Principal, Xamax Consultancy Pty Ltd, Canberra
Visiting Fellow, Department of Computer Science, Australian National University
Version of 17 December 1993
© Xamax Consultancy Pty Ltd, 1993
This document is at http://www.anu.edu.au/people/Roger.Clarke/EC/FAST.html
There are many different ways in which information technology is being applied to support the trading of goods and services. Can a single case study provide much insight into such a broad field?
In this instance, it can. The reasons are firstly that the features of the system spotlighted in this paper were designed to closely match the needs of a specific marketplace, and secondly that the designers conformed with principles of business process engineering which were only then in the process of being formulated by the leaders of management thought.
FAST's staff describe it as a procurement service (Neches 1993). It has been developed progressively between 1987 and 1993 by the Information Sciences Institute (ISI) of the University of Southern California, and funded by the Advanced Research Projects Agency (ARPA) within the Department of Defense. This paper is based on existing written materials, supplemented by a long day's visit to and examination of the system in December 1993.
The project's origins are an important factor in understanding its orientation. It was an outgrowth of a major and successful ISI program called MOSIS. This is a production service for custom and semi-custom VLSI chips. Since 1981 it has processed thousands of designs, for research organisations, defense contractors and corporations developing products for industrial and commercial markets. Design files are sent to ISI via magnetic media or email. The service's 'value-added' includes ISI's up-to-date knowledge of the fast-changing VLSI technologies and marketplaces, and quality assurance throughout the production process. Its primary competitive advantage lies in small runs, in support of both research and product development. Despite its success, MOSIS remains pre-eminent, if not unique.
MOSIS had resulted in ISI developing a supplier-client relationship with a wide range of research organisations which were heavily dependent on electronic componentry. FAST was conceived as a further service to these clients. The objective was to provide them with a service which relieved each of them of the burdensome task of keeping track of the many suppliers throughout North America, in order to procure occasional, relatively small orders. FAST accepts requests for quotation from its clients, seeks quotations from suppliers, provides the information to its clients. Further, it will manage the purchasing, logistics and settlement activities. The system is described in greater detail later in the paper.
During the last few years, a new approach has been emerging in the management literature. Its proponents claim that in the face of customer demands, competition and unrelenting change, gradualism is not enough, and that instead dramatic change in business orthodoxy is needed. The most important reference, Hammer & Champy (1993), contends that the dominance of organisational structure over organisational process must be reversed. "Many tasks that employees performed had nothing to do with meeting customers' needs ... Many tasks were done simply to satisfy internal demands of the company's own organisation" (p.4). "It is no longer necessary or desirable for companies to organise their work around Adam Smith's division of labor. Task-oriented jobs in today's world ... are obsolete. Instead, companies must organise work around process" (p.28, emphasis in original). By 'process', the authors mean "a collection of activities that takes one or more kinds of input and creates an output that is of value to the customer" (p.35). "Manufacturing ... is better called the procurement-to-shipment process; product development, concept-to-prototype; sales, prospect-to-order; order fulfilment, order-to-payment; and service, inquiry-to-resolution" (p.118).
Hammer & Champy identify a number of common features which are frequently encountered in reengineering (pp.50-64; see Exhibit 1). Of particular importance is the notion that 'processes have multiple versions', by which they mean that 'straightforward' cases are automated, 'medium hard' cases are performed by a case worker with significant computer support, and 'difficult' cases are performed by a case worker with help from specialist advisers as well (p.55). Progressively, it is contended that functional integration will extend all the way through the applications of successive organisations (p.197).
There has been much discussion in recent years as to whether information technology is resulting in a tendency towards flatter organisational structures and greater employee empowerment. Information technology is perceived by Hammer & Champy as "the agent that [enables] ... companies to break their old rules and create new process models ... an enabler ..." (p.47). Their claim is that flat structures and dispersed decision-making are essential outcomes of business process reengineering (pp.77, 189).
These ideas are, of course, not as entirely revolutionary as the tone of Hammer & Champy's 'pot-boiler' management text implies. See, for example, (strategic information systems theory), (activity-based costing), (outsourcing), and multi-faceted integration of the operations of manufacturers and their suppliers (Oesterle 1991). What the business process reengineering notion is doing, however, is to place the need and possibility for change firmly on the agenda.
Hammer & Champy depict business process reengineering as a process undertaken 'top-down' by the senior executives of a corporation (p.168), and involving a 'case for action' and a 'vision statement' (p.148-158). This is of course consistent with the notion of 'executive-as-leader', and planned and rational pursuit by corporations of their own best interests. But there are additional ways in which reengineering can occur, especially when it necessarily involves rationalisation of activities at a level higher than a single corporation, i.e. along an industry value-chain (Clarke 1991).
USC/ISI, in conceiving of FAST, recognised the opportunity to provide not only a means of electronic communication, but also a set of value-added services which would be attractive to particular kinds of organisations. In respect of particular classes of goods, FAST's clients outsource most of the procurement process, accepting that an additional intermediary is capable of performing the tasks both more effectively and more economically than they could achieve in-house. Hence FAST is characterised by its founders and operators as a 'broker' rather than as a value-added network service. The project's aims have been expressed as "demonstrating electronic commerce through the development of a prototype automated broker for standard parts and components. The major goal is to develop technologies that will improve procurement of items within the Department of Defense.
The architecture which underpins the broking operation comprises several elements:
The applications can be analysed into three components (see Exhibit 1):
During its early years of operation, all clients have had access to the Internet and have used standard ASCII text-editors; hence email using structured templates provided by FAST have provided a convenient and reliable basis for communications. X12-based EDI linkages with clients are in development.
The FAST database does not contain a list of suppliers' stock-holdings. What is does maintain is a model of suppliers and the classes of goods they supply (their 'linecards'), together with information about how to communicate with them. Three alternative linkages are currently supported: formatted email, hard-copy fax, and simulated on-line enquiry. The last involves access to the increasingly prevalent in-plant, on-line enquiry services offered by suppliers. Rather than using a human operator's time, however, the FAST team has invested heavily in software which emulates human operations at a workstation, and draws data down from these services into the FAST application. In addition, X12 linkages are being developed.
The basic sequence of activities in FAST operations is depicted in Exhibit 2. A request for quotation is initiated by a client. FAST parses the message, selects appropriate suppliers, and forwards it to those suppliers. It accepts quotations back from suppliers, analyses them, and passes them back to the client. These are non-binding, inventory availability responses, not contractual offers. The client selects its preferred quotation, issues an order to FAST, and FAST forwards the order to the supplier. The supplier issues an order acceptance (which is the point of contract formation), and FAST forwards it to the client. The supplier arranges delivery to the client. The client provides a receiving report to FAST, and FAST acts as agent for the client during the settlement process.
Substantial enhancements to this basic flow have been implemented. In particular:
In late 1993, FAST had about 130 clients, of whom a majority were University research centres, but large minorities were government agencies and corporations. The number of suppliers who were registered on the FAST database was about 2,400, of which more than half were small companies. During the first six months of 1993, FAST received 4,200 line-item requests, returned 6,000 quotes, and ordered 2,700 items for a total value of $1.3 million. The cumulative figures over the six years of operation are in excess of 30,000 quotes, 13,000 line-items and $5 million, the majority since the beginning of 1991.
Automated, semi-automated and largely manual modes
FAST is not appropriate to circumstances in which customisation is necessary, in either the goods themselves or in services associated with them. Nor is it appropriate to large-volume, or mission-critical goods, which necessarily involve direct contact between the supplier and consumer, negotiation on conditions, and specially prepared contracts. It is, however, highly advantageous for small-volume purchases of goods which are, or are similar in nature to, commodities, and which are available through a variety of channels. Its initial application has been to electronic parts, test equipment, optical components and supplies.
The efficiency of the mechanism is attested to by the large number of orders [how many?!] being processed each quarter [to smooth out the considerable variation in volumes] by three purchasing staff. Of these orders, [??% are being processed fully automatically]. FAST applies a small service charge to cover operations costs (currently 8% of order-value). [But does this cover the costs of the operations, i.e. 3 purchasing staff plus the Operations Manager?]. The development and maintenance of the system has been, and at this stage continues to be funded by a succession of ARPA grants.
Implementation and increasing use of ASC X12 standard documents as the basis for communications with both clients and suppliers.
A form of active agent, referred to as scenarios/agendas, are being progressively implemented. These are semi-automated procedures which analyse database state in order to detect and deal with or report exceptional conditions. The first application is the conformation of orders with those (many) suppliers which are not presently capable of issuing EDI confirmations.
The question arises as to the comprehensiveness of the list of suppliers who are recorded on FAST's database, and of those who are approached for a quotation. Although this has not presented any difficulties during the early years of operation, equity of access for small and remote suppliers may become of greater significance in the future.
Consistently with the USC/ISI role as an R&D organisation, opportunities are being sought for commercialisation of the ideas and the technology. Current possibilities include the tighter integration of FAST into the procurement environment available to purchasers at a major USAF base, a pilot project with a value-added network and a major financial institution, and a pilot with an information database vendor.
FAST's current design contains only limited security features. It is essential that more detailed consideration be given to the authentication of messages and their originators through such mechanisms as digital signatures and encryption, and non-repudiation of contractually binding messages.
The origins of FAST lie in customer service. The architecture does not, however, preclude activities oriented towards the supplier rather than the customer community. It is feasible for the existing infrastructure to support a significantly greater volume of business, and hence it may be possible, despite the small size of the team, to pursue multiple new business opportunities at the same time.
Anna-Lena Neches, Operations Manager of FAST, and her colleagues Bob Neches, Paul Postel, Craig Milo Rogers and Robert Wormuth, provided a great deal of assistance to me in my assessment of the FAST product, the process whereby it came into existence, and its prospects.
FAST (1993) 'FAST Electronic Broker: How To Interact With FAST' USC/ISI (March 1993)
Hammer M. & Champy J. (1993) 'Reengineering the Corporation: A Manifesto for Business Revolution' Harper Business, New York, 1993
Neches A.-L. (1992) 'FAS Electronic Broker' Presentation to Citicorp (October 1992)
Neches A.-L. (1993a) 'Government Application of FAST Technology, Tasks 1, 3 & 4: Semiannual Technical Report' USC/ISI (October 1993)
Neches A.-L. (1993b) 'FAST - A Research Project in Electronic Commerce' Electronic Markets, Issue 9-10, St Gallner Technische Hochschule, Switzerland (October 1993)
FAST's primary clientele in recent times have been their original target market, viz. electronics component purchasers among University and corporate R&D laboratories, together with various USAF bases, which have been using FAST since about 1987.
Recently, a review within the Air Force resulted in an internal initiative called MADES II being selected as the exclusive path for electronic commerce in the future. External schemes, primarily GATEC and FAST, are to be phased out from mid-1994.
Meanwhile, two other potential application areas are in development. One is the Citibank-EITech joint venture, which is a scheme intended to build an overlay of commercial features over the Internet.
The other is a 4-month pilot with the Defense Logistics Agency (DLA), intended to prove FAST's scalability to large volumes. DLA is a very large-scale operation, handling orders for about 1 million stock-items for many different organisational units (within DoD rather than the individual forces?), and having about 60,000 back-orders on hand at any one time.
Commencing in April, DLA will direct a small volume of live purchase orders under $2,500 through FAST. The focus at this level is on small business suppliers (which means $0.5-5 million dollar turnover, 100 employee companies). At this level there is no requirement that multiple quotations be received, and it is likely that FAST will in any case deliver more rather than fewer quotations than is presently the case.
The volume is planned to commence at 50 line-items per day during the first week, and be increased to 250 or higher, depending on the capacity of the FAST infrastructure and organisation to cope. The key to this is more likely to be the limited number of trained purchasing staff available to cope with the semi-automated and largely manual exceptions, rather than the network, processor and software capacity. If the pilot is successful, the intention is not to move to a wider implementation of FAST, but rather to transfer technology to the DLA.
A considerable expansion in the number of small vendors is necessary to support the DLA pilot (in such areas as rope, wire, separators - in Australian, washers, etc.). The target is to have at least 30 vendors in each Federal Supply Code. The shortfall is being addressed by consulting the U.S. government's PASS Small Business Register, and by seeking details from VANs (with limited success, because they have insufficient knowledge about the product-classes their customers supply).
Only one additional small business has been located which has an in-plant system, and only two which are EDI users and prepared to interface in that manner. It is therefore likely that, at least at the beginning of the pilot, a great deal of the volume will involve faxes from FAST to the vendors. The value which FAST can add is therefore primarily its database and associated know-how of the vendor market, plus the scope to upgrade the linkage beyond fax to structured email and EDI.
One additional challenge which has emerged is the handling of taxes. Although federal government agencies are exempt, there are States in which the vendor is not. Moreover, the schemes differ according to product classes, the location of the sale, the delivery point and/or the dispatch point, and whether the tax is calculated on purchase, sale or imputed sale value.
For FAST's Mosaic server, use http://broker.isi.edu.
CommerceNet is also operating a MOSAIC server, which includes information on FAST. It also provides formatted input-screens for FAST, which represent an improvement on the email templates used by many existing customers.
The contact at CommerceNet/EITech is Marty Tenenbaum, firstname.lastname@example.org in Palo Alto.
The only VAN with which FAST currently has a connection is GEIS, although this provides gateway access to two others, Harbinger and OrderNet, and could provide access to others as and when necessary.
Go to Roger's Home Page.
Go to the contents-page for this segment.
Send an email to Roger
Created: 22 November 1997
Last Amended: 22 November 1997
|These community service pages are a joint offering of the Australian National University (which provides the infrastructure), and Roger Clarke (who provides the content).|
| The Australian National University|
Visiting Fellow, Faculty of
Engineering and Information Technology,
Information Sciences Building Room 211
Consultancy Pty Ltd, ACN: 002 360 456|
78 Sidaway St
Chapman ACT 2611 AUSTRALIA
Tel: +61 6 288 6916 Fax: +61 6 288 1472