There are compelling reasons for using thin-client technology to deliver applications to consumers. With thin-client technology or a Web-based user interface (UI) to a server back end, consumers could pay a monthly fee for their applications that would include access to the applications, free updates, data storage, and tech support. Online gaming, with such popular games as Diablo II, could be a big application service provider (ASP) market, too. Already many online gaming companies have agreements with ISPs.

However, for equally compelling reasons, the consumer market is limited. First, at the practical level, people doing their home accounting or writing their novels don't want to lose data or time because they lose an Internet connection. Second, per-minute Internet charges (where applicable) make the cost of using an application appear higher because the clock is ticking while you're using the application.

Third, at least in the United States, the concept is culturally wrong. I won't mention any specific hot-button issues, but currently many of the biggest cultural debates in the United States concern a struggle between individual rights and the public good--and the individual rights side is winning. It's one thing for businesses to implement server-based applications from the top down, saying "we’re going to do it this way" and let end-users discover that their applications work at least as well with thin-client technology as they did when installed locally. It's another thing for consumers to voluntarily cede control of their applications by buying them as a service (one reason why I'm skeptical about the Microsoft .NET initiative). To acquire personal customers, ASPs are going to have to frame the service in terms of personal empowerment for the consumer.

A fourth reason that I think the consumer market is limited involves the market for Internet appliances, which are a natural for ASP services. Bryan Ma, an IDC analyst, estimates that the worldwide market for information appliances will exceed 89 million units by 2004. That sounds like a lot, but his estimate is worldwide, and the 1998 US census estimate for number of households in the United States alone is more than 100 million. In other words, 89 million Internet appliances is a long way from every household in the developed world having its own Internet terminal.

Thus, I see consumer use of ASPs much more likely when the applications are simple transaction-based stateless applications where the risk of data loss is minimal or when they are applications the consumers can't easily get otherwise. On the transaction side, Wyse Technology's Blazer technology is interesting. Blazer will let consumers scan local businesses, searching by item keywords and locations, to easily find goods and services that they need while in an unfamiliar city. (This kind of thing would be extremely useful at a convention center.) The catch, of course, is that this kind of kiosk would be more popular if it were free.

In terms of otherwise-difficult-to-get applications, students and business people--outside the context of their businesses--could be markets. Personal training could also be a big market. For example, a graphic artist who needs to boost her computer skills could buy access to Adobe Photoshop rather than buying the application (possibly requiring a more powerful computer than she already has or spending money for an application she won't use at home). Another possibility is business people who need an application, such as videoconferencing to hold a meeting, but whose company doesn't already subscribe to the service or have such an application.

The bottom line is that I agree that a consumer market exists for ASP services, but I’m betting that it won’t supersede locally installed productivity applications. Rather, the consumer market will be for simple transaction applications, where the risk of data loss is low and the need for personal control minimal; unusual applications not part of the normal at-home suite; or personal training applications.