Let's face it: Flat-rate online pricing is insupportable

The recent trials of American Online, Inc. (AOL) have pointed out the fatal flaws in the way Internet access is priced today. Egged on by Wall Street, which foresaw that AOL would lose business to Internet service providers charging $19.95 per month for unlimited access, Chairman Steve Case abandoned his usage-based pricing model for the same flat rate ISPs were charging.

The good news, at least for AOL's stock price, was that the company signed up hoards of new users. The bad news is all over the newspapers: busy signals, lawsuits and the potential of hundreds of millions of dollars in customer refunds.

To anyone familiar with the communications business, these events were fairly predictable. Economics 101 teaches us that we can plot the number of customers that will buy a product at every given price point. Normally, this is a fairly smooth curve, with volumes gradually increasing as price decreases.

But this is not true in the communications business. Rather than a curve, you get what I would describe as a cliff. Once price decreases to the magic price point, demand basically goes infinite. For Internet access, this price point has turned out to be a flat rate of $20 per month, whether you are online 20 minutes or 20 hours a day.

What we are witnessing is a finite resource - bandwidth - being priced as if it were a commodity. As the AOL experience shows, the economics of this situation are insupportable.

Before the Internet really took off, ISPs had to invest tens of millions of dollars in infrastructure before they knew if there would be much of a return on that investment. But they knew when the cliff effect kicked in, they would need that capacity. The trick was to invest enough to reach critical mass, but not so much as to break the bank. However, when the magic price point was reached, the ISPs were almost guaranteed not to have enough capacity. That is what has happened with AOL.

Flat-rate pricing may be a great marketing tool, but it does not take into account another tenet of the communications business: If a service provider must size its infrastructure to accommodate the amount of traffic, pricing must reflect this or the provider will eventually go out of business.

In order to adapt to its recent growth, for example, AOL has announced that it will need to spend at least $100 million, on top of a $250 million upgrade already under way, just to handle the additional traffic. And even this amount may turn out to be too little.

Until the breakup of AT&T in 1984, residential subscribers enjoyed flat-rate local calling, which was subsidized by business customers. But that pricing is slowly disappearing and may not be applicable to the Internet anyway. Long-distance telephone service has always been priced by a combination of time of day, usage and distance. However, this approach is not really appropriate for the Internet, because there is no meaningful way to measure distance.

Perhaps a better approach would be to follow the airline model and create classes of service. Those in "first class" would pay more but would get first shot at the available bandwidth; users in "coach" would pay less but experience more busy signals, at least during peak traffic times. This strategy could work for AOL.

Over time, a lot ISPs and their $20 flat rates are going to fall by the wayside. Once this happens, I doubt that AT&T or MCI Communications Corp., the natural heirs to this customer base, are going to replicate a pricing model as doomed as flat rates. Everybody should back off, let AOL admit it made a mistake and allow the company to reinstitute its former pricing structure where people paid for what they used.

Fong is a general partner of Mayfield Fund, a venture capital firm based in Menlo Park, Calif. He can be reached via the Internet at kfong@mayfield.com.