by Arnold Kling
Date: Fri, 9 Sep 94 11:12:01 EST
From: "Arnold Kling" 
Subject: draft of paper

Attached is a draft of a paper I'm going on the economic consequences of WWW for the conference in Chicago in October. I would appreciate any comments and suggestions. If you would like a copy of the final paper to be available for your students, let me know. It will probably be up in HTML format somewhere.


This paper attempts to begin a qualitative exploration of the economic consequences of the World Wide Web. It suggests that the marketing cost function using WWW has some unique properties. The economic changes implied by this unusual cost function are sketched out. Finally, the challenges to conventional economic paradigms are discussed.


The World Wide Web has many potential uses. This paper focuses on the use of the World Wide Web as a marketing tool in commerce. In this regard, the Web lowers costs, and it lowers them in some unusual ways. Accordingly, it ll lead to some significant economic changes. In section II of this paper, the effect of the Web on the "marketing cost function" is described. In section III, the implications for various industries and occupations are sketched. In the concluding section, the increased role of information and diminished role for physical capital are described as challenging for the standard economic paradigm.

Comparing the World Wide Web to other media: the shape of the marketing cost function

One way to attempt an economic analysis of the World Wide Web is to compare it with other media in terms of the cost function for marketing a product. Traditionally, a cost function relates cost to the quantity of output and to the price of inputs. In the case of marketing or advertising, however, "output" is not straightforward to define.

As a first approximation, we could define output as the number of consumers reached with a given volume of information. For example, consider a specific comparison between the Web and newspapers. Let us use 1 million subscriber-page-days (1 mspd) as our unit of measurement. That is, we look at the cost of reaching 1 million subscribers with one page of information for one day. My reading of the Standard Rate and Data Services media guide is that a full-page advertisement in a newspaper with a circulation of 1 million costs about $60,000 for one day. Thus, in a newspaper, 1 mspd costs $60,000.

On the Web, because the Internet provider industry is not yet mature, costs vary by choice of provider. See The Internet Letter (July, 1994). However, one of the providers that appears to be near the median in terms of cost charges $900 per year for what would appear to be more than a full page in a newspaper. If we take it that there are 1 million "subscribers" (Mosaic users), then the cost of 1 mspd on the Web is $900/365, or about two dollars and fifty cents. This is less that five thousandths of one percent of the cost of 1 mspd in a newspaper.

The million subscriber-page-day concept still does not cover some important facets of communication. One facet might be called flexibility: different users need different information. For example, if the only information that a customer needs from a mortgage company in order to decide to apply for an adjustable-rate mortgage (ARM) is an interest rate quote, then the mortgage company's information system does not need to do anything other than respond to requests with interest rate quotes. On the other hand, if one customer needs a rate quote, another customer would like to know whether he or she can qualify to buy a particular home, and a third would like historical information to help compare adjustable-rate mortgages linked to different indexes, this requires a flexible information system.

A second facet is interactivity. Interactive communication may be needed in order for the seller to understand the consumer's needs. More importantly, interactive communication is needed to process a transaction.

Other facets to consider include:

timeliness. Information may need to be updated frequently to be useful. attention. In theory, a television broadcast or a Web server can be available to a large population. But will it receive attention? market niche coverage. For many products, sheer size of the audience is less importance than coverage within a given market niche. cost to the consumer. The consumer's expense in time and money to obtain information varies by the medium employed. These factors will influence the effectiveness of different media.

It will help to have in mind a list of alternative media:

broadcast media salesperson telephone with salesperson automated telephone response system print media proprietary networks and online services other Internet protocols (e-mail, gopher) WWW

Broadcast media can communicate to many people, but these media cannot satisfy diverse information needs nor receive information to process transactions. Also, the 60-second spot is not suited to convey a high volume of information.

Sales persons can convey diverse information and process transactions, but a single sales person can work with only one customer at a time. The telephone can be used to communicate either a lot of diverse information to one person (with a salesperson) or a little bit of homogeneous information to a lot of people (with an automated response system).

Print media can convey a higher volume of information than broadcast media, but the cost of reaching a large audience may be higher. Also, information in print media can become outdated easily.

Until recently, most products and services required at least two media for marketing: an advertising medium, such as broadast or print media, to disseminate information to a broad audience; and an interactive medium, such as a sales person or a telephone, to process transactions.

Computer networks are unique among media in that they combine interactivity with low marginal cost of making the information accessible to an additional user. This allows computer networks to serve both as a vehicle for broad dissemination of product information and as a vehicle to process transactions. This is the first time in history that a single medium has possessed this capability.

Within the computer network category, the older Internet protocols include electronic mail, ftp, and Gopher. An e-mail infobot, Gopher, or an ftp site can provide a high volume of information to a large audience. However, these are like telephone response systems in that complex, diverse requirements for information are difficult to meet. The proprietary information services, to the extent that their interfaces rely on menus and text, are similar to Gopher and suffer from the same limitations.

Regular e-mail (as opposed to a mailbot) is like the telephone with a salesperson, in that it can provide flexibility but without the potential for cost saving that an automated system provides. That is, a salesperson needs to respond individually to each individual request for information.

WWW is the first computer network protocol to accomodate flexibility. While many people point to graphics as the factor that differentiates the Web from other protocols, it would be a mistake to under-estimate the importance of hypertext linking in providing flexibility. It is important to recognize the freedom hyptertext brings compared even with Gopher, which gives only either/or choices: up/down; menu/document. In the mortgage banking example used above, one can easily set up a Web server so that a screen to apply for an ARM can be reached from either a screen with mortgage rates, a screen with a qualification calculator, or a screen with plots of historical data. Menus and mailbots do not lend themselves to this sort of application.

To summarize, the World Wide Web appears to have the following properties:

low marginal cost of providing information to an additional user low marginal cost of providing additional information to a given user low cost of providing timely updates of information low fixed cost of setting up a server low cost of providing flexibility low cost of interactivity

From a seller's point of view, WWW is low cost in every dimension. However, from the consumer's point of view the cost is high. Part of this problem is temporary--the fact that reliable, easy-to-install Mosaic and TCP/IP are not yet available to individual home users. A more durable issue is that the consumer, faced with an ocean of information, may confront the problem of "water everywhere but nary a drop to drink." An important issue for the Web will be what it will cost the consumer in terms of time to find information.

Implications of the unusual cost function of WWW

The marketing cost function of WWW is unusual in that it combines low cost of providing flexibility and interactivity with low marginal cost of providing additional information to additional users. Relatively speaking, more of the cost of information exchange is placed on the user, who has to search for information. These factors have implications which are fairly straightforward.

1. The cost function analysis suggests that WWW will function very differently from broadcast media. One difference is that WWW is a substitute for sales people, while broadcast media are not a substitute for sales people. Sales people and broadcast media are complementary: the broadcast media reach the large audience, and the sales people provide flexible information service and transaction processing. A combination of mass media advertising and sales persons is needed to have a complete marketing capability. However, if the cost function analysis is correct, then WWW can perform the entire marketing function by itself.

2. WWW is a complementary technology to industries and individuals that organize and filter information. Because it lower the cost of providing information, WWW will tend to create a surplus, the proverbial "information overload." This threatens to make it costly for a consumer to locate and process information. Consumers will be highly dependent on agents of various sorts, including catalogs, search tools, and "best of" lists.

3. Overall, marketing costs should fall relative to production costs. For example, imagine what would happen to Ford's marketing costs if they could use a Web server in place of TV and showrooms.

4. Economies of scale in other industries may be affected. For example, if the fixed cost associated with entering an industry consists primarily of marketing costs, and these fall dramatically, then a major source of scale economies will disappear.

5. Redesign of the typical firm. Firms are built around the existing marketing environment, which places a premium on delivering information to internal users--sales people and executives. In the future, the firm may be redesigned to deliver information to external users--the customers.

6. More of the value of a product will be created outside of the firm. Consumers may create value by designing more of the product. Agents that filter information will create some of the value. This value "synthesis" will differ from the concept of value added, and uncertainty will be introduced into the relationship between value added and value capture.

These implications can be taken down to a more specific level. Industries and occupations will be affected.

One industry that would seem threatened is broadcast media. If products and services are marketed more efficiently on the Web, then TV and radio will enjoy less revenue. This will occur even if the inherent ability of these media to entertain consumers stays the same or improves, because for advertisers the relative cost of providing information on these media will appear to be higher.

Some service industries may be threatened by the low cost of providing information. Consumers may start to make "end runs" around professional lawyers, doctors, real estate agents, stock brokers, insurance agents, etc., because the demand for these professions is a function of consumer ignorance. In the future, consumers may make more of their own decisions in these fields, based on information that is more readily available. Eventually, one must consider the possibility that students will be able to make "end runs" around colleges and universities in order to obtain effective and credible education.

At an occupational level, clearly this line of thinking would indicate that there will be fewer sales clerks in the future, as automated order-taking expands on the Web. More interestingly, corporate MIS may be a threatened occupation. The thinking here is that corporate MIS exists to maintain proprietary systems that meet the needs of internal customers, while the focus in the future will be on reaching external customers using standard Internet protocols. Therefore, corporations with big MIS departments may find themselves beaten out by firms with a lower MIS overhead.

Corporate executive also may be a threatened occupation. Executives are skilled at managing the internal flow of information. Like the corporate MIS department, this function may be less important in the future.

The foregoing would suggest that I am predicting that all sorts of jobs will be "lost." I want to hasten to point out that conventional economics would say that an innovation that cuts costs and raises productivity will tend to raise economic well-being. It will create new industries, and it will create new opportunities. The gains will more than offset the losses.

New industries include more than just Internet service providers. They include products and services that would not have been economical with the marketing cost structure that existed prior to the Web. Narrow-interest movies and music, for example.

Industries that help filter information for consumers will be important. Something like an on-line magazine that takes a particular topic, organizes information, and offers product reviews could be a major industry. A service that surveys a consumer's tastes and past purchases and builds a statistical model to predict what products and information would interest that consumer may be very important. This will differ from traditional market research or direct-mail solicitation in that consumers are likely to play an active role in seeking out, evaluating, and choosing such services.

An occupation with new potential could be that of librarian. Every business may need to hire librarians to help design Web displays. A super-librarian tomorrow could be the equivalent of a super-salesperson today.

Instead of a business executive, perhaps there will be consumer executives. They will be skilled at providing consumers with information and at providing consumers with tools to convey their interests and preferences to firms.

Is economics a threatened occupation?

This paper has argued that the WWW represents a change in the shape of the marketing cost function. This in turn will cause economic shifts, some of which were sketched out in the preceding section.

One interesting note is that the concepts and predictions in this paper are not as precise as those that I recall reading in papers when I was in graduate school. Perhaps this means nothing more than that I am guilty of practicing bad economics. However, I would prefer to believe that the problem lies in the fact that conventional economics is one of the professions that is challenged by these new phenomena. Here are what I think are some of the challenges:

1. Capital is not a central actor in the drama. Capital really plays a pivotal role in economics as I learned it. We used production functions that employed capital and labor, we showed how income was distributed between capital and labor, we dealt extensively with the problem of capital allocation, etc. On the Internet, physical capital is not a major part of the story.

The way I like to put this is that in building the interstate highway system the challenge was constructing the physical roads and bridges, not in coming up with the rules of the road. With the Information Superhighway, one could argue that the situation is reversed, and indeed that this reversal is one of the major problems with the highway metaphor itself.

The traditional focus on capital would lead economists to focus on efficient use of hardware. However, one could argue that of all the issues on the Internet, pricing the use of routers and phone lines is hardly the most interesting, particularly when compared to the issues involved in pricing the content of information services.

2. Many of the issues involve information and marketing costs. In economics, we used to accomplish a great deal by assuming no marketing cost and perfect information. However, when we discuss a communications medium, we are digging underneath this assumption and trying to understand something different. It's like going below the level of atomic chemistry to study sub-atomic particle physics. Suddenly we face a lot of difficult, complex issues that previously we had conveniently ignored.

One aspect of this challenge is that terms do not seem all that well-defined, much less measurable. Terms like interactivity, flexibility, or attention seem less precise than terms like "labor," "capital," and "utility," although perhaps this reflects little more than the fact that economists have grown accustomed to working around the ambiguity of the latter concepts.

If economists can study the economics of computer networks with the same precision of language and theory that we use in our models of capital and labor, that would be ideal. However if economists continue to insist on mathematical precision, but precise models cannot yet be applied to the interesting issues raised by the World Wide Web, economics may be a profession that gets left behind by the computer/communications revolution.

Further Reading

In general, the best place to read about the Internet is on the Internet. Below is one published article and several World Wide Web sites that contain more information on the economics of WWW.

1. MacKie-Mason, Jeffrey K. and Hal Varian, "Economic FAQs About the Internet," The Journal of Economic Perspectives, Summer 1994, 75-96.

2. "University of Michigan Economics gopher,", maintained by Hal Varian.

3. "Computer-Mediated Marketing Environments,", maintained by Donna Hoffman and Thomas Novak.

4. "Thomas Ho's favorite Electronic Commerce WWW resources,", maintained by Thomas Ho.

5. (Brad Cox's) "World Wide Web (WWW) Home Page maintained by Brad Cox.

Arnold Kling is President, ASK Real Estate Information Services, Silver Spring, Md. His company developed the Homebuyer's Fair (, an on-line service to connect consumers with mortgage lenders and home sellers. He spent seven years with the Federal Home Loan Mortgage Corporation (Freddie Mac), where his accomplishments included design of the automated underwriting system. Prior to Freddie Mac, he spent six years as an economist with the Federal Reserve Board. He has a Ph.D in economics from the Massachusetts Institute of Technology. His publications include "Imperfect Information and Price Rigidity," (Economic Inquiry, January 1982), "Futures Markets and Transaction Costs," (in Myron Kwast, ed., Financial Futures and Options in the U.S. Economy), and "The Rise in Bank Failures from a Macroeconomic Perspective," (Journal of Financial Services Research, December 1988). He may be reached at
Arnold Kling
ASK Real Estate Information Services
1370 Lamberton Dr.
Silver Spring, Md. 20902
(301) 754-0730 (w)
(301) 649-4378 (h)
Virtual School Middle of Nowhere Brad Cox

Modification date: March 07, 2004 © Copyright 2004 by Brad Cox
Served by John Companies