The Community for Technology Leaders

Turnpikes or Freeways?

Fred , EMC •

Pages: pp. 4-6

Igrew up in New Jersey and returned there after school, so I'm used to the concept of toll roads, such as the NJ Turnpike and Garden State Parkway, as major thoroughfares. But I went to graduate school in California, where the freeway is king: few roads there require payment simply for their use, except for those with bridges.

Maybe it's no coincidence, then, that California is known for another attribute: its awful traffic jams. These free highways are usually packed with cars. Authorities have looked into solutions such as charging to open high-occupancy vehicle (HOV) lanes to solo drivers. (There are also exceptions to the rule, as some roads have had tolls from the time they were built, often as a way to recover the costs of construction.)

Rarely, however, does a freeway become a turnpike. People who are used to a road being financed at public expense (usually through gas taxes, so these expenses aren't entirely decoupled from the people who use them) take badly to the suggestion of instituting charges for each use, even during limited hours as a form of congestion pricing. This doesn't keep governments from periodically considering converting a free road into a toll road or charging extra for a subset such as the HOV lanes. I heard news of the latest such discussion today, to provide limited but high-cost access to the Lincoln Tunnel entering New York City. More than likely, this too shall pass. I mean, it will go away.

The Online Media Dilemma

On the Internet, the rules are different. Services transition between free and paid, sometimes multiple times. Naturally, people don't complain about something that used to cost money being offered for free, but they will grumble and perhaps take to the streets if it goes the other way. Witness various Facebook groups in which members proclaim they will drop it like a hot potato if they have to start paying to use it.

These examples aren't new by any means. Consider Slate, the online magazine founded in 1996 by Michael Kinsley with sponsorship from MSN ( In 1998, it started charging for content but switched back to free access supported by advertising revenue roughly a year later. Similarly, TheNew York Times has been a force on the Web since the mid-1990s, but it started charging for much of its content in 2005 ( Print subscribers as well as academics were granted access at no extra charge. Two years later, it dropped the fees with the expectation that increased advertising revenue would make up for the difference. In early 2010, it announced that in 2011, it will once again charge for access — this time, allowing all users access some number of times each month before they must pay for any subsequent access. Once again, print subscribers (such as myself) will get online access at no extra cost. Free access to print subscribers can be important: personally, I read the Times both in print and online; if it started charging me extra as a subscriber to use its Web site, I would drop the more expensive print subscription and stick to online.

Other models exist, of course. The IEEE Computer Society, which publishes this magazine, makes content available in print, online, or (usually) both. Subscribers to the print edition don't automatically get access to it online — that is, a separate subscription — but it's cheaper to get bundled print and online access than to buy both separately. I similarly subscribe to Consumer Reports, but I don't kick in a separate fee to use their Web site, primarily because I find other free sources for similar content. (Perhaps if I'm going to pay them for something, I should stop paying for print and subscribe to online access instead, but that's another matter.)

Still another model is to ask for contributions, rather than requiring them. Public radio and television take this approach, and I willingly support my local stations in appreciation for their excellent reporting and other content. According to a recent piece on National Public Radio, a newspaper in Florida now posts two links at the bottom of each online story, one to subscribe and one to donate.

Time will tell whether the Times' decision to return once more to a paid online presence will be any stickier than the last time. Paid access has worked well for one of its main competitors, the Wall Street Journal — but, of course, the readership of the Journal is a somewhat different demographic.

Differentiated Service

I began this column with a discussion of differentiated service on highways, having people pay extra to use the "fast lane." The Internet has some fast lanes too.

In fact, I remember attending an "Outrageous Opinions" session at a "Hot Topics in Operating Systems" workshop in the mid-1990s and presenting something I didn't think was outrageous at all: the idea that all the free services we were already accustomed to on the Web would at worst go away and at best wither under their own weight. I really was convinced that the growth rate of the user community would outpace the revenue that sites could derive from providing their services, and they would get overloaded. Then, as for-pay competitors stepped in to offer similar services to much more controlled user communities, people would pay to get faster responses.

When it comes to core Internet services such as search and email, I'm glad to have been proven wrong. Various service providers such as Yahoo, Google, and Microsoft have been able to provide these services with extremely high performance to enormous numbers of users. Between their extensive data and fast response times, there's no incentive for users to jump to another system; the only hope for differentiation comes from providing, for instance, better-targeted results.

However, the perils of a high growth rate do arise in other contexts. The past decade or two saw a steady transition from metered service (pay by the minute or the Kbyte) to all-you-can-eat service (pay one flat rate for unlimited usage). Some services, such as mobile phones, switch silently from one mode to another depending on location: I remember a grad student bemoaning a US$1,000 bill after having left his iPhone on while in Asia. But more insidious issues arise when you must choose between usage caps and unlimited usage and when "unlimited" usage actually is anything but.

The first of these issues comes from payment plans. I have a USB-based modem to get broadband on my laptop. I was faced with a choice between one price for what seemed like a fairly small usage limit and another, higher price for a higher limit I could be confident I couldn't possibly exceed. I opted for the higher limit, even though I rarely come close to the lower one, because I don't like surprises. Apparently I'm not alone: a recent study by the BBC ( showed how many people overpay for higher limits (including unlimited usage) they don't use. In fact, 75 percent of customers download less than 20 Gbytes per month, and half download less than 10 Gbytes per month.

The bigger news is the shift from completely unlimited usage to usage caps: irrelevant to most, but a shift from freeway to turnpike for the heaviest users. Comcast has instituted a 250-Gbyte cap, which it claims covers 99 percent of its users and which seems pretty reasonable to me. But there's a fear that they're preparing to move back from flat-rate to metered pricing, even if people hate it ( According to the BBC article, this might actually reduce costs for most people, so is it that bad? I suspect the people most affected will be those sharing data that, maybe, they shouldn't be. Others might be movie buffs who just can't get enough online streaming video. Part of me says these people should, in fact, pay their fair share of the cost or quit clogging the networks for the rest of us. And part of me is worried that, in a moment of weakness, I'll become like that grad student caught off guard by how much his data usage was costing him — I'm ready to buy my insurance against that by opting for the highest possible limits up front. No one likes to pay as they go.


The opinions expressed in this column are my personal opinions. I speak neither for my employer nor for IEEE Internet Computing in this regard, and any errors or omissions are my own. I thank Allison Douglis for her helpful feedback.

98 ms
(Ver 3.x)