Copyright Unbalanced: From Incentive to Excess (12 page)

Eli Dourado

I
N MARCH 2011,
the
New York Times
debuted its new online paywall to rumors that its development had cost the company as much as $40 million.
¹
Arthur Sulzberger, the
Times’
publisher, insisted that it cost much less, and
paidContent
reported that the real number was close to $25 million.
²
However much the paywall cost, many were aghast at how the
Times
could spend so much on a paywall that barely worked.
³
Before the paywall even officially launched, techies figured out how to disable it. They did so with three lines of JavaScript:

Others did it by injecting only two lines of CSS:

Even if one does not understand JavaScript or CSS very well, a passing glance at this code shows why it was so easy to get around the paywall. The
Times
implements the paywall so that the requested article is there all along. There are two sections of the page, called “overlay” and “gatewayCreative,” that obstruct the user’s view of the article. The “body” of the page is instructed to hide any content that does not fit “above the fold,” so that the user cannot simply scroll down and read the article. Simply instructing the browser to hide “overlay” and “gatewayCreative” and to allow scrolling on the page’s body enables the reader to get to the requested content. What’s more, with modern browsers, these instructions can be executed automatically any time a user is browsing nytimes.com. The end result is that anyone with very modest technical skill can configure his web browser to seamlessly bypass the
Times’
paywall. While the code used to implement the paywall has changed a little in the last year, it’s still terribly easy to bypass.

How could the
Times
be so incompetent? Is creating a paywall so very difficult? To make matters more curious, consider that the
Times
has had a completely functional and effective paywall (for example) for its crossword puzzles that long predated the main site’s paywall. The crossword puzzle paywall operates on a server-side basis and is not easily bypassed. Why would the
Times
implement a faulty news paywall on a client-side basis?

It seems clear that the
Times’
paywall was built to allow circumvention by those determined to circumvent it. The official exceptions to the paywall policy underscore this point. The
Times
originally announced that non-subscribing users would get 20 free articles per month, although this number has now been decreased to 10. In addition, the
Times
participates in Google’s “First Click Free” program. This means that users coming from Google search can access articles for free, up to five times per day. Furthermore, users coming from social media sites, such as Facebook, Twitter, and blogs, will not ever hit the paywall, even if they are over their monthly limit. It’s understandable why the
Times
would want to allow these exceptions. It offers free articles every month so that potential customers can try out the paper. It accommodates users coming from Google search because these readers would not otherwise have come to the site, so they do not represent lost business. And it does not exclude readers coming from blogs and social networks because it wants its stories to be circulated; people blogging and tweeting its articles is part of what makes the
Times
the newspaper of record, which increases demand for the
Times
.

All these official loopholes can be used to access a particular article without a subscription. For example, if you want to read an article and get stymied by the paywall, search Google or Twitter for the headline and you will breeze right on through to the content. But as the writers of the code above discovered in March 2011, the porous paywall goes further: it allows users who really want to do so to disable all access restrictions semi-permanently. The paywall is not
really
a mechanism for enforcing copyright. Rather, it is a mechanism for sorting and price discrimination.

Copyright is a monopoly, and the basic account of monopoly entails some economic inefficiency—what economists call deadweight loss. Suppose I am a widget monopolist, and after the fixed costs of getting my business running, designing the widget, and setting up my factory, it costs me $5 to produce each additional widget. The marginal cost of a widget is $5. If I am to cover my fixed costs and earn a profit, I
must
charge more than $5 per widget. If I charge less than $5 per widget, I would lose money on every widget. If I charge exactly $5 for a widget, then I am not making any profit on that widget, but I am also not covering my fixed costs, and I will go out of business. And, crucially, I maximize my profits not by just charging enough to cover my fixed costs, but by charging more, even if that means I sell fewer widgets.

Suppose that I maximize my profits by selling widgets at $8 each. Some consumers might be unwilling to buy at that price—but they may be willing to pay $6. This is a problem. Economic efficiency demands that every consumer willing to pay the marginal cost of a good consume it. The marginal cost of the widget is only $5, and the customer is willing to pay $6, which means that there are gains from trade—but the trade doesn’t happen. This means that we forgo $1 worth of gains between us, the difference between $6 and $5, and that is inefficient. Add up all the losses from all the customers willing to pay more than $5 but less than $8 per widget and you have the total deadweight loss of monopoly.

The deadweight loss is also potentially a forgone profit opportunity. What if I could charge $8 to everyone willing to pay at least that much, but
also
charge a lower price, still above $5, to everyone else? I would make the same amount of profit as before from the customers willing to pay $8, but I would also make some additional profit from the other customers. And consumers as a whole would be better off—those who are willing to pay $8 or more would be no worse off, but at least some other customers would now benefit from widgets they otherwise would not have consumed. Engaging in this
price discrimination
makes total welfare—the sum of the benefits minus costs to everyone—increase.
⁴

Creative content is just like a widget, only with a much lower marginal cost. Once a news story has been written, the cost of reproducing it in the Internet age is so low as to be negligible. Call it zero. This means that there are forgone gains from trade if
any
consumer who values the content
at all
is excluded from consuming it. Charging any price above zero to everyone means that there are some profits still left on the table. If the content creator can sort consumers into different groups based on willingness to pay and charge each group an appropriate price, it can earn higher profits and usually improve economic efficiency at the same time.

This is what the
Times’
porous paywall accomplishes: It induces some customers—those who derive enough value from the articles, enjoy affiliating with the
Times
, or don’t want to be bothered to circumvent the paywall—to pay what the
Times
defines as “full price.” Meanwhile, no other users really get excluded. Everyone else can read the articles they find through search or social networks, and if they really want to read everything, they can. Meanwhile, these users still get served ads, so the
Times
still makes money from their visits, and they are of course more likely to link to the
Times
in their own blog posts and tweets.

More than a year after its launch, to the surprise of many observers, the
Times’
paywall appears to be a great success. Back when the paywall launched, Reuters blogger Felix Salmon wrote:

What does all this mean for the New York Times Company? I can’t see how it’s good. The paywall is certainly being set high enough that a lot of regular readers will
not
subscribe. These are readers who would normally link to the NYT from their blogs, who would tweet NYT articles, who would post those articles on Facebook, and so on. As a result, not only will traffic from these readers decline, but so will all their referral traffic, too. The NYT makes more than $300 million a year in digital ad revenue, so even a modest decline in pageviews, relative to what the site could have generated sans paywall, can mean many millions of dollars foregone. On top of that, the paywall itself cost somewhere over $40 million to develop.

    Against all that, how much revenue will the paywall bring in? A very large number of the paper’s most loyal readers are already print subscribers, and get access to the website at no extra cost. So the new revenues from the paywall will only come from people who read the website a lot but who don’t subscribe in print.
⁵

One year later, the
Times
reported 454,000 paid digital-only subscribers, and most recently, in July 2012, the company reported 509,000 of them.
⁶
Subscribers pay $15 to $20 every four weeks, depending on whether they get tablet access, so at an average of, say, $17 per subscriber for 13 four-week billing periods per year, that is a rate of over $112 million of revenue per year from subscriptions.

And what about the loss of advertising revenue? In the first fiscal quarter of 2011, before the paywall was introduced, the company’s News Media Group, which includes nytimes.com, posted digital advertising revenues of $53.9 million.
⁷
In the second quarter of 2012, the most recent period for which data are available, digital advertising revenues in the News Media Group were $52.6 million.
⁸
Annualized, this represents a decline of around $5 million in digital advertising revenues against the estimated $112 million in additional subscription revenues.

As early as July 2011, Salmon had to admit that “the NYT paywall is working.”
⁹
This is not to say that the New York Times Company is now free of financial challenges; the newspaper industry faces more competition than ever from blogs and social media, and there is of course the recession. But all things considered, the
Times’
porous paywall gambit has paid off.

Why was the
Times’
success such a surprise? When people evaluate content and copyright issues, it is common for them to implicitly or explicitly rely on one of two extreme viewpoints. The first extreme view, held most prominently by the record labels, is to approach the content business like business in the world of atoms—to ignore the unique properties of informational goods. There is an assumption that you can’t make money by giving your product away, and that property must be defended. Consequently, those who hold this view see people who appropriate content without paying as on a level with criminals. The opposing extreme view, popular among the digerati, is embodied in Stewart Brand’s iconic phrase, “Information wants to be free.” According to this view, any sort of paywall or access restriction to content is doomed to failure because of the unique properties of the world of bits.

At root, the first view regards enforcement as something that is or should be accounted for as free. But to the contrary, copyright enforcement is costly. Indeed, the cost of deterring or preventing an infringement has skyrocketed in recent decades. For example, an online newspaper could adopt a strict paywall with ease, but it would be costly for it to police the Internet to ensure that bloggers do not quote its articles at length. Even if news sites were to engage in this kind of enforcement—and the record labels do something similar in their policing of YouTube—it would be virtually impossible to prevent legitimate readers from emailing copies of articles to their non-subscribing friends. As the cost of enforcement rises, it is socially efficient to aim to deter fewer instances of infringement. Indeed, in many cases it is also privately rational to do the same, at least if the offended party is the one paying the cost of enforcement. An extreme version of this argument is easy to follow: if it ever became literally impossible to deter any instances of copyright infringement, then it would be foolish to spend resources trying to deter them.

The logic of costly deterrence is amplified by the fact that in the content industry, there are often few benefits to deterring infringement. When you prevent a teenager from illegally downloading your album, you rarely gain a sale of the album; the teenager can always listen to something else. Furthermore, you lose a potential fan who might come to your concert, tell his friends about your music, or buy your T-shirt. The world of bits really is different from the world of atoms. Informational goods are non-rivalrous, so rights enforcement for them is less beneficial than it is for normal, rivalrous goods.