Abstract Are there  economic incentives for electronic commerce (e-commerce),  or is  it just  hype? This paper evaluates the cost-based differences between traditional markets (such as retail stores) and electronic markets (e-markets) both from the buyer's (demand side) perspective and the  seller's (supply  side) perspective. The implications that a shift toward  greater electronic market utilization have for  transaction intermediaries,  interactive  service providers (ISPs),  and government are discussed. We find that  there  are significant cost-based differences between traditional  and electronic markets for both buyers and sellers, and that electronic markets affect industry structures as well as future sources of organizational revenue.
 Kevwords." Electronic markets; Electronic commerce; Transaction cost economics; Industry structure: Consumer behavior; Business strategy
Introduction
  In the past few years, there has been an explosion of online commercial activity enabled by the Internet and  World Wide Web (WWW).  Activities range from  simple product advertising to  more complex systems  that  facilitate  electronic product ordering either directly from one company, or through  elec- tronic  markets. Commercial transactions  have taken place for centuries, but currently there is a revolution taking  place that  is  transforming  the  marketplace. This transformation  is  occurring because the  rela- tionship between organizations and consumers is in- creasingly being facilitated through  electronic infor- mation technology (IT). This is generally called elec- tronic  commerce (e-commerce), with a major com- ponent of e-commerce being electronic markets (e- markets). The number of products available online is growing steadily.  There are an  estimated  250,000 commercial Web sites now operating, and the  vast majority have been around for less  than  a year [1]. But, not enough  is  understood  about this  rapidly evolving phenomenon.  The number of losers  still exceeds the number of winners by 2 to  1 for Internet commercial ventures [l].
   The questions  that  arise from the current growth of electronic commerce are whether all commerce in the future will be facilitated electronically,  or  whether electronic  commerce's hype far exceeds its future impact. It is likely that the future  of electronic commerce falls  somewhere  in between the  two  extremes. Some industries  will be affected more dramatically by electronic commerce than others. The purpose of this paper is to evaluate the differences between traditional  markets (such as retail  stores), and electronic markets both from the buyer's (demand  side) perspective and the  seller's (supply side) perspective. We also discuss the impli- cations that  e-markets have for transaction  facilita- tors  (intermediaries, brokers and so forth), the rela- tively new industry  of interactive  service providers (ISPs), as well as state and federal government. Past work has focused  on the  theoretical  relationship, generally based on transaction cost economics analy- sis  [2],  between  IT  and  transaction  governance (markets vs. hierarchies) [3-8]. Our study involves a cost-based economics  analysis similar  to  previous work, but we compare traditional  markets with e- markets  instead  of  markets  with  hierarchies. Williamson states that  the  economic institutions  of capitalism (namely,  markets and hierarchies) have the  main purpose  and  effect  of economizing  on transaction  costs [2].  Our thesis  is  that,  in  many instances,  electronic markets enjoy transaction  cost advantages over traditional markets. Because of these transaction cost advantages, we can expect a contin- ued growth in online markets in many industries.
  The following  sections  describe the  findings  of our study. In Section 2, we discuss e-commerce and e-markets to  provide background for the  remaining sections.  Several e-market examples are presented. The remaining  sections  describe the  impact  of e- markets from three perspectives: buyers, sellers, and other  organizations  associated  with  commercial transactions.  In Section 3, we identify  the  impacts that e-markets have on industry  structures. We dis- cuss  traditional  retail  industry  structure,  industry structure for non-digital product e-markets, and in- dustry structure for e-markets associated with digi- tized products. In Section 4, we evaluate the charac- teristics of traditional and electronic markets from a buyer's perspective. We derive several revenue im- plications for  sellers  and other organizations from this analysis, as well as the analysis from Section 3. In Section 5, we evaluate the cost-based differences between traditional  and electronic markets from a seller's perspective. Next, in  Section 6, we discuss the  impact  that  e-markets have on other organiza- tions  including  transaction intermediaries,  ISPs, and state and federal government. Finally, in Section 7, we discuss our overall conclusions.
Electronic commerce and electronic markets
  There are  several  reasons  why electronic com- merce's time has come. One reason is that widespread electronic commerce has only become feasible within the past few years. It requires that both applications, as well as infrastructure be widely available to con- sumers. Examples of critical technologies  that  sup- port e-commerce include  hardware such as low-cost personal computers (with fast processors and large storage capabilities), high-speed modems, the  Inter- net, World Wide Web and Netscape Navigator. The Internet's World Wide Web is considered a strategic information technology with the  potential to change the  ground rules by which businesses interact  with their consumers [9]. The high expectations surround- ing the consumer potential of the Web are driven by its  perceived business advantages, by socio-demo- graphic changes, and by the  Web's unique features as a  direct marketing channel  [10].  The perceived business advantages include  greater inventory  con- trol, improved market reach, customization of prod- ucts and services, shortened time to market for new products and services, more efficient payment sys- tems,  and lower advertising costs [11-13].  The so- cio-demographic drivers include an increase in fami- lies with two working parents, single-parent homes, pressure to  free  up  limited  leisure  time,  a  more computer-literate population, crime, and cutbacks in store personnel [ 14]. Finally, Kalakota and Whinston have identified four principal reasons why electronic commerce's time has come. First, the  cost of pro- cessing many types  of financial and retail  transac- tions has been rising so rapidly that  it is imperative to  develop new ways to  handle those  transactions. Second, competition in  banking and retailing  has become so intense that only those organizations that can  provide superior  customer services,  which in turn  require  sophisticated transaction  management, will continue to grow and prosper. Third, consumers themselves  are feeding  the  fires  of competition by demanding more services and greater convenience in their  banking and shopping  activities.  Finally, the technology is  at last  in  place to  process electronic transactions at faster speeds more easily and at less cost than we can process paper transactions [15]. The shift  toward  electronic  commerce is  revolutionary because it involves linking  consumers to  electronic marketplaces, not just electronically supporting hier- archical transactions  within and between organiza- tions (commonly referred to as the problem of enter- prise integration). The involvement of consumers, in addition to  product/service providers, dramatically increases  the  potential magnitude of change. A sig- nificant portion of the GDP is consumer transactions. Past growth in enterprise integration systems missed these transactions.  The revolutionary nature of elec- tronic commerce provides adequate incentive to study electronic markets to  increase  our understanding of their impact on the  market's participants, traditional and newly created industries, as well as the economy as a whole.
2.1. Electronic commerce
  Broadly defined, e-commerce is  a modern busi- ness methodology that addresses the needs of organi- zations, merchants, and consumers to cut costs while improving  the  quality of goods and services,  and increasing  the  speed of service delivery. The term also  applies to  the  use of computer networks to search and retrieve information in support of human and corporate decision-making. More commonly, e- commerce is  associated with the  buying and selling of information, products, and services via computer networks today and in the future via any one of the myriad of networks that  make up the  Information Superhighway (I-way) [15].
  Electronic commerce involves  several layers  of infrastructure  including  the  information  superhigh- way infrastructure,  multimedia content and network publishing infrastructure, the messaging and informa- tion distribution infrastructure, and common business services infrastructure  [15].  It also encompasses a wide range of applications, designed based on, and supported  by, the  infrastructure.  Example applica- tions  include  electronic advertising, product order- ing,  delivery of digitizable products, payment sys- tems, and the focus of this paper, electronic markets.
2.2. Electronic markets
  Electronic markets are  the  foundation  of elec- tronic  commerce. They potentially encompass all of the applications discussed above. An electronic mar- ketplace (or  electronic market system) is  an inter- organizational  information  system that  allows  the participating buyers and sellers to exchange infonna- tion  about prices and product offerings. The firm operating the  system is referred to as the  intermedi- ary, which may be a market participant--a buyer or seller,  an  independent  third-party,  or a  multifirm consortium [6].  E-markets provide an electronic, or online,  method to  facilitate  transactions  between buyers and sellers that  potentially provides support for  all  of the  steps in  the  entire order fulfilment process. The business process model from a  con- sumer's perspective consists of activities that can be grouped into  three  phases: prepurchase determina- tion,  purchase consummation, and postpurchase in- teraction.  Each of these  phases can  be supported electronically in a complete e-market, but e-markets today generally support only the  prepurchase deter- mination activities.  Several e-market examples are described in Section 2.3.
2.3. E-market examples
  A number of electronic markets are available to consumers to buy products ranging from music CDs to automobiles. The following are current examples of products and/or services that are available through electronic markets.
2.3.1. Flowers
  Calyx & Corolla have used e-commerce to radi- cally alter the  way new cutflowers are moved from the  growers to  the  consumers.  Traditionally,  the value  chain  that  supplied  cutflowers  involved  a grower, jobber  to  transport  to  a  wholesaler, and finally a florist. From a survey of Boston florists in July  1995, the  price, including  delivery charge and tax,  for  an  example arrangement  of flowers  was US$60. Calyx & Corolla can provide an electronic market to  customers to  buy directly from  growers with the  flowers  being shipped  using Federal Ex- press. Their delivered price is US$54 [16]. Much of this is due to the elimination of some intermediaries between the  growers and the  customers. The price paid to  the  firm providing the  electronic market is generally lower than the  profits made by the  tradi- tional  wholesaler and retailer intermediaries.
2.3.2.  Clothing
  Similar to the cutflower example is an example in the shirt industry. The cost per high-quality shirt in a value chain that includes  a wholesaler and retailer is US$52.72. The elimination of these  intermediaries reduces  the  cost to  US$20.45, a reduction  of 62% [8].
2.3.3. Automobiles
  Thanks to  the World Wide Web, new car shop- pers have more options, including access to valuable information,  such as what a car really  does cost a dealer. As a result, consumers are increasingly lock- ing in better deals online. What's more, the trend has attracted the  attention  of some of the  biggest car dealers, financial  institutions  and insurance  compa- nies.  Electronic markets now exist to  enable  con- sumers to shop for and buy a new car, insure it, and take delivery without ever setting foot in a dealership [17]. A search of the directory of automobile dealers on Yahoo in October showed that 79 different deal- ers or locator services were listed [18].
2.3.4. Music
  Jason and Matthew Olim founded CDnow from the  basement of their  Ambler, Pennsylvania home. Jason Olim, a jazz fan frustrated by skimpy selec- tions  in  music shops,  came up with the  idea  of a cyberstore that could offer every jazz album made in the  US and 20,000 imports.  Shoppers place their orders with CDnow (cdnow.com),  which, in  turn, contacts distributors. Most disks are delivered to the customer's door in  24 h. Add in  advertising rev- enues, and CDnow expects to hit US$6 million in sales  in  1996,  triple  the  previous year's revenue, with 18% operating margins [1].
2.3.5. Books
  Books are another product that  consumers pur- chase online. One book seller on the Web is  Ama- zon.com Books. Their site  advertises  a  spotlight book, book of the day, titles  in  the  news, featured books, and books that  are hot this  week. Some of their  books are discounted as  much as 30%. By clicking  on book titles,  and  some  authors,  more detailed information can be accessed [19].  It is  no longer necessary to either go to a book store to buy a book, or to  find mail-order book stores  through  a print advertisement. Web advertising is  likely  to be more current than print ads.
2.3.6. Electronic magazines
   With no printing or circulation costs, online mag- azines once held the promise of low overhead and quick profitability. Now most Web publishers have amended their business models and expect years of losses before turning a profit--a model much closer to print publications. Although analysts and publish- ers expect mainstream advertisers to up their antes in Web ads, most e-zines are exploring alternative ways of making money in the  short-term, including spon- sorships, alliances and even subscriptions. Most on- line  publishers have a rosy  outlook, now that  the Internet has become a media focal point, and main- stream advertisers better understand the Net. Jupiter Communications, a  New York-based Internet  re- search company, predicts that  the  total  number of online consumers will jump from 13 million in  1996 to more than 35 million in 2000. Adam Schoenfeld, vice-president and senior analyst at Jupiter, said that the universe of ad dollars online--both on the Web and on dedicated online services--would grow to US$5.3 billion by 2000 [20].  A growing number of online consumers, as well as a growing amount of Net-based ad money, provides an environment where electronic magazines with good content may flourish in the future.
2.3. 7. Airline tickets
  Discount airfares you won't find anywhere else are popping up on the  Internet.  American Airlines and Cathay Pacific Airways are using their Web sites to reduce the thousands of seats that  are unsold on flights every day. American began selling fares on 20 routes  as much as 70% below the  lowest  fares consumers would be quoted through a travel agent or American's 800 number. Besides filling empty seats, airlines  want to  cut  distribution  costs by selling directly on the  Internet  instead  of through  travel agents [2l].
2.3.8. Stock and securities
  All of a sudden, innovations in technology, partic- ularly the Internet, are bringing profound changes to Wall Street that  hold a lot of promise, and a lot of peril, for the powerful firms that make their money in  the  securities  business.  For many people,  the Internet could replace the functions of a broker. For example, almost a dozen small companies are trying to  sell  their stock directly to  the public using Web sites like those run by Direct Stock Market and IPO Data Systems. And two small California companies, Real Goods Trading and Perfect Data, have set up electronic bulletin boards that allow their sharehold- ers to trade stock without a broker, dealer or market maker. Because it  allows traders  to find each other easily, the Internet may ultimately make it  possible to  have a  stock exchange that  exists  only in  cy- berspace,  with no trading  floor,  directly open to every investor with a computer and a modem [22].
  Three sets  of issues  and research questions arise from an analysis of these examples. First, what is the impact  that  electronic  markets have on the  costs relevant to a consumer's choice between traditional retail  markets and electronic markets? Second, what is  the impact that electronic markets have on seller costs, as well as the  structure of the  value chains needed to provide products? And third, what impact do electronic markets have on other organizations involved in  commercial market transactions? These three  issues are addressed throughout the remainder of this paper
Impact of e-markets on industry structure
  It is  apparent from the  examples above that  the diffusion of electronic commerce in an industry has an impact  on the  structure  of the  value chain in- volved in  supplying the products and/or services to the final consumers. This is mainly due to the disin- termediating effect of information technology identi- fied by Davenport in  his research on business pro- cess reengineering [23]. Although, in some instances, intermediaries  may be added to  transactions  facili- tated  through  an electronic market. Based on the examples above, we have identified two phases that industry  structures  potentially go through  as elec- tronic  markets diffuse across the  industry.  The de- gree of change is  determined by features  of the industry and its  products. This is discussed in  more detail at the end of this  section.
  An example of a traditional  market is  shown in Fig. 1. The industry  transformation  phases are de- scribed in relation to this  example. In  a traditional  market (for a non-impulse pur- chase), the customer searches out information about the  products available and their  prices, quality and features. This information comes from a wide range of sources including advertising sources, travelling to retail stores, and so forth.  At some point they stop their search because they realize that further search- ing  will probably not benefit them.  Once the  intbr- mation gathered has been analyzed, the  consumer decides where to  buy the  product. The product is then  either purchased and transported  home by the customer, or is delivered to them through a distribu- tion network.
  Electronic markets affect the  consumer purchase process. The first phase in the  translbrmation of the structure  of an industry  is  the  digitization of the market mechanism. This is described in Fig, 2.
  An electronic market provides a mechanism for reducing the  search costs (money, time  and effort expended to gather product price, quality and feature information)  for consumers. This also  reduces  the likelihood that sellers can charge significantly higher prices than their competitors because the consumer is unaware of the  other prices  (a  form  of regional oligopoly or monopoly). The result is that consumers can  buy products for  lower  prices,  intermediaries such as wholesalers are eliminated from  the  value chain, a new industry  that  provides access to  elec- tronic  markets is  created,  and firms  that  produce products can maintain a profit margin comparable to the traditional markets.
  The second phase in  the  transformation of the structure of an industry  is  the  digitization of the product itself as well as its distribution. Examples of digitizable  products  include  books,  newspapers, magazines, computer software,  movies and music. For example, if a consumer wants a new version ofNavigator software from Netscape, the  software can be downloaded from one of their sites on the Internet [24]. This eliminates the  need for Netscape to main- tain  an inventory of software  on CDs or diskettes that  must be physically shipped to  the  consumer. Another example would be either evaluating or pur- chasing anti-virus software from McAfee [25].  If a software  company charges for  their  software  then they  can  receive  payment before the  software  is allowed to  be downloaded. This can be especially easy, as electronic payment methods become more widely used in the future. This further transformed industry structure, that results from the  digitization of products and their distribution, is described in Fig. 3.
  The electronic market and distribution network enables a wide range of seller and customer activities to converge into one place including marketing, or- der  processing,  distribution,  payments  and  even product development processes that  involve several separate firms. This makes these activities easier and more convenient while also reducing  the  costs in- volved. Value chain costs can be further reduced by digitizing the industry's product. Examples of digiti- zable products were given earlier. Digitization of the product reduces inventory and packaging costs. Digi- tized products can then be distributed electronically to  the  consumer, minimizing distribution costs that would otherwise be paid to the firms in the distribu- tion  network and passed on to  the  final consumer. These cost-based differences are discussed in  more detail in  Section 5. Beyond the  cost benefits, cycle time  for order fulfilment is  minimized, which may result  in  improved customer satisfaction.  Digitized information can be distributed in minutes while ship- ping a product generally takes  days (or  longer  to some parts of the world). The characteristics of the phases in  the  transformation  of industry  structures enabled by e-markets are summarized in Table 1.
  The overall impact  of electronic  commerce on industry  structures is  not strictly cost reduction  and disintermediation. It is more complex than that. New intermediaries  and costs may be added to  a value chain, but in  many instances  the  potential benefits outweigh these costs. In the Section 4, we discuss a model that identifies costs relevant to differentiating between traditional  and electronic markets from  a buyer perspective.
Traditional and electronic markets: buyer cost perspective
  Electronic markets provide buyers with an addi- tional  sales  channel  through  which they  can  buy products.  Although there  may be certain  benefits derived by buyers in electronic markets (lower prices and search costs), it also increases the complexity of their  decision process by adding another option to consider. It may also add new forms of consumer risk. In this section we describe a model we devel- oped to compare the cost-based differences between traditional  markets (such as retail  stores) and elec- tronic markets.
4.1. Buyer's perspective relevant costs
  From the consumer's perspective (demand side of a transaction),  the  potentially relevant costs that we have identified  include:  (1) product price (PB),  (2) search costs (SCs), (3) risk costs (RCB), (4) distri- bution  costs  (DCs),  (5)  sales  tax  (TB),  and  (6) market costs (MC B).
  The product price is  the  sum of the  production costs,  coordination costs,  and profits of the  value chain that  provides the  product or service.  Search costs include the time, effort and money involved in searching for a seller who has the product demanded at an acceptable price with acceptable product fea- tures  and quality. The cost of the  time  and effort involved would be determined by the value the buyer places on their time and effort. Risk costs include the costs involved in minimizing transaction risk, as well as the costs associated with losing value in a transac- tion.  The risk dimensions typically  considered are: economic risk,  performance risk,  and personal risk [26].  Economic risk  stems from the  possibility of monetary loss  associated  with buying  a  product. Performance risk represents the  consumers' percep- tion  that  a  product  or service  may fail  to  meet expectations. Personal risk relates to the  possibility of harm to  the  consumer resulting  from  either  a product or the shopping process. An additional form of risk that is potentially important to Internet shop- pers is  privacy risk. Privacy risk reflects the degree to which consumers envisage a loss of privacy due to information collected about them as they shop [10]. Additional costs of concern include distribution costs, the  costs  associated  with physically moving the product from the  seller to the  buyer, and sales tax. Market costs are the costs associated with participat- ing in  a market. Traditional markets are assumed to be costless to  the  buyer, while e-market costs may include fixed access costs and/or transaction (varia- ble) costs paid to the firm(s) that operate the e-market,
4.2.  Comparison of buyer costs in  traditional  and electronic markets
  Assuming rational  decision-making, the  buyer's objective is  to  minimize the  sum of the  individual costs subject to the constraints that the product qual- ity and features, including how soon the product can be received, must be acceptable. Fig. 4 summarizes the findings of our evaluation of the costs relevant to a buyer's choice between traditional  and electronic markets.
  Prices in  electronic markets are generally lower than  in  traditional  markets. If they were higher then there  would be little  incentive  for  consumers  to switch to the newer e-markets. One explanation for why prices are lower is that search costs are lower. It is generally easier to gather relevant information, and compare a wider range of prices, in online environ- ments. This is  especially true  as  the  number of products offered online increases. In traditional retail markets, a buyer would have to either drive around town or call several sellers. This takes more time and costs more. Given this additional information in the e-market, buyers are likely to be able to find a price that is lower than in a traditional market. The ques- tion  then  is:  Why aren't all products purchased in e-markets? One reason is  that  not all products are available through e-markets. Another reason is that some products cannot be feasibly sold online. More important,  though,  is that there  are additional risks associated with buying online. Given that there  are transaction cost-based incentives for buyers to partic- ipate in e-markets, the next issue is whether there are incentives  for  sellers  to  participate.  Without any sellers in  the  e-market, there  would be no transac- tions.
Traditional and electronic markets: seller cost perspective
  Electronic markets provide sellers with an addi- tional sales channel where they can market and sell their  products. As with buyers, electronic markets provide sellers with certain benefits including access to a larger market and reduction of certain costs, but it also increases  the complexity of their  operations by adding a new potential sales channel to evaluate which changes the way they may do business in the future.  In  this  section,  we describe  a  model we developed to compare the cost-based differences be- tween traditional markets (such as retail stores) and electronic markets from a seller perspective.
5.1. Seller perspective relevant costs
  From the  seller  perspective (supply  side  of a transaction),  the  potentially relevant  costs that  we have identified  include:  (1) marketing (advertising) costs (ACs), (2) overheard costs (OCs), (3) inventory costs (ICs), (4) production costs (PCs), and (5) distri- bution costs (DCs).
  Marketing costs  are  the  costs  associated  with informing  the  consumer about the  availability and features of a seller's products or services. Advertis- ing channels in traditional markets include television, radio, newspapers, yellow pages, and so forth. Newer advertising  channels  include  push-based  methods (such  as  electronic  mail), and pull-based methods (such as electronic bulletin boards and the Web) [15]. Overhead costs include  the  more fixed costs of the business including  physical retail  space and ware- houses. Inventory costs include  the costs to  handle and hold inventory to deal with demand uncertainty for physical products. Production costs include  the variable costs  of producing  a  unit  of a  product including  labor and materials. Distribution costs in- clude the costs associated with moving the  product from the seller to the buyer.
5.2.  Comparison of seller costs in  traditional  and electronic markets
  Assuming rational  decision-making, the  seller's objective is  to  minimize the  sum of the  individual costs subject to the constraints that they provide the products and services demanded by their customers. Fig. 5 summarizes the  findings of our evaluation of the  costs relevant to a seller's choice between tradi- tional markets, e-markets with nondigitized products, and e-markets with products that have been digitized.
  Advertising costs are lower in e-markets than  in traditional markets. For example, the advertising cost per consumer for a Web page is  much lower than a television ad or a print ad (magazine or newspaper). This is  true  whether the  product is  digitized or not. Overhead costs are similar to advertising costs. Tra- ditional retail store markets require a seller to have a physical store, which they may either own or rent. In e-markets, a Web site may also serve as the  store front.  This is  especially true  when the  capability to order products electronically is  integrated  into  the Web site. Inventory costs are more closely related to the  product characteristics  instead  of the  consumer interface.  When products are digitized, they require an inventory level of only one unit, and the product is  stored on a computer disk. Production cost differ- ences are similar to the  situation for inventory costs. Physical products involve  significant  variable costs per  unit for  materials and labor.  Reproduction of digitized products generally involves the  copying of the computer file. Distribution cost differences are a little more complex. In an e-market with a digitized product, the product can be distributed electronically, perhaps through FTP, to the consumer. This is a very low-cost  distribution  method. Traditional  markets also have low  distribution costs for  sellers because the  consumer comes to  the  store and transports  the product  to  their  home  themselves.  An electronic market with a  nondigitized  product  still  requires physical  shipment  of  the  product,  for  example through  the  USPS or Federal Express. This is  the situation  with the  highest  distribution  cost  to  the seller.
  Several implications  for  business  strategy  (and potential sources of revenue) are apparent from our findings related to our e-market model and empirical study.  These  implications  affect  several  entities: product/service providers, transaction brokers, inter- active service  providers (market  makers), and also state  and federal government. The buyer costs rele- vant to each entity (as potential sources of revenue) are marked in Fig. 6.
5.3. Product/service provider revenue implications
  The  revenue  implications  for  product/service providers in  an e-market come from the  price and risk cost components, PB and RC B, of our buyer cost model. Essentially, sellers can compete using a low- cost  producer strategy,  and/or  they  can  compete using a strategy  by which they  differentiate them- selves from other sellers because they are less risky (more  trusted)  in  the  market. Competing based on reducing  buyer risk  costs,  when the  seller/buyer relationship  is  supported  electronically, can be de- scribed as an  electronic virtual partnership. This is described in Fig. 7.
  This is  interesting  because it  describes how, over time,  submarkets  may form within the  overall elec- tronic  market because consumer knowledge is  lim- ited,  and there  is  still  a cost to  gather information about new sellers.  Over time,  there  is  less  incentive for buyers to search the entire e-market for sellers of a product that has been purchased in the past. Unless a seller's price is significantly lower than prices from a trusted  seller,  the  switching cost will inhibit  the consumer from buying from the unknown seller.
 Revenue source  implications  for  transaction intermediaries
  Electronic markets also  affect potential revenue sources for other organizations that support commer- cial  transactions.  We discuss  the  implications  for transaction brokers, ISPs, and the government.
6.1. Transaction intermediary implications
  The implications  for  transaction  brokers  (e.g., stock  brokers, real  estate  agents,  intelligent  agent software  developers,  and so forth)  in  an e-market come from the search and risk cost components, SC B and RC B, of our model. In some situations, buyers may be willing to  pay someone  to  gather and/or analyze market information (a traditional broker role) related  to  their  purchases,  or they  may pay  for software that provides this  same functionality (deci- sion support  software or more advanced intelligent agent software). They will pay for someone or some- thing  (an  intelligent  agent  system?)  to  do  their searching. Also, consumers may be willing to  pay for  broker  services  such  as  risk  assumption.  For example,  consumers  may be willing to  pay for  a service  such  as  an  online  better  business  bureau where they  could check to  see if  there  have been complaints  against a certain  seller.  There are  also implications for distribution companies (such as Fed- eral  Express) that  arise  from  the  distribution  cost component, DC B, of our model. Package shipment companies  can  expect  continued  growth  in  their business related to increased usage of e-markets, but, as more and more products are digitized, this  growth may be reduced.
6.2. Interactive service provider (market maker) im- plications
  The revenue  implications  for  interactive  service providers come from  the  market and  search  cost components,  MC B and  SC B, components  of our model. Consumers may be willing to pay interactive service providers a portion of the  money they  save by buying products in  an e-market versus a tradi- tional  market to  gain access to  the  e-market.  Con- sumers  may also be willing to  pay for  access  to systems because they provide much more than just e-market  access,  for  example,  entertainment.  The growth of ISPs  clearly shows  that  consumers  are willing to pay for these services. The fixed cost that consumers pay to ISPs varies, but it  is  common to pay about US$20 per month.
6.3. Government
  Finally, the implications for state and federal gov- ernment come from the  sales tax component, T B, of our model as well as organizational revenue  gener- ated through e-market transactions. As more transac- tions move from traditional  markets to e-markets, it is likely that a smaller proportion of sales tax will be collected by state governments. Generally, laws exist that  require  the  payment of sales  taxes  even  on interstate  commerce,  but  collection  is  a  practical problem. This is especially likely since entry barriers into e-markets are low, increasing the likelihood that there  will be  an  increase  in  the  number of sole proprietorships and small businesses that  sell prod- ucts online to buyers around the world. It is harder to track a large number of small sellers. It is  also more difficult to  track  e-market transactions  that  involve buyers and sellers  in  different countries. For state and federal govemment, there  is also the problem of collecting tax  from  all of these  sellers  for  taxable income  generated  from  e-market  transactions.  As these  problems increase  with the growth of e-com- merce, we can expect a greater effort of government to collect the sales tax and income tax they are owed.
Discussion and conclusions
   Based upon our inductive analysis of several cur- rent  examples of electronic markets, and the  buyer and seller cost-based differences between traditional and electronic markets we have identified,  we make several observations and conclusions. First, we dis- cuss some factors  that  may inhibit  the  growth of electronic markets in the future. Second, we identify some  factors  that  affect the  level  of impact  that e-markets may have on industries.  And finally,  we present the overall conclusions of our study.
7.1. lnhibitors  to electronic market success
  Throughout our paper we have assumed that the impact of certain factors that  inhibit  the  future suc- cess of all e-markets,  and e-commerce in  general, will not sufficiently hinder the  growth of electronic commerce in  the  future.  If this  assumption is  not true,  then  the  study  of electronic markets is  moot since they may not exist in the future. It is important to acknowledge the existence of barriers to electronic market success. Three examples of inhibitors to elec- tronic market success are discussed below.
  First, the  lack of IT infrastructure  in some worm regions  is  a barrier to e-commerce participation by companies and consumers in these regions. In many countries, consumers do not have the  same level  of access to the Internet, World Wide Web, and so forth that  consumers  in  the  US have.  This is  a  major barrier to electronic market diffusion because even if consumers wish to participate in e-markets, they are physically unable to.  Even if access is  available, an additional barrier may be poor physical telecommu- nications. However, the increasing recognition of the importance  of telecommunications  to  national  and business infrastructure  has resulted  in  its  prolifera- tion  to  newly opened societies  and markets, most notably Eastern Europe and the former Soviet Union, and to rapidly expanding markets such as Egypt and Iran [27].  We should expect a continuation in  this trend toward greater access.
  Second, the  lel,el  of computer illitera O, associ- ated with the  worM's consumers that have access to IT  infrastructure  is  a  barrier to  e-market  success. Because of a lack of education about computers, or a lack  of willingness to  accept  new  technology,  a certain proportion of consumers are unable or unwill- ing to participate in electronic markets. As more and more children are introduced to computers in school, the  proportion of consumers  who potentially may participate in  electronics will increase  in the  future, Electronic markets are likely  to  be considered nor- mal instead  of novel for future generations of con- sumers.
  And third,  insufficient data and message security may inhibit  some companies  and consumers  from participating  in  e-commerce because they  feel  the level of risk is unacceptable. Confidence, reliability, and protection of information against security threats is  a crucial prerequisite for the  functioning of elec- tronic commerce [15]. Many initiatives are underway to  improve  security through  improved  data encryp- tion  and digital  signatures.  A specific  example is S-HTTP, a more secure version of HTTP that is used in the  World Wide Web. As the  level of transaction security for e-commerce related information transfer improves,  the  expected level  of e-market impact on industries,  and the  global economy in  general, will increase.
  An additional inhibitor to e-market success is  the fact that  a significant portion of all transactions  are not market transactions, but are hierarchical transac- tions.  Hierarchical transaction  governance is  often associated  with transactions  involving  high  asset specificity  [2].  Asset specificity  is  the  difference between the  value of an asset (machine,  employee and so forth) in its present use and its next best use. Transactions involving  high  asset  specificity  will continue to be governed by hierarchies because the firms  involved  generally need  to  maintain greater control over the  transactions  (perhaps through verti- cal integration  or long-term  contracts) to  minimize their overall risk.
  These inhibitors,  as well as other factors such as high market access costs, have resulted in the failure of some electronic markets. One example is  an elec- tronic  market for real estate. The National Associa- tion of Realtors' widely publicized information  net- work, created  two  years ago to  provide extensive real-estate information  on the World Wide Web, has run  out of its  US$12.9 million in  funding and is  on the  verge of collapse. Association officials and peo- ple in  the  industry  say  the  network fell  victim to overly  ambitious  goals,  some  free-spending  ways and unexpected changes in technology that  made it less  attractive to  its  primary customers,  real-estate agents.  Funded from the  national  association's re- serves,  Realtors Information Network, or RIN, had lofty  plans for keeping Realtors in  control of real- estate  transactions.  The for-profit subsidiary  would provide real-estate listings nationwide on the Web to consumers and would act as something of a propri- etary America Online for real-estate  agents. Agents who purchased the  system  would have  access  to information,  chat rooms,  real-estate  vendors and e- mail. Along the  way, network officials  misjudged their  audience.  Initially,  the  network tried  charging US$2 for  each home listing  on its  Web site.  But when competition charged less, it cut the price until it  stopped charging for  listings  at all.  Meanwhile, advertising for the site, which now has about 350,000 listings,  never materialized. At the  same time,  the proprietary system for agents bombed. In New Jer- sey, a pilot state, fewer than 1000 New Jersey Asso- ciation of Realtors' 36,000 members chose to  sub- scribe,  says Michael Ford, the  state group's presi- dent-elect and a national association director. Only four of the state's 18 multiple-listing services posted their listings on it [28].  This example highlights the need to understand the needs of both product/service providers, as well as consumers in  a market, espe- cially when start-up costs are in  millions. Market participants should not be charged anything to sub- scribe to a new electronic market until a large num- ber of product/service providers and consumers are participating, and both sides  see  the  value of the e-market. With limited  revenues  at the  beginning, new electronic markets need to  tightly  control the costs to create and run the e-market.
7.2. Determinants of electronic market impact
  Based on an analysis of the  e-markets described throughout  this  paper, it is  apparent that  electronic markets will impact some industries  more than oth- ers.  The question then  is:  What are  some  of the factors that determine this level of impact? We have identified six factors that each fall within one of four categories: product, industry, seller and consumer.
7.2. I. Product characteristics
  First, the form of the product is important. Digiti- zable products are particularly suited for electronic markets because they not only take advantage of the digitization of the  market mechanism, but also the distribution mechanism, resulting  in very low trans- action costs. It also enables the order fulfilment cycle time to be minimized. Examples of digitizable prod- ucts were described earlier. Second, the  magnitude of the product price may be an important determi- nant. The higher the  product price, the  greater the level  of risk involved in the  market transaction  be- tween  buyers and  sellers  who are  geographically separated and may have never dealt with each other before. Some of the  most common items currently sold through e-markets are low-priced items such as CDs and books.
7.2.2. Industry characteristics
  An industry  factor that  affects the  impact of e- markets is  the  level  of standards  that  exists in  an industry for describing products. A lack of available standards that both the buyer and seller recognize is a barrier to consummating sales electronically. Cur- rent description standards would generally be textual, but future  standards  could include  multimedia op- tions. As multimedia capabilities such as video, au- dio, and perhaps virtual reality (enabled by the  vir- tual reality markup language, VRML, in the WWW), are incorporated  into electronic market interfaces it will become easier to describe products to potential consumers.  A second industry  characteristic is  the need for a transaction broker. Electronic markets are most useful when they can directly match buyers and sellers. Industries that require transaction brokers, or third  parties,  may be  affected  less  by  electronic markets than  are  industries  where no brokers are required. Stock brokers, insurance agents, and travel agents may provide services that are still needed, but in  some cases software may be able to  replace the need for  these  brokers. This is  especially true  as more intelligent  systems that  assist consumers be- come available.
7.2.3. Seller characteristics
  E-markets reduce search costs that enables con- sumers to  find sellers offering lower prices. In the long run, this reduces profit margins for sellers that compete in e-markets, although it may also increase the number of transactions that take place. If sellers in  an  industry  are  unwilling to  participate in  this environment, then  the  impact of e-markets may be reduced.  In highly competitive industries,  with low barriers to entry, sellers may not have a choice. But, in  oligopolistic situations, sellers may determine the success of e-markets in an industry  if they want to maintain an environment of lower  volume, higher profit margin, transactions.
7.2.4. Consumer characteristics
  Consumers can be classified as either impulse, patient or analytical. Impulse buyers purchase prod- ucts quickly with little analysis, patient buyers pur- chase products after making some comparisons, and analytical buyers do substantial research before mak- ing  the  decision to  purchase  products or services [15].  Electronic markets may have little impact  on industries  where a  sizable percentage of purchases are made by impulse buyers. An example of this is grocery store purchases. A high percentage of sales in  these  stores  is  on  impulse.  Because electronic markets require a certain degree of effort on the part of the  consumer, these  markets are more conducive to consumers who do some comparisons and analysis before buying (the patient or analytical buyers). Ana- lytical buyers can use the  facilities available to ana- lyze  a  wide range  of information  before deciding where to buy.
  The determinants discussed provide a framework for estimating the  impact of e-markets on current or future industries. The more industry features (includ- ing product, industry, seller, and consumer character- istics)  associated with higher e-market impact,  the greater the  expected impact  of e-markets  on that industry. For example, the newspaper industry should be highly impacted by electronic markets. Newspa- pers  are  easily  digitizable, low-priced,  easily  de- scribed, and do not require a broker. It is uncertain whether publishers would want to  compete in  the lower profit margin environment, but they may not have a choice. What may reduce the  impact on this industry  is  that  consumers  generally  do  not  go through  an analytical process when buying newspa- pers. What may evolve is the equivalent of electronic newspaper subscriptions where readers pay a fee to receive access to a daily online newspaper, so they do  not need  to  complete a  daily transaction.  Al- though,  daily  purchases  of electronic  newspapers would enable consumers to buy only when they have the  time  or the  need for the  information. This same type of analysis holds for any industry in determin- ing the  expected level of e-market impact.
7.3. Ot,erall conclusions
  The main issue  we addressed  throughout  this paper was: are there  economic incentives  for elec- tronic commerce, or is it just hype? In particular we looked at the impact of electronic markets on buyers, sellers, and other organizations that support commer- cial transactions.  It  is  apparent that  there  are eco- nomic incentives  for these  entities to  participate in e-markets, although they  may introduce new forms of transaction risk. The benefits that  buyers receive from  lower  prices and  search  costs are,  in  many instances,  more than  enough to  offset the  potential additional risk, distribution and market costs. Sellers benefit from  potential new sources  for  revenue  as well as a reduction of many production and transac- tion-oriented costs. The impact of e-markets on other organizations that support commercial transactions is more mixed. They reduce the need for certain types of intermediaries  (such  as  wholesalers and  retail stores) while increasing buyer demand for new inter- mediaries (ISPs, online better business bureaus, and so forth). They also have potential tax  revenue col- lection implications for state and federal government.
  Our overall conclusion is that there are economic incentives  (both from reduction of costs as well as creation of new revenue sources) for electronic com- merce (and in particular electronic markets). It is not just a fad that will go away. Electronic markets are a new  institution  of capitalism and  they  are  useful because they economize on transaction costs in many instances when compared with other available trans- action governance mechanisms (such  as traditional markets).
References
[1] K. Rebello, L. Armstrong, A. Cortese, Making Money on the Net, Business Week, September 23, 1996. [2] O.E. Williamson, The Economic lnsitutions of Capitalism, Free Press, New York, 1985. [3] T.W. Malone, J. Yates, R.I. Benjamin, Electronic Markets and Electronic Hierarchies, Communications of the ACM 30, No. 6, June 1987. [4] T.W. Malone, J. Yates, R.I. Benjamin, The Logic of Elec- tronic Markets, Harvard Business Review, May-June 1989. [5] T.W. Malone, J.F. Rockart, Computers, Networks and the Corporation, Scientific American 265, No. 3. September 1991. [6] J.Y. Bakos, A Strategic Analysis of Electronic Marketplaces, MIS Quarterly 15, No. 3, September 1991. [7] V. Gurbaxani. S. Whang, The Impact of lnt~rmation Systems on Organizations and Markets, Communications of the ACM 34, No. 1, January 1991. [8] R. Benjamin, R. Wigand, Electronic Markets and Virtual Value Chains on the Information Superhighway, Sloan Man- agement Journal 36, No. 2. Winter 1995. [9] J.F. Rayport, J.J. Sviokla, Exploiting the Virtual Value Chain, Harvard Business Review, Nov.-Dec. 1995. [10]  S. Jarvenpaa, P.A. Todd, Consumer Reactions to Electronic Shopping on the  World Wide Web, Int. J. Electron. Com- merce 1, No. 2 (1997). [l 1] P. Berthon, L.F. Pitt, R.T. Watson, The World Wide Web as an Advertising Medium, J. Advertising Res. (Jan.-Feb. 1996). [12]  D. Burstein, D. Kline, Road Warriors, Dutton Books, New York, 1995. [13]  D. Spar, J.J.  Bussgang, The Net Rules, Harvard Business Review, May-June 1996. [14] J.A. Quelch, L.R. Klein, The Internet and International Mar- keting, Sloan Management Rev., Spring 1996. [15]  R. Kalakota, A.B. Whinston, Frontiers of Electronic Com- merce, Addison-Wesley Publishing, Reading, MA, 1996. [16]  L. Applegate, F.W. McFarlan, J.L.  McKenney, Corporate Information  Systems Management: Text and Cases, Irwin, Chicago, 1996. [17]  R.E. Calem, Auto Sales Are Booming on the Web, The New York Times: CyberTimes, Oct. 17, 1996. [18]  Yahoo. Business and Economy: Companies: Automotive: Internet Marketplaces. [19]  Welcome to  Amazon.corn  Books!,  URL = http://i www.amazon.com/. [20]  M. Glaser, E-Zines Find Road to Profits Is Long, Bumpy and Winding, The New York Times: CyberTimes, Oct. 18, 1996. [21]  D. Rosato, Airlines Turn Internet into Discount Haven, USA Today, May 7, 1996, URL = http://www.usatoday.com/. [22]  L. Eaton, Slow Transition in  Investing as Market Meets Internet, The New York Times: CyberTimes, Nov. 11, 1996. [23]  T.H. Davenport, Process Innovation:  Reengineering Work through Information Technology, Harvard Business School Press, Boston, 1993. [24]  Netscape  Communications Corporation,  URL = http:// home.netscape.com/. [25]  McAfee Associates, URL = http://www.mcafee.com/. [26]  L. Simpson, H.B. Lakner, Perceived risk  and mail order shopping for apparel, J. Consumer Studies and Home Eco- nomics 17 (1993[27]  S.E. Goodman, L.1. Press, S.R. Ruth, A.M. Rutkowski, The Global Diffusion of the  Internet:  Patterns and Problems, Communications of the ACM 37, No. 8, Aug. 1994. [28]  National Association of Realtors, Realtors Information Net- work, URL = http://www.realtor, com.