York Times’ output or the New Yorker’s distinctive style, you’ll have little trouble convincing a lot of readers to pay you a few dollars each month for the privilege of reading your content. Subscriptions are also a good model for sites that can combine “low costs and a niche audience,” like Stratechery, Ben Thompson’s one-man shop specializing in the intersection between business models and the tech industry.

The newspaper is one of the first media bundles, pulling together reporting, classified ads, obituaries, comics, sports, opinion, fashion and horoscopes into a single, inky document that was perfect for advertisers.

Unfortunately, most publishers fall somewhere in the mushy middle—too small and/or too few superfans to work at scale but too big and too generalist to work with the Stratechery model.

The underlying problem for publishers is a mismatch between content and attention. Content is effectively infinite, but attention is not. Moreover, attention is a zero-sum game—every moment spent reading something on the Internet is a minute not spent chatting with a friend, going for a walk, watching television, or eating a fancy meal. Subscriptions require an audience that is willing to open your app instead of anything else they could possibly be doing at that moment.

That’s a tall order.

It’s also not one that many think tanks or research organizations will be able to deliver. Maybe Brookings has the volume and/or superfans to pull of becoming a destination site. But not many do. And even the smallest think tank is still too big and too generalist to sustain the Stratechery model.

Bundling to the rescue?

But if subscription-style destination sites aren’t the answer, there’s still that second fork—the bundling option.

The most common sort of bundle here in the U.S. is probably the cable package. And, yes, I know that’s not the best sort of example on account of how everybody hates their cable companies. But the underlying economics are why those bundles still exist.

My other big hobbyhorse: Transclusion. More on that in another post.

Spotify is (perhaps) a less polarizing example. If you’re old like me, you probably remember buying music piecemeal, one album (or sometimes even just one single) at a time. Spotify bundles music. A subscription gets you access to (more-or-less) every piece of music ever recorded. For the superstar or the artist with a very niche audience, selling directly might be the better option. But for most musicians—and certainly for most consumers—the bundle makes a lot of sense.

Williams thinks the bundle model will work for text, as well. After all, Williams says, there are plenty of people in the world who are willing to pay for “news, journalism, analysis, opinion, essays, instruction, etc.” But what they won’t want to do is manage dozens of different subscriptions. Hence, the bundle. Here’s Williams:

There won’t be a Spotify of publishing — with literally everything you want. But there will be a Netflix and Hulu and Amazon, etc. — each with a substantial amount of things you want. You might also have your superfan subscriptions (Patreon-based individuals), and your company-expensed subscriptions (The Information), but most consumers will have one or two of the big bundles.

I’ll admit that I’m skeptical about the text bundle generally. There’s one really important disanalogy between text on the one hand and video and music on the other.

Text is really cheap to produce. Video and music are not.

Sure, technically I can produce video with my smartphone or capture audio with an inexpensive USB microphone. But even a cheesy B-movie (like, say, The Asylum’s MegaFault) has a $2 million budget.

By contrast, you can write one of the most famous novels of the late 20th Century by longhand in cafes while your infant is napping. With nothing more than a public library card, you can publish text that looks every bit as polished as something out of Backchannel or The Ringer, in large part because you can use exactly the same piece of technology that once powered those sites.

Because text is so cheap to produce, there’s no way a bundle (or even a collection of 2 or 3 bundles) will ever cover more than a small fraction of it. And that, in turn, means that anything I can get in your bundle is something I can get somewhere else—including, probably, for free.

Or almost anything, anyway.

The think tank advantage

Text may be cheap. But policy research isn’t.

For starters, it requires at least some degree of expertise to produce. It also requires access to data. Indeed, there’s a good argument to be made that it’s willingness to wade in and get your hands dirty wrestling data into shape that separates the policy wonk from the pundit.

But while there’s more data than ever, it’s not all free. Sometimes that’s literal (e.g., legal datasets from Lexis-Nexis or business data from Dun & Bradstreet). Other times, the costs are more indirect: ensuring compliance with privacy restrictions on health or education data, guaranteeing confidentiality of proprietary private industry data, or gaining clearance for secret national security data.

It’s far easier to clear all those hurdles when you have the backing of a think tank or a university or a government agency.

I’ve argued before that it’s tough for research organizations to offer highly-differentiated content. I still think that’s true, but I don’t think it’s as clear as I’d intended it to be. That is, I continue to believe that most think tanks and government agencies will have a tough time distinguishing themselves from other similar institutions.

But I do think there’s a pretty good case to be made that policy research is highly-differentiated from other types of nonfiction.

And that means a bundle built around rigorous policy research—a bundle of think tank content—could be something that would work.

Think tanks as platforms, part II

There are some relatively easy low-tech ways of creating policy bundles. They could form an old-fashioned webring (anybody remember those?), where various research organizations simply link their websites to each other. Or they could coalesce around an existing platform, such as Medium or Atavist. Both are easy and inexpensive. But both also have some pretty significant drawbacks.

1. Poor internal navigation. Webrings simply link existing websites together. Each site still has to be navigated/searched separately, meaning that users still have to do a lot of work to find what they are looking for. The same would be true for sites recreated in Atavist. Medium has more internal links between content (thanks to tags and and its out-of-the-box relevant posts feature), but it suffers from the fact that…
2. There are already a lot of non-policy organizations on existing platforms. Platforms are, by definition, open to all comers. Inevitably, non-policy work would get mixed in with (and, possibly overwhelm) policy research. And those other non-policy writers choose to work on those platforms because…
3. Existing platforms are optimized for blog posts and articles. Atavist is built for journalists. Medium is built for bloggers. And while research results sometimes work as longform narratives or even as relatively short blog posts, often they don’t. The real problem here is…
4. Platforms don’t let you create modular content. Blog posts and articles are artifacts. Policy communication needs campaigns. And while you can certainly use Atavist to produce a beautiful research comms artifact, you can’t use it to create content for a campaign.

There’s a better (albeit more expensive) option. Build a platform.

Imagine a CMS that’s optimized for the kinds of products that research organizations produce. We’re talking real content management here.

• The kind that provides researchers with an authoring experience that can entice them out of Word documents.
• The kind that integrates with a robust set of charting tools that allows researchers to create publication-ready data visualizations.
• The kind that keeps content modules separate and tags them with a robust system of metadata that makes it easy to find existing modules and remix them into new works.
• The kind with robust content APIs that allow you to publish HTML to your website, ebooks to Amazon, videos to YouTube and Facebook, and Q&As to Google Home.

Now take that one step further. Imagine you’re searching for the latest study on universal basic income, and as you navigate to the relevant topic, you see a blog post from the Economic Policy Institute, a long article from the Roosevelt Institute, and an interactive data visualization from the Urban Institute. Each piece is branded, but they are all available right there for the journalist rushing to hit a deadline and for the legislative aide scrambling to prep briefing slides.

Or one step further than that. Imagine you’re sitting at the Center for American Progress writing a blog post on college admissions. You’d really like to mention socioeconomic affirmative action in passing, but you don’t want to spend the extra time and words explaining the concept. So you simply grab the nice explainer module of the topic from The Century Foundation and drop it right into your article. You can focus on extending the conversation instead of covering the same ground.

A platform that’s optimized for policy research, designed for modular content, and shared by similarly-aligned think tanks has the volume to compete with the NYT. And it has the kind of differentiated content that leads people to seek it out as a destination. It’s a platform with a niche audience, certainly. But on the Internet, niche audiences are still pretty big in absolute numbers.

How do we get there from here?

That’s the $64,000 question. I didn’t have any solid solutions when I first wrote about this idea some 3.5 years ago. My takeaway then:

So I’m suggesting one of two things:

1. That a bunch of forward-thinking think tanks get together and fund a centralized digital team, one that then builds a custom CMS (or heavily customizes an existing open source CMS) that supports everyone.

2. That an existing agency (or a new one developed for this purpose) build such a CMS and then take on a lot of think tanks as clients.

The first one is radical and different and weird and ultimately unlikely to get buy in from an industry that’s pretty methodologically conservative.

The second one may represent too small a market for any agency to find worthwhile.

I’m not sure where that leaves the think tank industry. But I’m not super optimistic about its future.

I’m still skeptical about (1). Meanwhile, (2) faces a collective action problem—agencies will want some assurance that think tanks would buy a new platform; think tanks want some assurance that a custom platform would work and that other think tanks would also join; no single player has any incentive to be the first to commit.

And as any good policy researcher will tell you, collective action problems are some of the toughest nuts to crack.