Media Refragmentation and Funnel Capture

Plus! Disaggregation, Reaggregation; Separation of Powers; Multichannel, R(t); Central Bank Fiscal Policy; More

In this issue:

This piece was originally published exclusively for subscribers, but I’ve decided to make it free for all.

Media Refragmentation and Funnel Capture

Anecdotes become a trend when N=3, and so far, this has happened three times:

The now-classic story—somewhat ornery journalist finally gets mutually fed up with a publication, quits, and promptly ends up writing the same material for a more lucrative audience—is not just a small-scale story about the media business. It’s also a visible example of how shifts in transaction costs can fundamentally restructure industries.

The logic of bundling media has been explored in depth elsewhere. Any time there’s a product with low marginal costs and heterogeneous consumption preferences, it makes economic sense to bundle several semi-related products together and offer them at a fixed price to the large audience of people who like some mix of the products offered, but don’t like any one option enough to buy it à la carte.[1]

The logic of unbundling, at least the angry-resignation-letter variety, is not driven by consumption preferences but by HR and PR concerns. I’ve written in the past (see here and at more length here) about the “surface area” problem of scaling a company: as a company grows, the probability that someone associated with the company will do something unpopular rises. At the same time, the newsworthiness of this behavior rises with the size of the company.

For newsrooms, this pressure also shows up internally: Matt Taibi, Andrew Sullivan, and Glenn Greenwald all have views that some of their coworkers find scandalous. Any time a news organization employs a controversial, high-profile writer, it gives the organization a larger audience, which means more staffers, which means higher odds of a group of employees getting annoyed at their controversy-courting coworker.

(It’s probably not a coincidence that the political newsletters on the Substack leaderboard are so heavily weighted towards a) left-of-center writers who criticize the left, and b) right-of-center publications that don’t support Trump. In both ecosystems, the most venomous arguments are between people who fundamentally agree on the broad philosophical points but disagree on the current object-level debates informed by those points. This appears to be fractally true; libertarians and socialists spend an inordinate amount of energy criticizing other people who share the same identification but are either too impractical or too ideologically impure to merit their peers’ approval.[2])

So the cycle for big media companies seems to be:

  1. Hire a marquee writer with their own fanbase, or hire a writer who builds such a fanbase.
  2. Discover that the writer earns these fans in part by monopolizing a less competitive part of the political Venn Diagram.
  3. Say goodbye to them when the ideological gap gets too glaring.

This has given Substack an eclectic collection of top authors. Its most notable political publications are, by the standards of their audience, contrarian. But they’re also publications that reward writers for building their own audience, rather than renting one from an existing institution.

Very temporarily, the existence of Substack and other direct-to-audience monetization tools is a godsend for media companies. The internal fights are vicious (everyone who loses one of these fights is well-equipped to write a memoir, and also well-connected enough to broadcast their version of the story; internal squabbles at media companies quickly turn into PR counterinsurgency). An alternative monetization option is a good pressure release valve.

But in the longer term, it’s quite scary to media companies. Their distribution is being eaten by social media, and now their monetization is being eaten by Substack.

Good writers get avid social media followings, but don’t really monetize them, except by switching careers. (For the longest time, Ben Smith’s hundreds of thousands of Twitter followers did a lot for Buzzfeed’s ad sales and not so much for Ben himself—until he switched his handle from @buzzfeedben to @benyt and, presumably, cashed in on the distribution he’d accumulated.) Now, there’s a more direct way to create that audience: join a mainstream publication, cater to a peripheral audience of that publication, shift outside the company Overton Window (or wait for the window to shift on its own), then quit and take the audience with you.

Naturally, this has led to the rumor—quickly denied—that Substack is in merger talks with Twitter. It obviously makes sense. For anyone who wants to make a living from text, Twitter is the best way to get an audience and Substack is the default way to monetize it.[3] If they both continue to commoditize their nearest complement, eventually they’ll start bumping into each other; it makes strategic sense for them to team up right away.

But the surface area problem rears its head again. It’s one thing for Twitter to be an open platform for some groups; it’s another entirely for Twitter to earn its subscription revenue from promoting views that are, within a given social bubble, fringe. Twitter would be buying a revenue source, and paying to get rid of a strategic concern, but would also be buying endless PR problems. Since the core Twitter product is free, users can’t complain much when Twitter shuts them down. But if it’s a livelihood, they will complain, and they’ll use their mailing list to do it.

Some of Substack’s strategic decisions, like offering legal support to writers, represent a competitive advantage for the company’s current audience but a poison pill for acquirers. An aggressive commitment to neutrality, even at a high cost, is attractive to iconoclasts, but large companies don’t like to employ such people (if you want to be an iconoclast at a huge company, step one is to own a majority of the super-voting stock).

These trends end up being materially worse for media companies. And every media company that tilts towards politics has already been looking ahead to Tuesday with dread. Every media company is a middleman that connects talent to an audience, but when the cost of a direct connection is lower, the high-cost middleman business is an ongoing adverse selection problem. Anyone who can make it independently will try to, and anyone who can’t—anyone who can’t get an audience to click or can’t get a reader to pay—is hard to justify hiring. Like other businesses—retail, equity research, travel—a more transparent system with lower transaction costs obliterates any middleman optimized for a less efficient status quo.

[1] Bundling fails in cases where the cost is disproportionately time rather than money. In that case, getting extra information that’s not relevant is a cost, not a benefit. A savvy bundle-seller might try to customize what they deliver to customers, but this, too, is a challenge: the exceptionally busy ones are unlikely to spend a lot of time filling out surveys on their interests or painstakingly sampling sub-publications and selecting the ones they like. When time is the main cost, curation becomes overwhelmingly important, and the easiest way to expedite curation is to minimize the number of options.

Bundling is tricky with any other product where the buyer is hypersensitive to the qualitative traits of the product. If you bundled a weeklong Paris vacation with a mandatory week in Branson, or vice-versa, you wouldn’t get many takers.

[2] This explains another feature of discourse around independent media: Substack is sometimes dismissed as being a site for people of a particular political persuation, but that persuation is whichever one the speaker doesn’t share. To someone on the right, Substack looks like a venue for ex-Guardian columnist Glenn Greenwald, Goldman Sachs bête noire Matt Taibbi, and not-on-the-Trump-Train conservatives. To someone on the left, it’s a haven for Russiagate skeptic Gleen Greenwald, fellow Russiagate skeptic Matt Taibbi, and Weekly Standard/National Review conservatives. Intra-ideological squabbles are harder to see from the outside, so both sides exaggerate Substack’s partisan tilt. Political colorblindness consists of seeing shades of your own team’s colors, while the other side is a single consistent hue.

[3] There are other paired products like this. The under-monetized platform tends to have more reach, so it’s the best place to build an audience. The more heavily-monetized one trades broad reach for money. This showed up in the old blog-to-book pipeline, the politics-to-speaking-engagements revolving door, and in switches from sell-side research to the buy-side.


Disaggregation and Reaggregation in Fundraising

Politico has an interesting profile of Democratic fundraising organization ActBlue. Political campaigns have often been early adopters of advertising technology—the election cycle means there are fallow periods that keep people from getting too set in their ways, and in every race the underdog has an incentive to try novel marketing techniques since they have less to lose from failure. What ActBlue shows is that when marketing is more scalable, it gets more centralized: instead of having every political campaign reinvent online targeting on its own, ActBlue can offer a mass-customized solution for all of them. ActBlue isn’t trying to build a monopoly (there’s a large segment of the political fundraising market that they won’t try to touch), but economic forces push their market share higher regardless.

Separation of Powers

Airbnb plans to IPO soon, and, like many companies, it’s restructuring ahead of the offering. Airbnb’s new structure includes an endowment fund for hosts, backed by company stock. Platform companies have a political dynamic: they exist to facilitate interactions between interest groups, and their pricing and marketing decisions have distributional effects. In some cases, the utility-maximizing decision ends up distributing more of the gains to a group that’s already ahead, which is politically challenging—but being equitable would be a poor business decision. So the company has decided to use the efficiency ex ante, equity ex post approach: by giving away some of the upside after the fact, they can be optimally ruthless at maximizing that upside.

The Mortgage IPO Boom

The WSJ points out that six of the 30 largest mortgage lenders in the US have gone public, or plan to, this year ($). The mortgage business has gotten better over time as more mortgage companies have learned how to retain customers and convince them to refinance, which has made the market more efficient. But the mortgage industry looks suspiciously like it’s at the top of a very long cycle: rates are low, and refinancing responds to changes in rates, not to the absolute level; and there’s been a step-function increase in mobility, which means lots of first-time homebuyers. Clearly the owners of mortgage companies view this as something close to a cyclical peak.

Multichannel Retail

Walmart is testing a hybrid brick-and-mortar/e-commerce model at four locations, as retailers continue to blur the distinction between online and offline. One interesting note is how tech-enabled the physical store process has become:

Walmart will also be testing its in-house app designed to speed up the process of moving items from the backroom to the sales area… The app will be on a device that uses augmented reality to flag products that are ready to be moved to the sales floor, eliminating the need to scan each box separately… According to Walmart, this “first-time pick rate” program has already shown promising signs: successful first attempts at locating items has increased by 20% in difficult “categories.”

There are practical constraints to building a perfectly automated warehouse, but there are also constraints—costs, error rates, variable timing—to relying on humans. At scale, ubiquitous smartphones can push those costs lower, by making more of the packing process an interaction between humans and software rather than humans and the real world.

Recursive R(t), Redux

In September, I wrote about recursive R(t), the fact that widespread coverage of the pandemic means people will adjust their behavior in a way that reverses positive and negative trends in infections. The latest case study in this is England’s lockdown. Tyler Cowen notes that adherence to masking and social distancing norms was weak in parts of the UK that did not plausibly have herd immunity. In July, the daily infection rate for the UK as a whole had dropped over 90% from the late-April peak, and, as in so many other domains, people reacted to the rate of change in risk rather than the absolute level of risk.

Case Studies in Central Bank Fiscal Policy

Nathan Tankus points out that the Economics Not-Quite-Nobel is a cash prize issued by a central bank, and thus technically constitutes fiscal policy performed by a central bank. The usual separation of powers is that central banks are academic and technocratic, and can quickly react to crises but only through a limited set of tools. Legislatures have broader powers, but spend more time arguing and are less persuaded by data. Most of the time, this is not especially problematic: in normal economic circumstances, fiscal and monetary policy are fairly close substitutes for solving aggregate problems, although they have different distributional effects. But 2020 is definitely a year in which central banks would have liked the ability to make direct cash transfers to a large number of people. Instead, we have a situation where the economy is booming for sectors that are closely tied to central bank policy (anyone who issues bonds or stocks, anyone who owns residential real estate), while it’s still a deep recession for the rest of the economy. A Fed that could award every US citizen 1/1,000th of a Nobel Prize each month for the duration of the Covid recession would have much more monetary firepower for the next crisis, and would have ensured a more normal-looking outcome to the current extraordinary circumstances.