Business models: usage-based pricing.

Every summer, Microsoft's top management invites its Wall Street camp followers to a high-level forecast of the company's upcoming fiscal year. Major markets for Office and other cash cows have become saturated, chairman Bill regularly points out; upgrade rates are declining, and the company's historically high growth rates are bound to collapse.

And every year the analysts chuckle knowingly, because they've heard this "saturation" tale before. Miraculously, the slowdown never seems to happen. A year passes, and Microsoft ends up reporting another blow-out year that shatters all previous sales records. "We knew you were goofing," the analysts say fondly.

In the long run, however, Microsoft's doom-and-gloom predictions are almost certainly right and the Wall Street cynics are almost certainly wrong. At least in the U.S., the software industry really is running out of fresh customers. Most home PC owners these days buy relatively little software beyond what's bundled with new machines, and the business segment has become almost entirely a replacement market--hardly a hot growth opportunity for software products that don't rust or wear out.

So where will software companies get their new revenue growth? One answer that looks especially promising these days is usage-based pricing. Microsoft itself has been floating a proposal to let its biggest corporate customers adopt a mainframe-like "subscription" plan that provides automatic updates in return for a fixed annual fee. KnowledgePoint, a small developer of human resources titles, has been experimenting with a service that delivers customized job descriptions over the Internet for $8 each. Intuit continues to invest heavily in creating a Web-based financial services network. And the online gaming companies have proven there's at least a modest market for pay-for-play entertainment software.

All of these experiments reflect a fundamental change in the relationship between software developers and their customers. Instead of selling a fixed-price copy of a product, developers with usage-based pricing models collect payment for the utility value of the software--a job description, game play, content, financial services, and the like. As long as the software continues to deliver that utility value, the cash register should keep ringing.

At least that's the theory. In fact, usage-based pricing is such a new notion that it's tough to find anyone with a track record of success. Part of the problem is customer psychology: PC software buyers have been conditioned to think of software as a shrink-wrapped commodity, and--as Intuit in particular has found--they don't automatically embrace software companies as service providers. At the same time, developers haven't shown much insight into the kind of ongoing content and services that keep the meter running for pay-as-you-go customers. Clearly, the transition to a new pricing model will take huge amounts of education and testing on both sides.

Still, we're convinced that the search for an alternative business and product model will become increasingly desperate as more markets become saturated and more customers lose interest in buying upgrades. As we approach that point, here are some of the key issues that are likely to emerge:

Will usage-based pricing cannibalize sales of full-priced copies?

Customers aren't dumb: If paying a few dollars for sporadic usage is the better deal, sales of full-priced copies will suffer. But usage pricing also encourages people to try new software, and some of those trials are bound to turn into sales of full-price units.

KnowledgePoint chief executive Michael Troy says retail sales of his job description software "spiked" after he began selling individual descriptions on the Web.

Another less-visible problem with usage-based pricing is revenue deferral. Instead of getting paid before the product is even installed, the developer collects over a much longer lifetime of use. The result may be smoother cash flow and longer product life cycles--but it's likely to take longer to recover upfront development and launch costs.

What about tech support? Installation and start-up questions almost always generate the biggest chunk of tech support calls, so there's a risk that people who pay for just a few hours of usage will push support costs through the roof. (That was the unhappy experience of companies that distributed try-before-you-buy software on encrypted

CD-ROM disks a few years back.) However, if the software runs on a remote server and isn't downloaded to the user's PC, the cost of support probably drops substantially.

How will customers pay for usage? Right now, processing a so-called "microtransaction"--an electronic payment under $10--is so costly that the developer may net less than 50% of the revenue that small usage fees generate. Until there's a good microtransaction solution for the Internet, we expect most usage pricing will be based on a sustained relationship with the customer--an open corporate billing account, an annual subscription, or perhaps a "membership club."

We also expect to see some aggregation of usage-based titles, perhaps through local ISPs. Customers might sign up for a game channel, a home productivity channel, or (for business users) a desktop publishing tools channel. And the developer would be paid for syndication rights, exactly the way TV networks and cable operators now pay for a movie or TV series. Obviously, that's a blue-sky scenario, but it does work perfectly well with usage-based pricing.

Is there a role for the reseller channel? Probably. We suspect most developers will use outsourcers to handle back room delivery and transaction processing for usage-based products (just as they are beginning to outsource electronic software distribution). We think most customers, especially corporate accounts, would be very comfortable with the notion of paying a CompUSA or a Software.net for software usage. In fact, one of the biggest obstacles to usage-based pricing today is the absence of a reseller channel and delivery infrastructure.