Arnold Kling     Essays | Short Book Reviews | Favorite Links | Internet Bubble Monitor | Home
Life After the Dumb Buyer's Death

"Arguing in My Spare Time" No. 3.09

Arnold Kling

March 18, 2000

The future of the software industry finally is starting to come into focus. It will be much less monopolistic and much more constrained in its profits than what we have been used to seeing.

The key source of profits in the software industry in the past twenty years has been the dumb buyer. A dumb buyer is an organization with a large software budget and a decision-making process in which network effects are influential and political considerations overrule cost-benefit calculus.

Today, the software marketing model consists of direct sales of products to dumb buyers. The dumb buyer chooses operating system software, calendar-management software, and other products. Many dumb buyers also develop their own proprietary applications, often using components that they purchase at high prices, such as database software and application server software.

The new element in the picture is the Application Service Provider (ASP). An ASP provides applications over the Internet, rather than as software installed on a desktop or internal network. My prediction is that over the next five to ten years the ASP will emerge as a layer in between the developers of fundamental software tools on the one hand and the business users/analysts who determine the requirements for business applications on the other.

ASPs have the potential to replace massive purchases of generic software with Internet-based versions of these applications. Once companies can use an ASP for calendars, and email, they will be less willing to pay for per-seat desktop licenses of these products.

However, the really interesting opportunities for ASPs involve replacing proprietary software development. Instead of building an inventory management system for your company, you can use an ASP.

Replacing proprietary software is a very different model than replacing mass-market software such as email or spreadsheets. The difference has important implications for the ASP industry.

  1. ASP's will concentrate on niches, rather than sell to every industry. There is too much industry-specific terminology and business logic for an ASP to be completely general. Some ASP's may focus on the real estate industry. Some may focus on dentist offices. Some may specialize in banks. Indeed, the market could divide further, with some ASP's specializing in small banks while others focus on large banks.

  2. ASP's may tend to be owned, wholly or in part, by industry trade groups or consortia. In this respect, they will tend to look like "revolutionary ventures" in Amar Bhide's taxonomy. (Bhide's book, "The Origin and Evolution of New Busineses," is a remarkably insightful analysis of the process of business innovation. Although his research pre-dates the Web, I found his book more useful for understanding the economics of innovation on the Internet than any other book on the market.) If you were starting an ASP for an industry, a wise marketing strategy would be to get the endorsement of a trade group and/or some of the top companies in that industry. A useful tactic for getting that endorsement will be giving an equity stake to the endorsing group.

  3. ASP's will act as system integrators. They will develop some components, but they may license components from other companies.

  4. The niche ASP business will be highly labor-intensive. Large enterprises will expect the same level of customization and service that now is provided by their internal staffs. Small proprietors will want plenty of customer support and training. Everyone will want the ASPs to develop interfaces to legacy systems. The "sell-it-as-is" model that has provided enormous profit margins for the Microsofts and Oracles of the world is not going to prevail for ASPs.

  5. Pricing of ASP services will be highly unsettled. The conventional wisdom is that companies will "rent" their applications. But what will be the unit of rental? Per use? Per week? Per year?

  6. Bundling also will be an interesting issue. It is possible that the niche ASP will adopt a model like the one I proposed two years ago in "Software Clubs." There will be a natural tendency for the niche ASP to want to offer a full-service bundle to its customers. Will buyers go along with bundling, or will a doctor's office want to choose the payroll solution of one ASP and the billing solution of another?

Microsoft has sent some signals that indicate that it is considering trying to morph into an ASP. This would be a challenge. I am expecting that niche ASPs will have to work very closely with particular industries. Microsoft is used to providing little support or customization for its products. Instead, they are used to serving a mass market with low prices. Also, the burden for training and support typically falls on the internal Information Systems staff of the enterprise that buys Microsoft software. If Microsoft becomes an ASP, they will be expected to provide direct customer support.

A more plausible scenario is one in which Microsoft supplies basic components to ASP's. They might supply the web server and the database software, for example.

However, a challenge for Microsoft as a supplier of components is that ASP's are going to be smart buyers. If Unix-based solutions work best, then the ASP is not going to have to worry about some bureaucrat who wants to simplify life by becoming an all-Windows shop. ASP's also are going to be wary of getting locked into monopoly suppliers.

Another weakness for Microsoft as an ASP may be its lack of political dexterity. I thought of this when it announced on March 16th an alliance with Freddie Mac and some major mortgage lenders. They plan to develop a more efficient mortgage process for the HomeAdvisor service that Microsoft now plans to spin off. Other potential ASP's for the mortgage industry include, with Fannie Mae on its board, and MERS, the service provider most aligned with the Mortgage Bankers Association.

I should mention that because of previous affiliations, I am a shareholder in both Freddie Mac and Therefore, do not quote me on the following.

Microsoft and Freddie Mac have each gotten themselves into positions where they need a victory by Governor Bush in November to gain respite from hostile political appointees. They also have yet to demonstrate the sensitivity required to reassure an industry coalition. A few years ago, Bill Gates' idea of tact was to say, "Banks are dinosaurs. . .just kidding!" You probably are going to have trouble positioning yourself as an ASP if you cannot stop fantasizing out loud about turning your customers into road kill.

Again, about the preceding paragraph: don't quote me.

As the dumb buyer gradually becomes extinct, the software industry that I foresee will have the following players:

  1. Business buyers. These will be technically-savvy business users, or business-savvy technical executives.

  2. Niche ASPs. Each industry will have its own specialized providers of Internet-based applications. My guess is that most industries will support several ASPs, each with somewhat differentiated products. The niche ASPs will be guided, and in some cases owned, by coalitions of leading companies within their industry.

  3. Developers of software tools. These companies will provide generic infrastructure, such as database management software, web servers, communication tools for information applicances, and so forth. However, for most of these products, their customers will be ASP's rather than businesses. The ASP's will not be as susceptible to marketing pressure. The dumb buyers will disappear.

    Assuming that the market evolves in the direction I have sketched out, the correct answer to the question "Who will be the next Microsoft?" is "none of the above." ASP's will be confined to niche markets and forced to offer labor-intensive services, which will constrain their profitability. Developers of generic software tools will be competing in the Internet environment, where standard protocols tend to limit monopoly power (on the Net, no one knows you're a Mac).

    Most importantly, generic tool developers will suffer from the relative absence of dumb buyers. The Microsoft model will not be reconstituted.