Beyond Internet Business-as-Usual

A CHI 98 Workshop

Patrick Steiger, Markus Stolze, Michael Good

Originally published in SIGCHI Bulletin, 30 (4), October 1998, pp. 48-52. Copyright © 1998 by Patrick Steiger, Markus Stolze, and Michael Good.  All rights reserved.

Introduction
A Model of Electronic Commerce
Buyer Support
Open Issues for Buyer Support
Merchant Support
easyBusiness Website and Other Results
Workshop Participants
References
About the Authors
Authors’ Addresses


Introduction

The number of Internet users and the volume of on-line business is growing rapidly. For example, the Spring 97 CommerceNet/Nielsen Media Demographics and Electronic Commerce Study reports that nearly a quarter of the US and Canadian population over 16 years of age (more than 50 million persons) have recently used the Internet — twice the number reported in the Fall of 1995 [4]. Another study by International Data Corp., a research firm in Framingham, Mass, reports that $1.2 billion in goods and services were sold directly over the Web in 1996, a figure which they expect to increase to $91.1 billion by the year 2000 [2]. Despite this success, Internet commerce as we know it today must become better adapted to the growing needs of buyers and merchants on the Internet to better exploit the opportunities of electronic media.

The goal of the workshop was to understand the needs of buyers and merchants in electronic markets, to discuss the state-of-the-art in Internet commerce, and to identify open issues that require more research.

A Model of Electronic Commerce

During the pre-workshop activities, we identified major players and components of Internet marketplaces. Figure 1 depicts the electronic commerce (e-commerce) services in the core of the system wrapped by the user interface, which supports the interaction between market participants and the services as well as among participants. The products and services traded can have many different characteristics. We proposed various ones (such as volatility, tangibility and durability) in a questionnaire sent to all workshop participants. From the responses we concluded that the product properties listed in Figure 1 are the most distinguishing factors.

Model of Electronic CommerceFigure 1: Model of Electronic Commerce: Components

The e-commerce services listed in Figure 1 will be used differently by different types of buyers. For example, buyers who evaluate rationally the advantages of different product features (i.e. rational buyers) need a different kind of support than in a situation in which they know exactly what they need and simply want to get through the ordering process as efficiently as possible (i.e. to act as a habitual buyer). Emotional buyers will be more interested in interactive experiences than hard facts, and social buyers will be looking for exchange with other buyers when shopping. Similarly, merchants will want to offer different functionality depending on whether their main strategy is to offer a large selection of products, individualized customer support, or products at the lowest price.

In addition to the component view of Internet commerce we also developed a process model that lists the main tasks of buyers and merchants. The model builds on work by Schmid [3] and Guttman et al. [1]. Schmid identified three phases of electronic commerce: information, negotiation, and settlement. Guttman et al. compared several consumer buying behavior models and defined six fundamental stages: Need identification, product brokering, merchant brokering, negotiation, purchase and delivery, service and evaluation

At the core of the process model proposed here is a seven-step cycle performed by the buyer. The seven steps shown in Figure 2 are: Analysis of Needs, Search for Suppliers and Product Catalogs, Product and Supplier Evaluation, Negotiation, Contract Signing, Settlement, and Consumption.

The process model also lists other tasks that are performed by buyers and merchants in no particular order. Tasks shown with a dark background are performed by the merchant (e.g. Consulting), and tasks with a white background are performed by the buyer. The light-gray shaded area contains common tasks with close interactions between buyer and supplier.

Furthermore, tasks that are listed outside of the marketplace border are performed individually without buyer-merchant interaction. For example, merchants manage their product information for the electronic catalog “off-market”, without direct interaction with customers. Contrary to this, “Search for suppliers and product catalogs” is a task that is performed by the buyer, but with support from merchants who supply the necessary information. Customers may analyze their needs on their own or may be supported by the supplier.

Process Model of Electronic CommerceFigure 2: Model of Electronic Commerce: Processes

Buyer Support

During the workshop we discussed findings, support mechanisms, and examples of systems that are relevant to supporting the various shopping tasks.

Andrew Howes and Jinwoo Kim presented their respective results from studies in which they investigated how product finding can be made more effective. Howes studied product finding performance in virtual-reality-based shops and compared this with shops having a textual interface for navigation. He found that it takes more time for new visitors to locate items in the virtual-reality-based shop, but that visitors also have slightly better recall of products offered in this type of shop.

Kim compared the effectiveness of strict hierarchical index link structures and “space metaphor” link structures, which use words like “up”, “down”, and “neighbor” to suggest an analogy between the navigation in the information structure and physical navigation. In his study of buyers who had to find multiple items from a shopping mall he found that spatial metaphoric aids were perceived as being easier to navigate than other types of links.

Martin Rosewitz and Fernanda Viegas presented two cataloging tools that support product evaluation. Buyers can express their preferences for certain product features such as price or screen size when shopping for TVs. The tools support buyers by determining which of a number of products best matches the stated preferences. Rosewitz explained how fuzzy logic mechanisms can be used to deduce well-matched preference functions for product evaluation from simple preferences stated by the buyer. Viegas focused on supplying buyers with visualizations to help them understand product properties and to rank products according to stated preferences.

Michael Good presented an SAP-based system geared specifically towards habitual buyers in business-to-business settings who need to buy operating resources such as office materials. The system assumes a simplified shopping process that consists only of Product Finding and Settlement. The challenge in designing such a system was to make it easy enough for users to create valid purchase requests within minutes without the help of additional documentation.

Tom Gross presented a system for collaborative browsing that can be used for addressing social needs of buyers. The system supports same time-different place collaboration between buyers. One buyer takes control and “steers” the web navigation, the other person “follows” and his or her browser displays the same page as that of the person steering.

Stanley Przybylinski also looked at collaboration issues within a concurrent design environment, and Govert de Vries discussed collaborative and community issues in online games.

Other interesting systems for buyer support mentioned during the workshop include

PersonaLogic (http://www.personaLogic.com), a tool that, similar to the systems presented by Rosewitz and Viegas, supports product evaluation,

Jango (http://www.jango.com), a tool that supports merchant selection and product selection.

Kasbah (kasbah.media.mit.edu), a Web-based multi-agent classified ad system where users create buying and selling agents to help transact goods.

Dell computer configuration catalog (http://www.dell.com) which supports product configuration and product selection, and, Electronic catalog software from vendors such as Aspect Development (http://www.aspectdv.com) and Saqqara Systems (http://www.saqqara.com).

Open Issues for Buyer Support

Patrick Steiger took examples from the insurance domain to remind participants of the importance of needs analysis, and that it should be performed independently from the discussion of particular product features. For example, stating preferences for particular travel insurance coverage only makes sense once the general need for travel insurance has been identified.

An issue that attracted repeated discussion was the question of how to deal with complex products. Peter Boersma explained that complex products not only have attributes, but also rules about matching products, compatibility constraints, and versions.

Markus Stolze added that, from a user perspective, there are two things that make products more complex: configurability and fine print. Configurable products are more complex to handle because what looks like a single product is actually a whole family of possible product configurations that buyers have to select. None of today’s cataloging tools offer support for buyers who want to compare configurable product from multiple providers. Fine print is another complexity dimension that is not addressed by today’s cataloging tools: for example, comparable descriptions of warranty services for a computer will require product descriptions that go beyond simple lists of attribute-value pairs.

Another issue that received much attention was how to create e-business systems that satisfy not only the rational needs of buyers (i.e. needs analysis and product comparison) but also the emotional and social needs of buyers. Antoinette Littel contributed to this discussion by giving examples of individual and social needs and how some of them are currently addressed by pioneering e-shops (cf. Tables 1 and 2).

Merchant Support

A number of commercial systems are available that support merchants performing store setup and store maintenance tasks. Examples of such systems are IBM Net.Commerce, Microsoft Site Server Commerce Edition, and Pandesic.

Two areas where participants felt more work needs to be done are customer retention and support for building multi-merchant marketplaces. Mechanisms for customer retention help merchants keep customers in their shop and encourage them to return.

Eric Arthur reported on a new set of add-on tools to the Microsoft Site Server Commerce Edition tool suite. These tools support merchants by providing customer profile information to define rules for conditional advertising display and targeted mailings. In addition, participants felt that services such as newsletters, chat services, and coupons should be investigated in more detail.

Martin Bichler argued that current shop-building tools do not address the needs of brokers who need support in setting up multi-merchant marketplaces, that, for example, offer auctioning as the market policy. He explained how their CORBA-based framework for electronic auctioning addresses this problem.

Table 1: Examples of Individual Shopping Needs as presented by A. Littel
Individual Needs Commonness Example Website Description
Learning about new things **** http://www.klwines.com Newsletter
Bargain-hunting * http://www.dakinfarm.com
http://www.flightbrokers.nl
Find the pig contest (50% off)
Airline tickets
Getting physical exercise
Relax, kill time * http://www.landsend.com Read stories
Stimulate your senses (hear, smell, touch) * http://www.thunderboltrecords.com Hear CD tracks
Avoid depression, loneliness * http://www2.garden.com 24 hour chat
Not being seen by friends and family ************* Provided by most websites Isolated sessions
Table 2: Examples of Social Shopping Needs as presented by A. Littel
Social Needs Commonness Example Website Description
Enjoyment of negotiation
Communication with people having the same interest * http://www.barnesandnoble.com Live chat with authors
Meeting with friends and family
Watching people: their clothes, their behavior
Enjoying status by giving orders and being served
Doing things with friends

Robert Guttman and Alex Kleiner, in their contribution, argued that auctioning is too limited a mechanism for negotiation as it focuses on the negotiation of price. Hence auctions are a zero-sum game. They pointed out the need for more flexible negotiation schemes to achieve a win-win situation. This is a way to keep medium-sized merchants economically viable, too. For this, negotiations have to accommodate merchant and third-party-supplied information about product features, add-on services, and supplier features.

Other components of marketplaces identified during the workshop include the mechanisms for intelligent information integration for distributed catalogs, as studied for example in Stanford’s work on the CommerceNet initiative (http://www.commerce.net and http://logic.stanford.edu/cit/commercenet.html).

Discussion participants agreed that standards are of central importance, particularly for merchants and merchant support systems. Various standard efforts are underway. These standards cover such topics as XML Document Type Definitions and the Open Trading Protocol (http://www.otp.org). A review of all relevant standards is beyond the scope of this report. Various ones are described in [1]. A good place for more information about political issues of e-commerce is http://www.ecommerce.gov.

easyBusiness Website and Other Results

One result from the workshop was that the authors set up a seed website on e-commerce that focuses on design knowledge, systems, standards, components, and tools to support easy business over the Internet. See http://www.zurich.ibm.com/~mrs/easyBusiness/.

The results of the workshop have been presented and discussed in a CHI `98 Informal SIG, and a mailing list has been assembled to facilitate communication among interested people.

Workshop Participants

Erik Arthur, Microsoft, Redmond, WA, USA
Peter Boersma, General Design, Amsterdam, NL
Eddy Boeve, General Design, Amsterdam, NL
Martin Bichler, Fisher Center/ Haas School of Business, University of California, Berkeley, CA, USA
Govert de Vries, Kodak Imagination Works, Palo Alto, CA, USA
Michael Good, SAP Labs, Palo Alto, CA, USA
Tom Gross, University of Linz, Austria
Robert Guttman, MIT Media Laboratory, Cambridge, MA, USA
Andrew Howes, Cardiff University of Wales, UK
Jinwoo Kim, Yonsei University, Seoul, Korea
Alex Kleiner, MIT Sloan School of Management, Cambridge, MA, USA
Antoinette Littel, Royal PTT Nederland NV (KPN) Research, NL
Gareth Miles, Cardiff University of Wales, UK
Stanley Przybylinski, Industrial Technology Institute, Ann Arbor, MI, USA
Martin Rosewitz, FORWISS (Bavarian Research Center for Knowledge-Based Systems), Erlangen, Germany
Patrick Steiger, IBM Research, Zurich, Switzerland
Markus Stolze, IBM Research, Zurich, Switzerland
Fernanda Viegas, MIT Media Lab., Cambridge, MA, USA

References

[1] Guttman, R. H., A. G. Moukas and P. Maes (1998). Agent-mediated electronic commerce: a survey. Knowledge Engineering Review, June 1998 (in press).

[2] Pananet’s Cyber Stats, Insights, and “Insites”. Available from http://www.panasites.com/pananet/insites.

[3] Schmid, B. F. (1997). Requirements for electronic markets architecture. Electronic Markets 7(1): 3-6.

[4] Target Marketing, July 1997, pg. 30. Available from http://www.commerce.net/work/pilot/nielsen_96/

About the Authors

Patrick Steiger has been member of the e-business solutions group at the IBM Zurich Research Laboratory since 1995. He is currently a Ph.D. student at the University of Zurich, Switzerland. His research is focused on increasing consumer sovereignty on the Internet by providing tools to make their needs explicit. He is working on a prototype for personal risk management.

Markus Stolze is a research scientist at the IBM Research Laboratory in Zurich, Switzerland. His research combines artificial intelligence technology with human-computer interaction concerns and is aimed at the design of intelligent interactive systems for network computing. Stolze received a Ph.D. in Computer Science from the University of Zurich for his research on methods for fitting knowledge-based systems to workplace environments.

Michael Good is a Senior Developer in the Advanced Technology Group at SAP Labs, Inc. He was the principal developer of SAP’s Employee Self Service purchasing product, which provides companies with an Intranet substitute for paper purchase requisitions integrated with electronic catalogs.

Authors’ Addresses

Patrick Steiger, Markus Stolze
IBM Research Division
Zurich Research Laboratory
Säumerstrasse 4
CH-8803 Rüschlikon, Switzerland

email: {pst, mrs}@zurich.ibm.com

Michael Good
SAP Labs, Inc.
3475 Deer Creek Road
Palo Alto, CA 94304-1316 USA

email: michael.good@sap-ag.de