Running the numbers on the NYT paywall

By Felix Salmon
January 21, 2010
John Gapper's column on the NYT paywall makes me realize that a lot of people fail to appreciate exactly what's at stake here. Gapper seems to think that online subscription revenues can make newspapers profitable again; they can't. In fact, insofar as the paywall makes any sense at all, it does so only as a tool to boost print subscriptions.

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

Apologies for being something of a one-trick pony today, but reading John Gapper’s column on the NYT paywall makes me realize that a lot of people fail to appreciate exactly what’s at stake here. Gapper seems to think that online subscription revenues can make newspapers profitable again; they can’t. In fact, insofar as the paywall makes any sense at all, it does so only as a tool to boost print subscriptions.

Gapper notes that the Guardian’s parent company lost $93 million in the last fiscal year, despite having a website attracting 35 million unique visitors globally, and 13 million domestically. He reckons that charging some of those subscribers could make the Guardian’s problems disappear:

Outsell, a research group, reported this week that only 6 per cent of US online readers say they would pay online news sites if they charged.

If we are to take the figure at face value (which I don’t think we should), then The Guardian could get 2.1m people to subscribe to it online, making it highly profitable at a stroke.

The logical flaws here are huge. Let’s just count a few:

  1. Gapper is assuming that if 6% of online readers would end up subscribing to some site or other, that’s the same as saying that any given site can persuade 6% of its readers to subscribe. But that’s not the same thing at all.
  2. Gapper is taking a poll of US internet users, who are more likely to pay for things online than other nationalities, and extrapolating the numbers globally.
  3. Gapper is assuming that foreign readers (say, Guardian readers in India) are just as likely to subscribe as domestic readers (Guardian readers in the UK).
  4. Gapper is assuming no overlap at all between the 6% of readers who would pay online, and the percentage of readers who already pay for a newspaper subscription in print form. The NYT has already said that its print subscribers will get an online subscription for free, and I’m sure the Guardian would do likewise.
  5. Gapper is ignoring that putting up a paywall will always reduce advertising revenue, which means that in order for the new subscription revenue to make the newspaper “highly profitable at a stroke”, it would have to not only make up existing losses but also cover the drop in online advertising revenue.

Once you understand what Gapper can do to statistics like these, then, it becomes a bit easier to make sense of something like this:

Rates for online display ads have been falling steadily as competition has proliferated, with most sites now finding it hard to get more than $4 per 1,000 impressions on their pages (or $14m for the 3.5bn hits on all US newspaper sites monthly).

I have no idea where Gapper’s getting his $4 CPM figure from, but it’s clearly much closer to being a minimum than an average. Papers might find it hard to get more than that, but you can be sure that the NYT, for one, is succeeding all the same. In fact, the WSJ reports today that nytimes.com is pulling in $100 million in revenues annually — more than you might think the US newspaper industry as a whole was making, if you read Gapper too literally. Indeed, in the third quarter of 2009, the New York Times Company made $79 million from its internet businesses, of which $68 million came in advertising revenues. Newspaper advertising was $39 million — and that was in a very weak quarter. On an annual basis, I’d be surprised if nytimes.com didn’t make significantly more than the $100 million that the WSJ is talking about.

It’s also worth noting that Gapper has managed to confuse CPMs — the amount of money that an advertiser pays per 1,000 pageviews — with RPMs, or the amount of revenue that a publisher receives per 1,000 pageviews. There’s nearly always more than one ad unit per page, which means that RPMs are some multiple of CPMs, depending on how much of a newspaper’s inventory is sold.

Gapper, then, is systematically overestimating the upside of subscription revenues, while underestimating the magnitude of advertising revenues. Erick Schonfeld, by contrast, is much more realistic, concluding that total subscription revenues from nytimes.com would optimistically reach only $9 million per quarter, or $36 million per year. With the New York Times Company making the best part of $300 million a year from online advertising, it’s hard to see that the extra revenue boost would really be worth it.

The point here is that with the powerhouse nytimes.com site front and center, the New York Times Company as a whole is a major online media player, serving up billions of high-prestige pageviews and building strong relationships with every major online advertiser and media buyer in the country. Even under the most optimistic scenario, a majority of the NYT’s loyal readers will desert it when it moves to a paywall. And with those readers gone, media buyers are by no means guaranteed to stick around.

Gapper makes great play of the fact that websites can target ads more accurately when readers are registered, but you can’t target ads at readers who no longer exist. And the NYT is a mass-market general news publication: it’s not the kind of place where high-end business-to-business advertisers will pay $90 CPMs to reach C-suite executives. Or if it is, the numbers involved would be so small that they wouldn’t make a visible dent in its overall online advertising revenues.

What’s a realistic number for how many people will pay to subscribe to nytimes.com? David Carr says that the NYT wants to target “10 percent or so” of the 17 million current readers of nytimes.com. That’s 1.7 million people. Subtract the print subscribers who will get nytimes.com for free, and you’re left with 1 million, more or less. How many of those could you dare hope to persuade to subscribe? One third? Once again, just as Schonfeld did, we get to somewhere in the region of $35 million a year, assuming a subscription price of $100 each per year. For a company with annual revenues in the billions, the hit to the value of the brand alone has to be bigger than that.

There is one other dynamic at work, here, however, and that’s the price of the print subscription, which has proved to be surprisingly inelastic: David Carr in fact lauds the way in which “The Times has shown a great ability to leverage prices once they have custody of a consumer”. But pushing existing consumers to the limit of what they can pay only makes it that much harder to attract new ones.

The NYT aspires to be a national paper, which makes sense, since it either already has or is never going to get most New Yorkers as print subscribers. But the annual subscription rate, if you live at say 1600 Pennsylvania Avenue, is $769.60, and it’s really hard to get people to pay that kind of money, in the middle of a recession, for a newspaper they’ve lived happily without for all their lives, and which they can get online for free.

So I see the decision to implement a paywall not as an attempt to build a significant revenue stream for the website alongside ad revenues, but rather as an attempt to shore up — and maybe even increase — the print subscriber base. Even a Monday-Friday subscription, at $384.80 a year, brings in much more money than any online subscription ever will. Yes, it probably costs more than that to print and deliver the paper. But for the time being at least, print advertisers are still willing to pay top dollar for a full page in the physical New York Times. And so long as that’s the case, the NYT will do anything to keep its physical circulation numbers as high as it can — even, it seems, if that means dealing a serious blow to nytimes.com.

Update: Gapper responds with a 950-word blog comment. Good for him!

Comments
7 comments so far

Though the model you meticulously describe so well makes sense, it would have been better if you had simply described the model in mathematical terms for any given publication, since the NYT is one of the worst publications in the world. The NYT has repeatedly reinforced its poor reputation by confabulating news, taking a blatantly liberal slant, and hiding behind the despicable use of “undisclosed sources”, so no matter what “n” is, I would never pay them a dime. I read the NYT only as a last resort and much prefer the Economist, FT, or just about any other source of news. If the NYT did not exist, the world would be a better place. IMHO. Did I say the print edition of the NYT was a waste of trees?

Posted by netvet | Report as abusive

“But the annual subscription rate, if you live at say 1600 Pennsylvania Avenue, is $769.60″

Blimey. And they wonder why people don’t read newspapers any more. A Guardian sub will set you back just over £300. Of course, the Guardian loses money and the NYT makes money. But a subscription to the profit generating Sun would cost a third of the Guardian’s cost.

Posted by GingerYellow | Report as abusive

“Reuters: all NYT pay wall, all the time!”

I can’t blame you for taking this seriously though Felix: the success or not may determine future business models for the entire industry.

IMHO Schonfeld is on the money, or rather, lack of it.
NYT just isn’t that good. The % of people willing to pay will be astonishingly low, particularly as the charge can’t be expensed unlike the VAST majority of WSJ and FT subscriptions.
Once I hit my NYT limit I will simply use another source.

What you meta-analysis has shown once again is that statistics can be used to prove anything.
Your destruction of the Gapper assumptions is a wonderful example of how to look beyond the headlines, thanks.

Posted by TinyTim1 | Report as abusive

Dear Felix,

I forgive you for being a one-trick pony. I appreciate that you have spent the time and effort to respond at such length. So let me reply to one or two points:

“The logical flaws here are huge. Let’s just count a few:

Gapper is assuming that if 6% of online readers would end up subscribing to some site or other, that’s the same as saying that any given site can persuade 6% of its readers to subscribe. But that’s not the same thing at all.”

No, I’m not assuming that. You quote me as saying at the start of those two paragraphs: “If we are to take the figure at face value (which I don’t think we should)”. That is a warning not to take literally what you go on to take extremely literally, as if I am making some exact prediction.

I don’t know how many readers the Times could get to subscribe. I am simply making the point that surveys in which a large majority of readers say they would not subscribe are sometimes taken as evidence that a freemium model of some kind will not work. But actually, only a small proportion of readers need to become subscribers for it to be successful. That is all.

“Gapper is taking a poll of US internet users, who are more likely to pay for things online than other nationalities, and extrapolating the numbers globally.”

See above.

“Gapper is assuming that foreign readers (say, Guardian readers in India) are just as likely to subscribe as domestic readers (Guardian readers in the UK).”

See above.

“Gapper is assuming no overlap at all between the 6% of readers who would pay online, and the percentage of readers who already pay for a newspaper subscription in print form. The NYT has already said that its print subscribers will get an online subscription for free, and I’m sure the Guardian would do likewise.”

See above.

“Gapper is ignoring that putting up a paywall will always reduce advertising revenue, which means that in order for the new subscription revenue to make the newspaper “highly profitable at a stroke”, it would have to not only make up existing losses but also cover the drop in online advertising revenue.”

No, I’m not ignoring the fact that readership will probably drop. I argue elsewhere in the piece that this trade-off is not zero-sum and pointing out that some publishers have found that a core of paying readers can not only provide subscription revenues but are more valuable to advertisers.

“I have no idea where Gapper’s getting his $4 CPM figure from, but it’s clearly much closer to being a minimum than an average.”

I’m getting it from talking to ad networks and online publishers, who report that $3 to $4 CPM is about the rate that ad networks are now providing to publishers. Incidentally the New Business Models for News project at Cuny assumes CPMs for a hyperlocal online news project of between $1 and $12.

“It’s also worth noting that Gapper has managed to confuse CPMs — the amount of money that an advertiser pays per 1,000 pageviews — with RPMs, or the amount of revenue that a publisher receives per 1,000 pageviews. There’s nearly always more than one ad unit per page, which means that RPMs are some multiple of CPMs, depending on how much of a newspaper’s inventory is sold.”

No I haven’t. Sadly, I had a lot less space to express this than you have used to reply so I condensed it. However, as you say in your last sentence, it depends on how much inventory is sold. I was making a rough estimate – again based on talking to online publishers – that the trade-off between the number of ad placements you have on a page (more than one) and the percentage of inventory actually sold (less than 100 per cent) tends roughly to net out.

“And the NYT is a mass-market general news publication: it’s not the kind of place where high-end business-to-business advertisers will pay $90 CPMs to reach C-suite executives. Or if it is, the numbers involved would be so small that they wouldn’t make a visible dent in its overall online advertising revenues.”

The first sentence is true, the second is tendentious. The NYT and other general newspapers do not have a big business-to-business advertising base, and CPMs for some general news may be hard to raise substantially. However, a publication such as the NYT has many niche areas – science and health coverage, technology coverage etc, where CPMs could well be boosted.

“For a company with annual revenues in the billions, the hit to the value of the brand alone has to be bigger than that.”

You are confusing brand value with revenues. A high brand value may help a company to charge a higher margin, and therefore to increase revenues – Apple and Coca-Cola being good examples of companies that do so. But you have to charge something to gain the full benefit.

“So I see the decision to implement a paywall not as an attempt to build a significant revenue stream for the website alongside ad revenues, but rather as an attempt to shore up — and maybe even increase — the print subscriber base.”

I agree that is one of the motives – the NYT has perversely been encouraging print subscribers to cancel by offering its news free online. But it is sensible. Eventually, print will become obsolete and the NYT will become an entirely digital product but it makes sense to maximize cashflow from print to pay for the expensive conversion to digital.

Always a pleasure to debate.

John

Posted by johngapper | Report as abusive

It always amazes me that digital conversion equates to expensive. A college student can do this for next to nothing. It is when bespoke programmers hardwire your systems that you sit with an eternal and expensive problem.

Posted by Ghandiolfini | Report as abusive

Just to note that $93m is a bit too high for the Guardian loss – it was distorted by a non-trading writeoff and losses at Emap (net of the profits of Trader Media). The best starting point is the divisional loss of £38m, which was in turn affected by £8m cost of the site move (these numbers from memory sorry). So the loss is more like $45m – I don’t think it makes much qualitative difference, but the $93m number makes it look like the Guardian could never possibly make a profit again, whereas the lower number gives the (I think more accurate) picture that it’s probably small-negative over the cycle.

Posted by dsquared | Report as abusive

In the end, the failure of paywalls is a distraction. It’s just another experimental drug for a patient that died years ago. The reality is that all mass media/mass marketing business models – from magazines to broadcast TV – are no longer viable, and newspapers are just the first to go: http://digitalthinking.posterous.com/the -great-paywalls-of-news-corp

Posted by digitalmem | 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 http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/