The economics of the NYT paywall

By Felix Salmon
January 20, 2010
Preston Austin has managed to squeeze the economics of a NYT-style paywall into one tweet, but it's compressed, so let me expand it into slightly more than 140 characters.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Preston Austin has managed to squeeze the economics of a NYT-style paywall into one tweet, but it’s compressed, so let me expand it into slightly more than 140 characters.

The way that it seems the NYT paywall is going to work, visitors to will have a free allowance of n articles per month. To read the n+1th article, they will have to pay a subscription fee F. After that, they can read as many articles as they like for the rest of the month.

If a visitor to normally reads N articles per month, then the key number in their mind will be N-n. If reading that number of articles is worth more to them than F, they’ll pay the fee. If on the other hand N-n is small, or perceived value-per-article is small, then they won’t pay. Specifically, if the average value to the reader of any given article is v, then they’ll pay the fee when v(N-n)>F.

It does get a bit more complicated than this. For one thing, there’s the day-of-the-month issue. If you run into the paywall towards the end of the month, that’s a clear signal from that you’re not a particularly heavy user and therefore not the kind of person who’s likely to get a lot of value out of subscribing. Or, to put it another way, it’s a signal that this month at least, N-n is going to be low. In general, the NYT is unlikely to make much money from people paying the full fee F to read just a few extra articles at the end of the month.

On the other hand, as Jim Surowiecki says, consumers tend to like bundling. From a purely economic perspective, there’s no point in subscribing to before you reach your limit of n free articles. But from a more behavioral perspective, it takes a weight off your mind when you know that you’re a subscriber: you don’t have a little voice in the back of your head asking “are you sure you need to read that?” before you click on any link to In general, the mere existence of the paywall will make life happier for subscribers than non-subscribers, who will always feel somewhat constrained in how they use the site. For some people, the peace of mind associated with being a subscriber will in and of itself be worth F, even if they don’t read n articles per month.

More generally, it’s pretty clear that v is higher for subscribers than it is for non-subscribers. The value of reading any given article lies not only in the content of that article, but also in the ability to read the non-article components of the web page, to be able to follow links from that page to related or unrelated stories on the rest of the site, and in general to feel at home while visiting the website as a whole.

In his defense of the the meter system, David Carr writes:

By building a metered system, the executives have installed a dial on the huge, heaving content machine of The New York Times. Access can be gradually ramped up or down depending on macro trends in the market.

The implication here is that when advertising is strong, access can be ramped up, and when advertising is weak, it can be restricted, in an attempt to maximize subscription revenues.

I suspect that the main dial here is going to be n rather than F: it’ll be much easier to change the number of articles that people can read for free than it will be to change the price of a monthly or annual subscription. In a way, the best of all possible worlds, from a revenue perspective, would be to launch the new system with great fanfare and a low n, as well as a low annual-to-monthly subscription ratio. Then, when the low-hanging fruit has been picked and you’ve locked in a decent number of subscribers, quietly dial n back to a very high number, so as to minimize the pain caused to non-subscribers, and maximize total advertising revenue.

On the other hand, the experience of the FT suggests that there’s a strong temptation to go the other way: it has been dialing down n to a very low level, as it becomes increasingly addicted to online subscription revenue. It’s a bit like the airline fallacy that charging to check bags increases revenues: it doesn’t, but executives, especially at public companies, have a tendency to look at short-term revenue line-items rather than the bigger picture.

Finally, as John Gapper notes, there’s another calculation going on here: at the margin, implementing an online paywall is a good way of preventing print subscribers from cancelling their subscriptions on the grounds that they can get the same content online for free. (Print subscribers to the NYT will get online access for free, rather than having to pay extra for that, as they do at the FT.) Rather than saving the cost of a print subscription, P, they will now only save P-F. That’s an argument for a high F, of course — but the higher you peg F, the harder it becomes to convert the millions of visitors to your website into paid subscribers.

The NYT is therefore going to have to work out how to maximize revenues in a highly complex and dynamic system: the level of F which maximizes online subscription revenues won’t be the same level of F which maximizes print subscription revenues. And the NYT will also have to think long and hard about how transparent it wants to be about the level of n.

My prediction is that when the NYT paywall arrives, n will be in the 15-20 range, and F will be set somewhere around $15/month and $99/year. What that will do for the NYT’s total revenues, however, I have no idea.

11 comments so far

Regarding the point in your antepenultimate paragraph, I would guess that most people who are willing to pay much of a P-F to get the hardcopy rather than (in addition to) the website are current subscribers; I think it’s likely that there are a lot of people willing to pay the current subscription fee, and a lot of people who don’t actually value it enough to justify the cost of printing and distribution, and not a whole lot of people in between.

Of course, whether I’m right or wrong about this, I’m mostly leaving this comment as an opportunity to use the word “antepenultimate”.

Posted by dWj | Report as abusive

Regarding the behavioral aspect, the “little voice in the back of your head asking “are you sure you need to read that?” before you click on any link to” It would seem that the NYTimes would have to be offering content that can not be found elsewhere in order for this question to be relevant. If the same information can be found in an article elsewhere, people will see the NYTimes headline and seek the same story from another news source.

Posted by DaveTheFirst | Report as abusive

Terrible idea. The New York Times is a great tool for improving literacy and educating people of all ages. To deny service to those who aren’t well-off, which this would do, would be depriving people of factual and well-written journalism. It would also steer people to the sites that remain free. Those sites are often aggregators like Drudge Report and Huffington Post – which rely on sensationalist blogs and tend to ignore the facts – or wire/24 hour news services which have very short summary articles. In this economy people need access to the New York Times as it is a FREE voice of moderation in this crazy world.

Posted by ClassOf07 | Report as abusive

I like the idea, though we’ll have to see how well it works. I like the idea because media’s response to the economic issues has been to cut, cut, cut. The notable exceptions are the Times and The Economist because they have committed to producing more content that people want to read. If the test of media is that it be useful and thus attractive to you, then they need to produce better content and better content would be good.

Posted by jomiku | Report as abusive

I’m glad you’ve picked up on the metering point, because you seemed to have missed it altogether on your first post, which isn’t like you. I’ve always been bewildered when bloggers and the like say that a paywall means you’ll never get any links or google hits. I’m not aware of a single news site that doesn’t let you read a damn thing without paying; you always get either a limited number of articles free (a system which, my understanding is, the FT has made work) or you show the first few paragraphs of a story, or both.

The NYT clearly know this, and they intend to set n, in your example, to a number that is higher than the number of stories the average web-browser consumes in a month. You say “the key number in their mind will be N-n,” but you don’t seem to acknowledge the fact that for most users N-n will almost certainly be negative. The “long tail” applies to users as well as content: 90% of users account for a small proportion of the traffic. To these users, this change isn’t even going to be noticeable.

I think David Carr makes it completely clear that it’s the number of articles, not the cost, which they see as the variable part of the “meter.” I take your point that the FT has kept n low, but the FT is always going to have an easier time getting people to pay for a subscription, because people don’t pay with their own money, but their company’s. The NYT will need to, and I think will, keep the meter higher, and probably reintroduce some Times Select-style content which is never free, probably the stuff like property and style which is basically just read by rich folks.

As to the end-of-the-month stuff: what? Unless I’ve missed it, I’m sure the NYT hasn’t linked its monthly subs directly to the calendar month. Most monthly-billed services, like mobile phone plans, have a variable billing date.

You’re usually so sharp, Felix, but I can’t help but feel you’re so eager to condemn this move you’re scratching around for reasons to do so.

Posted by ravcasleygera | Report as abusive

One other unrelated point, which isn’t your problem, but I’d be grateful if you could pass it onto the relevants. What on earth is the point of providing a way for people to sign in to the site through another account, like Google or Facebook; if, having done so, you still have to provide your name, a username, a password, and basically all the information you’d have to provide if you simply registered properly? The point of such services is to save you from having to enter precisely that information. Your implementation should automatically create an account attached to whatever external account the person used to sign in, with details like username set automatically and invisibly.

Also, OpenID should be added to the list of sign-in methods.

Like I say, I realise this isn’t your problem; but you’ve never hesitated to point out the problems at other news sites’ websites, so…!

Posted by ravcasleygera | Report as abusive

Rav, the sign-in problem is a known issue, and we’re in the process of fixing it. Apologies.

The assumption across the industry seems to be that n will be fixed. What if n is a variable depending on the time of the week, the time of the month, or even user behavior?

As you pointed out, NYT has an disincentive to charge users if they’re just grazing the paywall, especially at the end of a month, because the success rate will be very low and they risk losing that reader entirely. Instead of adjusting n across the board, they can simply adjust n for readers of a particular bend/use pattern. The software exists to do this.

Disclosure: The company that I work for, Atypon, develops and supports the application that the Financial Times uses to power’s paywall and digital commerce. New York Times Digital is also a customer.

Posted by kevinacohn | Report as abusive

One of the best reactions to news of the planned paywall.

I’ve written up a much expanded economic analysis of the paywall on my blog nalysis-of-nyt-paywall/

Many many more things need to be taken into account to truly understand how a reader will interact with a metered payment plan. For starters, the value of an article to a reader is not constant, AND the reader has to make a prediction about the future value of article’s that haven’t been read yet.

Posted by albertsun | Report as abusive

Finally, a serious analysis for a serious change in the revenue model. My hopes is that the subscription model will result in even more in-depth articles in order to draw in serious readers, a la The Economist. I’ve felt that over the last few years, articles were getting a bit sparse.

Now to seal the deal, they need to start paying us for recommending articles to our friends. I’ll admit, this isn’t my idea, I got it from a 5/19/10 article at
that talked about how to make the information subscription revenue model work by using freemiums, social networking, and referrals. (I’ll put the direct link at the bottom of this comment.)

I’m always referring articles I like to friends who I know will be interested. If those friends read, pay, and/or subscribe, I’d like to get my cut.

Felix, keep up the good work. I’d pay for your thoughts any day. (But I enjoy getting them for free as well!)

Here’s that direct link I promised:
http://e-marketingforsensiblefolk.blogsp model-work.html


/2010/05/making-info-subscription-model- work.html

Posted by whystop | Report as abusive

Nice prediction on cost. The only place you were off is that there is no annual subscription.

Today, in their quarterly posting, nytimes reported 100,000 subscribers, which if it held up would be more than $5M per year (minus loss in advertising).

Posted by pjwst6 | Report as abusive
Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see