In the third wave of the Internet Economy Indicators research (released in June, 2000), we made a distinction between two types of dot com companies - digital and physical. Digital dot coms are Internet based companies such as Yahoo, Ebay and America Online, whose products and services are digital in nature and are delivered directly over the Internet. By contrast, the physical dot coms sell physical products (e.g., books, CDs, toys) that are shipped to consumers. Based on this classification, we found that for 1998, Information Technology (IT) capital (computer hardware, software and networking equipment) contributed significantly to financial performance measures for the digital dot coms. However, the IT contribution for physical dot coms was uniformly insignificant across all financial measures. We attributed the result involving physical dot coms to the fact that beyond the digital customer interface, such companies were not fundamentally different from traditional business operations. In fact, we found that nearly 80 percent of the physical dot coms held and managed inventory, and handled packaging and shipping processes by themselves, citing customer service excellence as the primary reason. By contrast, the digital products companies manage content inventory directly through their Web sites and related applications.
We argued that the physical dot coms need to adopt more digitization in their extended value chain in order to operate as virtual organizations, whose core competency was customer relationship and knowledge. Did the scenario change for 1999 and 2000? Our analysis for 1999 suggests that both publicly traded digital and physical dot coms improved their financial performance significantly over 1998, and are headed in the right direction at a rapid pace. Average annual revenues increased 261 percent and 313 percent for digital and physical dot coms respectively, implying spectacular growth of online commerce activity. Perhaps more impressive was the growth of gross income, which increased 150 percent and 307 percent for the digital and physical dot coms respectively. The number of employees increased by 78 percent and 185 percent for digital and physical dot coms respectively. Sales per employee grew at 65 percent and 21 percent for digital and physical dot coms respectively. These findings are summarized in the tables below:
Publicly traded digital dot coms: 1998 - 1999
Publicly traded physical dot coms: 1998 - 1999
Were the physical dot coms productive in their use of labor, Information Technology and non-IT capital? In contrast to our earlier result that the physical dot coms did not benefit from their investments in IT capital, we find that IT did contribute significantly to revenue and revenue per employee for this group of companies. Our interpretation is that these companies are getting better at leveraging the technology to build customer relationships and in coordinating with fulfillment partners. That is, they are now relying more on others for logistics and fulfillment functions, and are focusing on their core business of creating value for the customer through the Internet.
The analysis also suggested that both digital and physical dot coms are rapidly improving their performance over time. For example, even if the digital dot coms did not increase their number of employees, IT and non-IT capital, they would still enjoy a 21.2 percent increase in sales in one year. They would also experience increases of 24.47 percent, 27.22 percent and 28.4 percent in revenue per employee, gross income and gross income per employee respectively. The physical dot coms are expected to see corresponding improvements of 20.4 percent, 21.28 percent, 29.67 percent and 25.7 percent in their revenue, revenue per employee, gross income and gross income per employee respectively. These numbers suggest that both types of dot coms are getting better at selling, fulfilling and providing customer service without increasing the amount of resources such as people or capital. This is certainly a positive signal amidst pessimism regarding the future of dot coms.
What about dot com performance in 2000?
Given the precarious conditions that exist in the venture capital market today, it is natural to ask how dot coms have performed in 2000. While the above analysis was restricted to publicly traded dot com companies, most dot coms are small, privately held companies. We identified a total of 3125 "dot com" companies in the U.S. Internet economy, defined as layer 3 or 4 companies who do at least 95 percent of their business online. While 100 percent online sales may appear to be a more appropriate criterion for defining a dot com, we note that even pure Internet based companies receive some phone and fax based orders. Accordingly we set the online sales threshold at 95 percent. From the first to the second quarter of 2000, the total revenues associated with these dot coms increased a healthy 18.7 percent, while gross profit margin increased 1.3 percent from 33.1 percent to 34.4 percent. By contrast, a "top-ten" list of traditional companies (chosen from retail, financial services, hardware, software, telecommunications and aerospace sectors) experienced a 5.1 percent reduction in gross margin from 34.5 percent to 29.4 percent from the first to the second quarter of 2000, indicating a possible slowdown in the overall economy. Under these circumstances, the 1.3 percent increase in gross profit margin for all dot coms tells a very different story from the one usually portrayed in the media based on the failure of a few large dot coms. Indeed our analysis for 1999 and 2000 suggest that as a group, the dot coms have performed admirably. While a group of dot coms with bad business models have already gone out of business, it is unreasonable to suggest that all dot coms will shut down in the near future. It is ironic and unfortunate that the capital markets have virtually closed their doors to the entire group in spite of spectacular growth in revenue, revenue per employee, gross income and an overall improvement in the productive use of resources.
Ying Fang and Anjana Susarla collaborated with Drs. Whinston and Barua on this research.
The Research Team
Professor Andrew Whinston
Dr. Andrew Whinston is the Hugh Roy Cullen Centennial Chair Professor in Information Systems at the Graduate School of Business in the University of Texas at Austin. He is a Professor in the departments of Economics and Computer Science as well. Dr. Whinston is a Fellow of the IC2 Institute, Austin and is the director of the Center for Research in Electronic Commerce (http://crec.bus.utexas.edu), for several years a pioneering research facility in Electronic Commerce. Under his stewardship, the Center identified the potential of electronic commerce early on, and made significant contributions in theoretical aspects of business and technological practice in this new frontier, and developed cutting edge applications that facilitate and demonstrate strategies for this marketplace. His current research spans various realms of Electronic Commerce, its impact on business protocols and processes, on organizational structure and corporate networks, electronic publishing, electronic education, and business value of IT. He is the author or co-author of over 300 technical papers and 30 books. His recent book titles are "Frontiers of Electronic Commerce," "Readings in Electronic Commerce," "Electronic Commerce: A Manager's Guide," "The Economics of Electronic Commerce," "Handbook on Electronic Commerce," "The Internet Economy: Technology and Practice," and "Electronic Commerce and Revolution in Financial Markets."
Dr. Whinston received his Ph.D. in management from Carnegie Mellon University in 1962. He served as an Assistant Professor at Yale University from 1961 to 1964 in the Department of Economics and the Cowles Foundation for Research in Economics. In 1964 he joined the faculty of the Economics Department at the University of Virginia as an Associate Professor and in 1966 was appointed as Professor in the Business School at Purdue University with a joint appointment in Computer Science.
Dr. Anitesh Barua
Anitesh Barua is an Associate Professor of Information Systems, Spurgeon Bell Fellow, Bureau of Business Research Fellow, Distinguished Teaching Professor and Associate Director of the Center for Research in Electronic Commerce (http://crec.bus.utexas.edu) at the Graduate School of Business, the University of Texas at Austin. He received his Ph.D from Carnegie Mellon University in 1991. His research interests include the business value of Internet related Information Technologies, measuring economic aspects of the Internet Economy, and the efficiency of electronic markets. His research has been sponsored by both government and private organizations including the National Science Foundation, Cisco Systems, Dell Computer, Ernst & Young, IBM Research, and Intel Corporation. 45 of his research articles have appeared in refereed journals, conference proceedings and edited book chapters. Dr. Barua has received several awards for his research and teaching, including the William W. Cooper Doctoral Dissertation Award from Carnegie Mellon University in 1991, National Science Foundation Research Initiation Award in 1992, the University of Texas College of Business Administration (CBA) Foundation Teaching Award in 1994, ANBAR Citation Excellence in 1997, the CBA Teaching Innovation award in 1999, membership in the University of Texas Academy of Distinguished Teachers in 2000, and the El Paso Energy Foundation Faculty Award in 2000.
Dr. Barua has appeared on multiple occasions as an Electronic Commerce expert on major television and radio programs in the U.S. and abroad including ABC, CNN, CNBC, Jim Lehrer NewsHour, Associated Press Radio, CBS Radio and Voice of America. Dr. Barua has also appeared as an expert witness before the House Ways and Means Committee, and has briefed the staff of the Joint Economic Committee on the Internet economy.
Jay Shutter is President of the Momentum Research Group (MRG), A Citigate Cunningham Company. The Momentum Research Group specializes in high technology research and has a practice that focuses on brand momentum and the impact of the Internet Economy on business strategy, brand strategy, marketing and communications.
At Citigate Cunningham, Jay is responsible for developing new research products and services that address the impact of technology and the Internet on businesses and consumers. Formerly Director of Research at InformationWeek and prior to that, Director of Research at IntelliQuest, Jay developed the technology industry's first business research panel as well as an industry leading customer registration business. A frequent presenter at research and marketing conferences, Jay is a recognized expert in such areas as panel research, database research, segmentation studies, brand tracking and Internet research.
Jay has degrees from the University of Texas at Austin and Southern Illinois University at Carbondale.
Dr. Brant Wilson
Dr. Brant Wilson, a recognized expert at marketing research and analysis, is Founder and President of Customer Value Systems, Inc. Formerly Director of Business Planning at Compaq Computer Corporation, Dr. Wilson founded Compaq's worldwide marketing research department, expanding it to strategic planning function. In 1988, he developed the computer industry's first electronic customer registration program at Compaq and pioneered a highly successful customer tracking system. More recently his targeting strategies have helped CVS clients achieve more than four times the expected response rate on promotions, leading to dramatic profit improvements.
Dr. Wilson is a frequent speaker at marketing and business conferences, on such topics as product strategy and positioning, brand loyalty, niche marketing, customer segmentation and profitable marketing strategies.
Prior to working at Compaq, Dr. Wilson held senior positions with McCann-Erickson Advertising and Anheuser-Busch, Inc. the world's leading brewer. Dr. Wilson received a B.S. from the University of Richmond, a Masters at Texas A&M University, and a Ph.D. from Washington University in St. Louis.
Jon Pinnell is President and Chief Operating Officer of MarketVision Research where he has primary responsibility for research operations and the firm's Cincinnati offices. MarketVision Research is a full service research supplier who provides innovative design, execution and interpretation to deliver fresh consumer insights and market knowledge to clients in North America and worldwide.
In addition to his corporate responsibilities, Jon leads the firm's Marketing Sciences Group. Jon's areas of expertise and research interest include the application of quantitative techniques to marketing models of decision making, especially those areas relating to brand and category choice decisions.
Jon is actively involved in pioneering new research methods. In 1995 Jon was a co-recipient, with Valentine Appel, of the Richard Lysaker Award, presented by the Advertising Research Foundation for the best new research on media.
Jon has presented papers at leading conferences for practitioners and academics. Before joining MarketVision, Jon was the Director of Marketing Sciences at IntelliQuest, Inc., the leading research supplier for the technology industry in Austin, TX. Jon received a BS from the University of Kansas in Marketing and Statistics and an MBA from the University of Texas at Austin in Marketing Research.
Appendix A: Defining the Internet Economy
Research in Electronic Commerce has almost exclusively focused on the number of Internet users, demographics, and various aspects of online buying and selling. However, as we looked at how to better understand the impact of the Internet across all business sectors, our focus turned to measuring the size and growth of the Internet Economy. Our first challenge was to determine how to define this emerging economy.
The first step in defining the Internet Economy was to build a conceptual framework and taxonomy. The Internet Economy can be conceptualized as a collection of IP-based networks, software applications and the human capital that makes the networks and applications work together for online businesses, and agents (corporations and individuals) who are involved in buying and selling products and services in direct and indirect ways. There is a natural structure or hierarchy to the Internet Economy that can be directly traced to how businesses generate revenues. Based upon this type of structure, we broadly classify the Internet Economy into infrastructure and economic activity categories.
The infrastructure category is further divided into two distinct but complementary "layers": the Internet infrastructure layer, which provides the physical infrastructure for Electronic Commerce, and the Internet applications infrastructure, which includes software applications, consulting, training and integration services that build on top of the network infrastructure, and which makes it feasible for organizations to engage in online commerce.
The economic activity category is also subdivided into two layers: electronic intermediaries and online transactions. The intermediary layer involves the role of a third party in a variety of capacities: market maker, provider of expertise or certification that makes it easier for buyers to choose sellers and/or products, search and retrieval services that reduce transaction costs in an electronic market, and other services that facilitate the conduct of online commerce. The transactions layer involves direct transactions between buyers and sellers like manufacturers and e-tailers.
While we could take the position that e-tailers are also an intermediary between the consumers and manufacturers, the illustration above highlights the difference between an e-tailer like Amazon.com and an electronic intermediary in the purest sense of the term. An intermediary would also specify where a book or CD could be found at the lowest price or shortest delivery time or some combination of criteria specified by the consumer. By contrast, Amazon.com only displays its own catalog, prices, availability and lead-time. Of course, it is true that for the case of e-tailers the difference between the two topmost layers could be a matter of degree.
Layer One: The Internet Infrastructure Indicator
A physical economy critically depends on an efficient infrastructure involving transportation, energy, raw materials, and skilled workforce. Likewise, the growth of a digital economy depends on the ubiquitous presence of high speed and intelligent electronic networks, and the ability to share any type of content between all agents in the economy. Accordingly, the Internet infrastructure layer includes companies that manufacture or provide products and services that make up the Internet network infrastructure. This layer includes companies that provide telecommunications and fiber backbones, access and end-user networking equipment necessary for the proliferation of Internet-based Electronic Commerce. This layer includes the following types of companies:
Layer Two: The Internet Applications Infrastructure Layer
Products and services in this layer build upon the above IP network infrastructure and make it technologically feasible to perform business activities online. In addition to software applications, this layer includes the human capital involved in the deployment of Electronic Commerce and E-Business applications. For example, Web design, Web consulting, and Web integration are considered as a part of this layer. This layer includes the following categories:
Layer Three: The Internet Intermediary Indicator
Internet intermediaries increase the efficiency of electronic markets by facilitating the meeting and interaction of buyers and sellers over the Internet. They act as catalysts in the process through which investments in the infrastructure and applications layers are transformed into business transactions. While much has been written about a large-scale disintermediation in the transformation of the physical to the digital economy, the Internet necessitates a new breed of intermediaries whose roles are naturally information and knowledge intensive.
In the physical world, intermediaries are distributors and dealers, whose primary role is to increase the efficiency of distribution and to lower buyer transaction costs by locating close to the customer population. By sharp contrast, physical proximity is not an issue on the Internet; online search, evaluation, communication, coordination, assurance of vendor and product/service quality are the important aspects in the Internet Economy. Internet intermediaries play a critical role in filling information and knowledge gaps, which would otherwise impair the functioning of the Internet as a business channel. This layer includes:
Layer Four: The Internet Commerce Indicator
This layer includes companies that generate product and service sales to consumers or businesses over the Internet. This indicator includes online retailing and other business-to-business and business-to-consumer transactions conducted on the Internet.
It is important to note that many companies operate at multiple layers. For instance, Microsoft and IBM are important players at the Internet infrastructure, applications, and Internet commerce layers, while AOL/Netscape has businesses that fall into all four layers. Similarly, Cisco and Dell are important players at both the infrastructure and commerce layers. Even though the four-layer Internet Economy framework makes it time-consuming to separate revenues for multi-layer players, the framework presents a more realistic and insightful view of the Internet Economy than a monolithic conceptualization that does not distinguish between different types of activities. Further, the multi-layered approach lets us analyze how companies choose to enter one Internet layer, and later extend their activities to the other layers.
Each layer of the Internet Economy is critically dependent on every other layer. For instance, improvements in layer one can help all other layers in different ways. As the IP network infrastructure turns to broadband technologies, applications vendors at layer two can create multimedia applications that can benefit from the availability of high bandwidth. Companies at layers three and four can benefit from improvements in both layers one and two - providing media-rich content to consumers as well as new digital products and service (information and software delivered online). This interdependence also exhibits itself in the form of alliances where conduit and content providers or applications vendors and e-tailers join hands to create bundled offerings that are valuable to consumers.
Appendix B: The Growth of the Internet Economy
To appreciate the unprecedented growth and implications of the Internet economy, we need to address specific issues related to the nature of transformation enabled by the Internet. These are discussed below.
How do we understand a technological breakthrough such as the Internet? What are the essential differences in understanding the Internet economy and the bricks and mortar economy?
As enumerated in our previous research (see http://www.internetindicators.com/), the Internet economy's infrastructure consists of two layers: global high-speed IP-based networks and applications (layer 1), and consulting, training, and integration services (layer 2). Each Internet economy layer has a complementary relationship with every other layer. For example, with advances in layers 1 and 2, firms at layers 3 and 4 can provide media-rich content to consumers as well as offering new digital products and services (information and software goods that are delivered online). This interdependence also exhibits itself in the form of alliances in which conduit and content providers or applications vendors and electronic retailers (e-tailers) join hands to create bundled offerings that are valuable to consumers.
How do we understand the wide range of consequences that the Internet has across various sections of the economy?
The Internet is not an isolated technology phenomenon; it is a "general purpose technology", at the heart of which are sweeping transformations affecting virtually all forms of economic and societal organization. Many economists point out that a major technological breakthrough such as computing or electricity comes into widespread use only after the emergence of an extended trajectory of incremental innovations. Pervasiveness of a General Purpose Technology (GPT) is its ability to be used in many downstream applications because it provides a generic function, such as the developments in communication technologies that can be used by households, for broadband communication within industries and a variety of online stores.
The technological dynamism of a GPT is its potential to support continuous innovation efforts and learning, which allows for large increases in the efficiency of the GPT over time. As suggested by economic historians such as Rosenberg, inventions such as the railroads had widespread consequences on the organization of the economic systems, bringing non-local markets into the purview of producers and reorganizing trade. What would be considered a perishable commodity could become a non-perishable commodity with the advent of the railroad. Similarly the Internet has created services out of products. The Napster-Bertelsmann agreement, it is pointed out, has far reaching consequences not only on the music industry, but on the entire popular perception of music.
Is there a historical precedent for the major technological innovations to go through a period of experimentation before they gain widespread acceptance?
Let us compare the technological breakthrough of today, the Internet, with the breakthrough of the nineteenth century, the electrical dynamo. Farsighted engineers had envisioned the role of electricity in transforming the organization of home and workplaces at the turn of the century, but it took a long time before electricity came into widespread use.
Historically, factories were reluctant to adopt electricity because of the seemingly unprofitable replacement of still serviceable capital for new production techniques. This accounts for the lags in diffusion. It is not enough if there are a few visionary entrepreneurs who can see the potential of a new technological breakthrough. For all the complementary innovations could be carried out, it required the creation of a new generation of workforce, organizations learn the new technology and this creates valuable human capital. This spills over into other firms as the entire economy can learn from one firm's example. Now in the present, companies may still prevail with pre-Internet era procurement and customer servicing.
The existence of strategic complementarities in the adoption of new GPTs by the various sectors of the economy may generate temporary lock-ins or create implementation cycles (Aghion and Howitt, 1998a). The arrival of the networking infrastructure and software applications will induce firms to engage in risky experimentation on a large scale with start up firms, not all of which will succeed. There is also the phenomenon of 'social learning'. The way a firm typically learns to use Internet technology is not by discovering the merits of the Internet on its own. More often than not, a firm learns from the experience of other firms in a similar situation, other firms who have successfully overcome the glitches of implementing Internet enabled business processes. Preliminary analyses of successful adopters of Internet technology tend to support the notion of there being technology 'leaders' and 'followers' (see http://crec.bus.utexas.edu/ for the study entitled "E-Business Value Assessment").
The evolution of the Internet is not an incremental process due to the complementary organizational and technological changes needed. The Internet ecosystem carries with it radical innovations, and smaller innovations -defying any attempt to understand its evolution in a smooth and gradual way.
Can we get an accurate assessment of the consumer value of the Internet?
Many critics of the new economy interpret the productivity of the Internet in terms of falling prices of computers. Their study concludes that the decline in computer prices triggered a massive substitution of computer services for other inputs. However, we need to go beyond the phenomenon of declining prices. A historical example of this could be found in the consequences of the introduction of electric power in factories, which cannot be modeled as a response to a change in the price of power designed to reflect the technological requirements of steam. Even a zero price of steam could not have led to the radical redesign of the plant, which is the main source of efficiency gain under electricity (David, 1991). With Internet technology there is an apparent productivity gain because the Internet causes businesses to substitute net based transactions for processes that were performed by labor, say in keying in of purchase orders. To the extent that we substitute physical transportation for digitally encrypted transmission of information, there is a productivity gain. The true value of Internet technologies comes from the reorganization of industry to tap the full benefits of the Internet, the manifold consumer benefits to transacting on the Internet, and the way that the Internet can alter coordination and communication within firms.
Major innovations such as the railroad, the dynamo or the steam engine were all technological leaps that profoundly altered the structure of economy and society. A measure of changes brought about by such changes should include the changes in economic structure and rates of growth across structural transformations before one can arrive at a true picture.
How is the growth of the Internet different from that predicted by economic growth theories?
Traditional growth theory is based on the factor substitution ideas that we have outlined earlier. Growth theorists argue that the reason for economic growth is learning by firms, capital accumulation in the economy or incremental technological change. Economic growth is only a result of better technology or capital accumulation. While this is certainly true, what these models are missing, in a sense, is the extent to which these processes are interdependent and acting in tandem to influence business processes. In theories of economic growth, markets are not a variable. The tremendous changes in the organization of markets, electronic commerce networks and B2B markets, to name but a few, are not variables that are analyzed by growth theorists. In the more recent theories of endogenous technological change, the process of accumulation of knowledge is assumed to be incremental. Thus, these theories seem to build a picture of an economy where market agents strive to improve what already exists.
On the other hand, we argue that there can be short-term lags and lock-ins before the Internet can fully realize its potential. The long run economic growth in the digital economy is not only driven by information acquisition capabilities enhanced by the Internet, but also by the dynamic interaction between the Internet and market structures. With the reduction in communications cost, there is more coordination - a process facilitated by the trajectory towards open standards. For the individual customer, the greater opportunities for search and customization of product offerings makes for greater convenience and satisfaction. Moreover, the Internet acts as an efficient market infrastructure that reduces asymmetric information and aids in the diffusion of innovation, thus serving to enhance the productive capacity in the economy. Consumers and producers both gain from the enhanced productive capacity in the economy, that manifests itself as greater output and higher consumption rates at the aggregate level.
Who are the beneficiaries of the Internet economy?
The unique value proposition of the Internet lies in its ability to mitigate uncertainty and asymmetric information. Economic decisions are characterized by time lags and uncertainty about the future. Effective coordination requires aligning the incentives of agents who are separated spatially and temporally. As we argue, the Internet can help improve the productivity of the entire economic system, and not just enhance the internal productivity of the firm.
An analysis of the productivity of the Internet technology must include the consumer benefits of such advances in networking technologies such as the ATM. As we have argued earlier, advances in Internet infrastructure is the engine to the enormous benefits of Internet technology, many of which are yet to be realized. The process of development of complementary innovations, i.e. layers 3 and 4 is still underway. The future will see far more advances in Internet technology enabling smarter appliances, more efficient financial intermediaries, growth of B2B markets and yet to be discovered uses of Internet technology that has altered and will continue the surge in productivity brought about by the Internet.
Appendix C: Researching a Large Company - A Case Study
Measuring the Internet Economy requires strong knowledge of traditional market research techniques, expertise in technology markets and a firm understanding of the Internet itself. While this would seem to be relatively straightforward, it was not. The unique challenge of this study lies in the nature of the data to be collected. Since many companies have not even begun to think about tracking their Internet revenues separately, the process for collecting solid estimates requires even more careful preparation than the typical market research project.
The following case study outlines the process used to collect information from some of the larger companies covered in the project. This case study presents the steps taken to prepare for the interview and highlights challenges that may arise in a typical interview.
Appendix D: Frequently Asked Questions
What is the Internet Economy?
The Internet Economy is made up of companies directly generating all or some part of their revenues from Internet or Internet-related products and services. These companies are the Internet infrastructure and Internet applications players, such as Cisco, Dell, IBM, HP, Oracle, Microsoft and Sun Microsystems, whose products and services make it feasible to use the Internet for electronic commerce. For example, IBM sells servers and PCs that are used to gain access to the Internet. Similarly, 3Com sells modems and Cisco sells routers, all used to gain access to the Internet. The Internet-related revenues from these companies, for example, are included in the estimates for the entire Internet Economy.
Then there are companies selling products and services over the Internet. This includes pure Internet-based sellers like Amazon.com and eToys.com as well as bricks-and mortar and catalog companies, such as LL Bean and Alaska Airlines, who are also conducting a part of their business on the Internet. Further, electronic intermediaries or Internet middlemen (such as eBay) act as catalysts by facilitating the interaction between buyers and sellers. So the overall Internet Economy is made up of the revenues of infrastructure and applications players, electronic intermediaries and online sellers.
It is important to point out that the Internet Economy is not just a collection of "high-tech" companies. It includes any company that generates revenues from the Internet. For example, a part of the revenues generated by traditional telecommunications companies are counted in this economy since they carry IP traffic over their miles and miles of copper, coaxial and fiber lines. We did not, however, count all revenues from all technology companies. Not even 100 percent of Cisco's revenues were considered to be part of the Internet Economy, since all networking devices are not attached to the Internet.
Q. How did you measure the Internet Economy?
There were two important steps to how we measured the Internet Economy. The first was to develop a conceptual framework and taxonomy for attributing revenues and employees in the Internet Economy. Since the manner in which businesses operate in the Internet Economy can vary substantially, we decided to group companies based upon core business activities.
The Internet can be thought of as a network of networks, made up of many component parts. This may consist of networking hardware, networking software, servers, PCs, Web software, Web designers, Web operators, and the companies that are actually doing business on the Web. There is a natural structure or hierarchy to the Internet that can be directly traced to how businesses generate revenues. Based upon this type of structure, we use the following four-layer architecture to group companies, revenues and employees. We call these four layers the Internet Economy Indicators.
Layer 1: The Internet Infrastructure Indicator - the gross revenues and attributed employees from companies that manufacture or provide products and services that make up the Internet network infrastructure. This layer includes companies that provide telecommunications and fiber backbones, "last mile" access, Internet dial-up access and end-user networking equipment necessary for the proliferation of Internet-based Electronic Commerce. It also includes PC and server manufacturers, modem manufacturers and other manufacturers of hardware necessary for the Internet to function.
Layer 2: The Internet Applications Indicator - the gross revenues and attributed employees from companies that provide electronic commerce applications (e.g. IBM, Microsoft, Sun), Internet consulting services (e.g. MarchFIRST, Scient), multimedia applications (e.g. RealNetworks, Macromedia), Web development software (e.g. Allaire, Vignette), search engine software (e.g. Inktomi, Verity), Web-enabled databases (e.g. Oracle, IBM DB2, Microsoft SQL Server, Sybase, Informix), and online training services (e.g. Sylvan Prometric, Assymetrix). This also includes consultants who define and develop Internet applications for their clients (e.g. Cap Gemini E&Y , EDS) and the integration of Internet applications with other IT/enterprise platforms (e.g., IBM, Compaq, Andersen Consulting). Products and services in this layer build upon the network infrastructure and make it technologically possible to perform business activities online.
Layer 3: The Internet Intermediary Indicator - the gross revenues and attributed employees from companies that increase the efficiency of electronic markets by facilitating the meeting and interaction of buyers and sellers via the World Wide Web. This layer includes online brokerages, Internet ad brokers (e.g. DoubleClick, 24/7 Media), portals/content providers (e.g. Yahoo, Excite), market makers in vertical industries (e.g. VerticalNet, PCOrder), content aggregators (e.g., CNet, ZDNet), and online travel agencies.
Layer 4: The Internet Commerce Indicator - the gross revenues and attributed employees from companies that generate product and service sales to consumers or businesses over the World Wide Web and Internet. This layer includes online retailing, pay-to-use content and other business-to-business and business-to-consumer transactions conducted on the World Wide Web and Internet.
The second step we took to measure the Internet Economy was the actual data collection methodology. In-depth research was conducted on companies that participate in one or more of the four layers of the Internet economy. Phone-based interviews were conducted with 1600 of the smaller Internet players. 370 in-depth interviews were conducted with the largest companies playing in the Internet Economy. For the largest companies, we analyzed company annual reports, product literature and web sites to develop a solid understanding of how they fit into one or more of the four layers. Universe estimates for Internet companies were developed from several Web-based databases and final Internet Economy revenues were estimated using a combination of enumeration and statistical sampling and projection techniques.
Finally, extensive secondary research was used and included reports from IDC, Gartner/Dataquest, Forrester, Dell'Oro and other research vendors. The Internet proved to be the most effective source of information on companies involved in the Internet economy.
Measurement Related Questions
Q. How did you handle researching companies that had mergers and acquisitions between Q4 1999 and Q2 2000?
When this occurred we counted the company information as if the companies had existed in the end state for both periods. This way we could provide a consistent comparison. In some cases, the mergers caused revenue to leave the Internet Economy. For example, when a US-based company was sold to a non-US firm, we did not continue to count the revenue after the sale. On the other hand, when a US-based firm bought a foreign company, we counted the incremental revenue in the Internet Economy only for the period in which the revenue was a holding of a US-based company.Q. What accounting rules were used to assess the revenues of large holding companies whose subsidiaries are Internet focused?
Standard accounting rules and procedures were used to evaluate these revenues. In most cases, this meant that if more than fifty percent of the voting shares of an Internet company is owned by the parent company, that share of revenue and related expenses are included in the consolidated statement of the parent company. This allowed us to discern in many cases how much revenue was relevant for analysis, and to avoid double-counting revenues in both the original company and the parent.
In some cases revenue was not reported according to these rules, so we needed to adjust the revenue accordingly. For example, in some cases one hundred percent of the revenue was reported in both the parent statement and the Internet company. In these instances we assured that it was only counted once.Q. Are the estimates restricted to the U.S.?
The estimates are based on the worldwide Internet-related sales of US-based companies. Cisco's foreign Internet sales are included in this number because Cisco is a US-based company. We took a conservative approach to revenue inclusion.Q. Do the Indicators include measurements of Internet usage by number of users or volume of traffic?
The Internet Economy Indicators do not measure Internet activity by user or volume of traffic, but measure Internet activity from a revenue and employee perspective. It is a supply-side data collection approach rather than a demand-side approach.Q. How accurate is the estimate of the Internet Economy?
Based upon our four-layer taxonomy, we believe that the estimates for each layer are very accurate. To compile our estimates, we completed research on over 1600 US-based companies that currently operate in the Internet Economy. In addition, we estimated revenues and employees from 370 of the largest technology companies by reviewing quarterly and annual reports, researching company web sites and product/service offerings and calling to solicit estimates of Internet revenue from the companies directly. Finally, we leveraged existing market data that included secondary research reports from a wide variety of academic sources and business research firms. Appendix E shows the wide variety of sources utilized.Q. How are the four indicators combined into one employment number and one revenue number?
The four indictors are combined into one number by estimating and eliminating the overlap and double counting of revenues across all four layers.Q. Are the Internet Economy Indicators comparable to GDP or GNP?
It is not a straight comparison. The gross domestic product (GDP) or gross national product (GNP) is based on value of goods and services produced in the US, minus any direct or indirect taxes. The Internet Economy Indicators only measure total revenues and jobs.Q. What if a company generates revenues in more than one indicator layer? Are the revenues broken out by layer?
Yes, but there is a complexity in terms of assigning revenues and employees to each layer, since companies don't report revenues or employees this way. Across the board we were faced with the challenge of identifying what portion of a company's revenues were Internet-related. For some companies, it was simple because they were pure Internet companies. Companies like Amazon.com, Vignette, and eBay can be considered one hundred percent Internet companies. In addition, many of these Internet 'pure-plays' only operate within one of the Internet Economy layers.
For other companies, attributing revenues to the Internet was much more complicated. These companies are not one hundred percent Internet companies, and they generate revenues in more than one of the Internet Economy layers. For example, IBM generates revenues in the infrastructure layer (layer one) from its ISP business as well as from PC, server and networking equipment. In the applications layer, IBM generates revenues from it's E-business software and consulting services. In the Internet commerce layer, IBM generates revenues from its commerce site.
For companies playing in multiple layers, in any layer we only count revenues that are attributable to products and services in that layer. In order to combine the revenues across all layers, we take out the Internet commerce revenues for players whose products/services are counted at infrastructure and/or applications layers and who also sell some or all of these online.Q. Does the estimate of the number of employees include Internet-related employees at non-Internet companies?
Yes. The research did not focus solely on 100-percent Internet companies, or companies that generated 100 percent of their revenues from the Internet. We looked at all companies that were generating all or some part of their revenues from Internet or Internet-related products or services. Therefore we counted Internet-related employees from traditional bricks-and-mortar companies like Barnes and Noble or Lands End.Q. Is it double counting to include finished PC's and servers as well as their operating systems? Won't the value of the operating systems (with a mark-up) be reflected in the price of the box?
Yes. For example, we only counted revenues associated with servers and desktops leaving out operating systems. Similarly we did not count OEM modems as they are a part of the PC, and took care to consider branded modems that are sold separately. In addition, microprocessors, memory, disk drives and other OEM components are counted in the PC or server, but taken out of the revenues of their original manufacturers. An exception to this is the Electronic Commerce revenue of these OEM manufacturers, which are counted in Layer 4, but not in Layer 1.Q. How do you account for the revenues from computer manufacturers? Do you count all of their revenues?
No. Since we know that not all home PCs are purchased and used for Internet access and because we know that not all corporate PCs have access to the Internet, we adjusted the revenues attributed to the Internet Economy based upon the estimated percent of PCs sold that were purchased for Internet access.
For example, we know from IDC research estimates that approximately 90 percent of all PCs purchased for the home in 1999 are being used for Internet access. In the corporate world, that estimate is approximately 60 percent of all PCs purchased. Therefore, we accounted for PC revenues using the percentage of those PCs that were used for Internet access. These are conservative estimates.Q. Are there examples of companies who have revenues from a product or service counted in more than one layer? If so, how do you account for this in the total number without double counting?
Yes, there are companies that have revenues counted in more than one layer. However, this only occurs between layers one through three and layer four. For example, adjusted revenues from Dell Computer's sales of PCs are included in the infrastructure layer and are also accounted for in the Internet commerce layer due to online PC sales. When we estimate the total revenues from the Internet Economy, we remove from the commerce layer the double counted revenues for any company that also has revenues from the same product or service in the other layers.Q. Explain how this is a global study.
It measures the worldwide revenues of U.S.-based companies.Q. How did you handle changes in data collection and/or research methodology compared to the previous phase of the project?
In the current study, however, the data is collected quarterly, so we gathered comparisons of Internet revenue for 1Q '00 vs. 1Q '99 and 2Q '00 vs. 2Q '99, along with Internet employees during this same period. Over time, these estimates can be updated quarterly and annual totals derived, based on subsequent data collection. It is expected that each phase will engender improvements in the availability of Internet revenue information and better information on the companies participating in the Internet economy.Q. How did you define your universe?
Based on our knowledge of the types and categories of companies that make the Internet possible, we began compiling a list of the companies that we know participate in the Internet Economy. We operated on the following premise: If you are participating in the Internet Economy, you had better be on the World Wide Web.
Using that premise, the Web was the ideal tool for building a database of companies that participate in the Internet Economy. Using the layered conceptual model, we generated extensive lists from Yahoo, Excite, HotBot and other directory services. We searched through the lists to ensure that we did not miss the big players who may account for the larger share of revenues in a subcategory within a layer.
Since most Web directories only provide URL links, we realized that it would be extremely time consuming to link to each company's web site to get their contact information. In order to expedite the development of a contact database, we wrote database code to pull the company name and URL from the search directory listings. We then wrote additional code that matched each URL against the Internet Domain Registry database. What could have taken a potential 1,000 man hours was accomplished in 16 hours using the Web and web-based technology.
Using the same "you've got to be on the Web" premise, we found an extremely valuable database of secure site URLs from a UK-based company called Netcraft. This database was used to create the universe estimate for companies conducting Internet transactions. This was the most appropriate and accurate sampling method to conduct the first large-scale supply-side estimate of Internet commerce revenues.Q. Did you count 100 percent of the revenues from all networking equipment manufacturers?
No. Using available estimates of the percent of corporate PCs that are networked and were purchased for use on the Internet, we adjusted the actual revenues from networking equipment manufacturers attributed to the Internet economy. For example we know that approximately 50 percent of all PCs purchased for businesses in 1998 have Internet connectivity. Using these types of deflators we attributed networking equipment and Internet access device revenues as follows:
Q. Why did you choose to measure sales revenue and employment? What's next? What will it take to create a proper index to measure this Internet economy?
Revenues are relatively easy to measure and understand, and is consistent with previous studies. Jobs are an equally important indictor, but are more difficult to measure because of the overlap between Internet and non-Internet jobs. To create an index comparable to GDP, we would have to measure value added within each of the layers. While it is an extremely complex task, it can be done with the help of sufficient resources and extensive data collection effort.Q. What types of companies contributed most to the Internet Economy?
Given the nascent state of the Internet economy, it is not surprising that a few infrastructure players like Cisco, IBM, Sun, Compaq, Dell and Oracle contributed most significantly to the Internet economy. We expect this trend to continue; however, with the proliferation of the infrastructure, bricks-and-mortar companies selling products and services will most likely play an increasingly significant role at the commerce layer. For example, it is clear that the Internet revenue from companies like Wal-Mart, Citigroup, Sears, UPS, General Electric and other corporate giants will start to impact the Internet Economy and some of this impact even begins to show in the current study.Q. What percent of the Internet Economy revenues is U.S. versus worldwide?
The Internet Economy Indicators focused strictly on U.S. companies and their total worldwide Internet-related revenues. The average breakout of U.S. versus worldwide sales for some of the larger companies in the study is approximately 75 percent U.S. and 25 percent non-U.S. However, since the majority of the Internet Economy is made up of small privately held companies, a much larger portion of the revenue number is U.S. based. We estimate the ratio at 80 percent U.S. and 20 percent non-US.Q. How much effort went into this research?
A large team of research experts from across the United States, using the Internet as a communications and research tool, combined to conduct the research.
The research team consisted of the professors and graduate students from the University of Texas at Austin Business School, research consultants from Citigate Cunningham and outside research experts and analysts, including Dr. Brant Wilson of Customer Value Systems, Inc, a Houston-based marketing research and consulting company. In addition, primary research expertise was brought in from MarketVision Research, which specializes in econometrics-based research. Jon Pinnell, the lead marketing scientist from MarketVision is a recognized expert in research in the technology industry and also in statistical sampling and modeling
Appendix E: Secondary Resources
Overall Internet Economy Trends
Research Firm Offers Map of Digital Economy.
Internet Enabled Job Creation.
The UCLA Internet Report: Surveying the Digital
The Digital Economy Fact Book, Second Edition.
None of Your Business.
Not everyone's e-namoured with the Net.
Privacy Fears Online.
Guarding Their Privates, See?.
I've Got a Secret.
US Firms Invest Heavily in Internet Strategy.
How the Internet Changes (Almost) Everything
Go figure: Web-influenced offline sales dwarf
Where B2C Will Be, 2002.
Net Markets, Positive Net Opportunities.
$6.3 Trillion in B2B by 2005.
Consumer E-Commerce Growth Expectations May Be
Home Office Households Approach 37 Million in 2000 as Internet Turns
"Work Anywhere" into "Work Everywhere".
Enterprises Turn to E-Marketplaces for IT Staff.
Layer 1: The Internet Infrastructure Indicator
Networking at the Speed of Light: Optical technology moves data -
and attracts money - fast.
XDSL and 1 Million Businesses.
The Consumer PC Market: Downhill from Here.
Where Fiber Fits.
DSL Growth Slows, Cable Modems Get Help from
AT&T Makes Major Broadband Deal in China.
Mobile Internet Platforms Emerge.
Streaming Media Use Grows to New
Layer 2: The Internet Applications Infrastructure Indicator
Sizing Supply Network Apps.
Layer 3: The Internet Intermediary Indicator
Dot-Com Sales, Marketing Spending Plunge.
Q2 Online Ads: Another $2 Billion Spent.
American Express' One-Use-Only Web Card Address Online Buying
Online Investing: It's Not Just Buy and Sell.
Web Advertising: Not Just Clickthroughs
Are E-Payment Methods Ready for Global E-Commerce?
Americans - Two-Fisted Media Users.
Layer 4: The Internet Commerce Indicator
Transaction Fees Becoming Important.
eCommerce Grows Up.
B2B Will Hit $1.26 Trillion in 2003.
eAuctions Bid Up Online Revenues.
Less Looking, More Buying.
B-to-B Buying Picks Up.
Special Report: Web Smart 50.
Management Update: The Top 50 E-Tailing Sites Provide Surprisingly
Pure Plays Face Trouble in E-Commerce
Junk That Catalog And Get On The Web: W.W. Grainger's Richard Keyser
tells how he sells factory parts in cyberspace.
Bricks-and-Mortars Challenge Amazon's Lead.
US: Where the Web Shopper Are.
The Great Online Purchasing Migration.
Small Business, Big Online Success.
High Numbers for the High Holidays.
Have Yourself a Scary Little Christmas: eTailers will have to work
like elves to survive the holiday season.
eHoliday Bells Ring: Cha-Ching.
Websites Dress Up, Holiday-Style.
eHoliday Season Begins.
The Information Society Index (ISI) - measuring the global impact of
information technology and the Internet.
Income Disparity Within Web Population.
It's a Woman's World Wide Web: Women's Online Behavioral Patterns
Across Age Groups and Lifestages.
US eMail Usage on the Rise.
US Internet Penetration Reaches High.
Some People Just Don't Like The Net.
The Awkward Age, Online.
The Digital Divide
Digital Divide Dissipating?
African-Americans Narrow the Digital Divide.
The Diminishing Digital Divide.
The Internet Economy Outside the US
Asian Internet Growth - Hyped or Hyper.
Chilly eCommerce in Canada.
eCommerce Enthusiasm Goes Global.
Internet Changing the Way Canadians
European Users: The More They Surf, The More They'll
Putting Worldwide Back in WWW.