The New York Times announced this week that starting next year it will erect a paywall around its content. Visitors to the site will be able to view a set number of articles for free each month, but after they reach that limit they will be invited to subscribe before they can view any more.
Consensus among the digerati seems to be that this will never work. (See here and here.) That’s an easy position to take since micropayments and TimesSelect-type paywalls have failed in the past. But is it really that clear? It seems to me that what the NYT is trying to do is not so impossible.
If the NYT website’s readership is like anything else, there’s probably a power law at work. A small minority of readers make up a sizable percentage of pages read. Put another way, there’s probably a small minority of users that read a considerable amount more articles than the average user. They need to set the number of monthly free articles high enough so that the bottom (say) 98% of readers never even notice that there’s a paywall, but the top 2%, which are presumably devoted NYT fans would be affected.
Now, that’s easy to say, but figuring out the balance might be very tricky. Whatever it does, the NYT doesn’t want to affect the ad-revenue-generating traffic it’s getting. How does it do that? First, it seems to me, is that the NYT has to be realistic about how many people it can get to subscribe. It’s going to be a tiny, tiny number, but that’s money it’s leaving on the table right now. So the number of monthly free articles needs to be in the 100+ range, not the Financial Times‘ 10 a month. Second, it needs to set a reasonable subscription price. The top readers are probably devoted fans, but that doesn’t mean they’ll pay anything, and if they don’t the NYT will lose not only the subscription, but the ad revenue it now generates from those folks.
So that’s all to say that while it will no doubt take thoughtful tinkering to strike the right balance, there’s no economic reason the NYT could not segment its customers in this way. Am I wrong?






