Why do paywalls work?

When The New York Times launched its paywall in 2011, I called it unsustainable. The paywall was porous; specifically, it didn’t meter visits coming from social media. That meant, I argued, that it would exempt more and more people over time and therefore that it wouldn’t work.

A lot has changed since then, and I think it’s only fair to admit that I was wrong. The Times has more than 3 million digital-only subscribers, for a print+digital total of 4 million. The Washington Post has 1.5 million digital subscribers. I’m lucky to work for a publication that is also thriving thanks to its focus on subscriptions. As Ken Doctor wrote in his 2018 year-end roundup at Nieman Lab, “the reader revenue revolution is real.”

So how did I miss it?

For one thing, my piece criticizing the Times paywall focused heavily on its planned social-media exemption. That exemption is gone. The Times paywall is less porous and much smarter today than it was in 2011, and that partly accounts for its success.

But that doesn’t really explain anything. I wrote that The Times’ assumption that “restricting access [from] blogs and social media would be counterproductive” was “quite true.”

The reason I thought that this assumption made sense helps to explain how my vision for the future of media diverged from the version we got, and offers a bit of a puzzle going forward.

The reason you had to give access to your articles for free to visitors coming from social or blogs was simply that in both cases the biggest problem users faced was that there was too much to pay attention to. If you cut someone off, they’d happily resume browsing and quickly settle on something else that was free and nearly as good. If that sounds off, or like I’m devaluing professional journalism, recall that 2011 was just 2 years after Clay Shirky published Here Comes Everybody. The idea that there was an endless supply of great stuff published by smart people for free was starting to catch on. Everyone was an expert on something (or enough were), and everyone had access to a printing press. All you had to do was filter out the good stuff, and that’s what social media was starting to do.

I summed up this attitude in one of my first posts for this blog:

There are many reasons why consumers resist paying for content.  One of those is the reality that you don’t need to do so in order to be well-informed.**

The truth is that when the NYT starts charging for content I can go to the next best free paper.  If all the papers started charging, I could read only blogs.  These blogs can freely quote the lede and important ‘grafs from news stories on their way to offering commentary.  So I wouldn’t really lose much, if anything, by not reading the Times.

Why didn’t it turn out that way?

I’m honestly still not totally sure, but here are a few possibilities:

  1. The digital ad market collapsed. I’ll return to this in a second, but will note for now that one popular explanation for this is the Facebook/Google “duopoly.” Another is simply that an endless supply of content drove prices down.
  2. Filtering “the good stuff” turned out to be harder than I thought. Our filters have gotten really good in some ways, but remain very poor in others. Social media is still doing a poor job of filtering out harassment, hate speech, etc. And the sophistication of disinformation efforts proved greater than what many people thought about even a few years ago (certainly including me).
  3. Social media just didn’t turn out that well. My thoughts on this are here and here.
  4. Brands just matter a lot more than I thought. Probably true.
  5. People cared more about the kinds of content only journalists can produce than I anticipated — and would pay for it. From early on, I was clear about the fact that certain kinds of content are not easily replicable by free competitors. In particular, beat reporting and investigative reporting are expensive, and not many other organizations outside of journalism are set up to do them. But investigative reporting was relatively rare, and more to the point it seemed unclear whether people cared much about the important stuff. It seemed like the serious, expensive reporting was being subsidized by the other stuff. (I wrote about this here and here.) So, basically journalism had a competitive advantage in the stuff people mostly didn’t care about. And the stuff people did pay for — the features, the profiles, etc. — was suddenly forced to compete against a mountain of free content. But maybe I just underestimated people. Or, at least I underestimated a few million people? And perhaps the Trump era has simply caused more people to care about this stuff, and therefore to be more willing to pay.

Again, I was clearly wrong. But I do still think that many people in media underestimate the significance of the mountain of free content, and how it’s affecting their businesses. Clearly, the ad duopoly has affected publishers. But what if the digital ad market were more splintered? The fact remains that audiences are easier to find, and there’s more content to advertise against. There’s no universe I can think of where digital ads wouldn’t have gotten a whole lot cheaper, simply because of additional supply of content and superior targeting technology.

This has had a rather strange effect on the journalism market. Let’s drastically oversimplify go hypothetical for a second and bucket content into three groups. The first is the highest quality, the most expensive to produce, and can therefore sustain subscription businesses — at least potentially. (Think The Times, The New Yorker, etc.) The second one is cheaper to produce, but still requires journalists and has a harder time getting anyone to pay. (I’m thinking of more aggregation-heavy operations, but insert your least-favorite ad-driven digital media company here.) The third bucket is essentially user-generated content, most of which now lives on a few big social-ish platforms, like Facebook, Instagram, Twitter, and YouTube. Forget the stuff that buckets 1 and 2 publish on these platforms. The content here requires no journalists to produce, and is entirely ad-driven, thanks to massive scale.

All of these content providers are competing for a supply of attention that is in some sense limited. I figured that Bucket 2 — the digital, ad-driven, lower-cost outlets — would be in a better position in the market than Bucket 1 — higher-cost, premium outlets — because of both a better cost structure and a commitment to free content. The endless supply of content from Bucket 3 would make paying unattractive, and in that world Bucket 2 would usurp Bucket 1.

So here’s where we get to the strange effect of the tanking digital ad market. Instead, Bucket 1’s worse cost structure had a counterintuitive, positive effect. These publishers realized earlier on how hard it would be to become sustainable on ads alone, and so pivoted to reader revenue sooner. When the ad market tanked, Bucket 2 was suddenly in trouble. And Bucket 2 failing is basically the best thing that can happen to Bucket 1. I said before that readers won’t pay because of all that content, most of it from Bucket 3. But that’s not quite right; the free competitors in Bucket 2 were a big part of the story. The idea was that if you don’t want to read Bloomberg, there’s always Business Insider. If BI needs to charge — or if it were to disappear — Bloomberg’s position is strengthened.

OK, I’ve spent 1,000 words going in a big circle. I missed that people would pay for news because… I missed that people were willing to pay?

My point, though, is that at least part of why I was wrong was that I didn’t predict that the content supply shock would hit the ad side of the market earlier and more dramatically than the user side. The endless supply of ad spots lowered digital ad rates, which is consistent with my whole point about endless amounts of free content. In the short-term, this strengthened subscription publishers by hurting their free competitors.

I realize that I’m at risk at this point of responding to a failed prediction by amending my failed model rather than changing my belief. That’s a typical sign of confirmation bias and bad forecasting. So it could well be that my whole theory of the content supply shock is just wrong all the way through.

Nonetheless, I would not be surprised if the content supply shock eventually catches up to the user side of the market. By that I mean that peoples’ willingness to pay will still be affected by how many alternatives there are, and though the alternatives to paying may not look great now, that could easily change.

Imagine, for instance, that digital media startups knew that the ad market would tank. Many of them might have combined their lower-cost approach not with a socially optimized strategy but with a focus on subscriptions. What might Vox.com look like if it had started from Day 1 with reader revenue in mind? In fact, platforms like Substack and Patreon are making it easier for digital-first publications to get reader revenue. So, while The Times and Bloomberg are getting a brief respite as their free digital competitors pivot to reader revenue or fail outright, they’ll soon face a new wave of competition from digital competitors set up to compete without so much focus on advertising.

Moreover, I expect we’ll see new and clever ways to aggregate amateur content and free, subsidized professional content (from think tanks, universities, etc.) that will create new Bucket 3 competitors and will attract many readers as a cheaper alternative compared to subscriptions to premium journalism.

In other words, the reader revenue revolution is real. But subscription businesses shouldn’t get complacent, because it may not be indefinite.



Leave a comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s