A Comparison of Two Electronic Auction System Initiatives in the Dutch Flower Auctions

Eric van Heck and Pieter M. Ribbers

Dr.Ir. Eric van Heck
Erasmus University Rotterdam
Rotterdam School of Management
Department of Decision & Information Sciences
P.O. Box 1738
3000 DR Rotterdam
The Netherlands
Telephone: +31-10-4082032
Telefax: +31-10-4523595
Email: e.heck@fac.fbk.eur.nl

Prof.Dr. Pieter M. Ribbers,
Tilburg University
Department of Information Systems and Auditing, and
Center for Economic Research (CentER)
P.O. Box 90153
5000 LE Tilburg
The Netherlands
Telephone: +31-13-4662188
Telefax: +31-13-4663377
Email: p.m.a.ribbers@kub.nl>

WORKSH1.doc, July 25, 1997, ReSubmitted to Wirtschaftsinformatik

Abstract

This article discusses an unsuccessful and a successful introduction of an electronic auction system in the Dutch flower industry. The unsuccessful introduction of an electronic auction system was the development and implementation of the Sample Based Auction (SBA) by Flower Auction Aalsmeer (FAA). Growers send a sample of the product. Growers, buyers, and auctions use electronic data interchange (EDI) to share all the information required in the transaction process. The successful introduction of an electronic auction system was the entrance of Tele Flower Auction (TFA) into the Dutch flower industry, enabled by Information Technology (IT). Indeed, the development and introduction of TFA is one of the initiatives in response to import restrictions by the traditional Dutch flower auctions. TFA is an electronic alternative that enables buyers to trade at a distance; this alternative is currently exploited by an import organization called East African Flowers (EAF). This article aims to provide a better understanding of the successes and failures of electronic auction systems. It provides a descriptive framework for analyzing the merits of electronic auctions. It uses that framework to evaluate the SBA case and the TFA case. The results of the analysis and the framework itself illustrate the various complex issues that arise in the design and implementation of electronic auction systems.

Keywords

electronic auction system, electronic market, flower auction, flower industry, success factors

1. Introduction

Since 1991 there has been an ongoing debate in the Dutch flower industry, about: (i) the decoupling of price discovery and logistical processes in the Dutch flower auctions; (ii) the increasing imports of foreign flower products at the Dutch flower auctions; and (iii) the use of new Information Technology (IT) in this industry.

The rapid growth of the flower industry created increasing demands for logistic support, space in the auctions and complaints about nearby traffic congestion. Flower Auction Aalsmeer (FAA) was close to their limits in terms of complexity, capacity and room to expand. The answer to these limits was the introduction by FAA in 1994 of a Sample Based Auction (SBA) for potted plants.

After a referendum in September 1994 the growers, who are the owners of the auctions, decided to ban foreign grower participation in the auctions during the summer. These efforts to reduce foreign access to the traditional Dutch auctions, led buyer organizations and foreign growers to announce the creation of competing auctions. Indeed, the development and introduction of Tele Flower Auction (TFA) is one of the initiatives in response to import restrictions by the traditional Dutch flower aucti ons. TFA is an electronic alternative that enables buyers to trade at a distance; this alternative is currently exploited by an import organization called East African Flowers (EAF).

This article aims

• to describe the development, implementation, and effects of two electronic auction system initiatives in the Dutch flower industry, e.g. the Sample Based Auction (SBA), and the Tele Flower Auction (TFA);
• to explain through an indepth analysis a better understanding of the reasons for the failure of the SBA, and the success of the TFA.

This article further shows that IT enables new ways of competition and coordination, thus changing the ways in which individuals and organizations exchange goods and services.

The article is organized as follows. In section 2 the descriptive framework is presented. This framework will be used to describe the Sample Based Auction case, and the Tele Flower Auction case. In section 3 some characteristics of the Dutch flower industry are presented. In section 4 the concept and development of the Sample Based Auction, as a reaction to consumer orientation and decoupling, is described with the help of the framework of section 2. In section 5 the concept and development of the Tele Flower Auction, as a reaction to import restrictions, is described with the help of the framework of section 2. In section 6 both cases are analyzed, and the lessons learned are presented. Conclusions are formulated in section 7.

2. Literature Review and Descriptive Framework

There are many theoretical and empirical studies of auctions [Cram95;DaHo93;HePo88; Milg89;RoHa94], and electronic markets [Bako91; ClRe93; ClWe90; GuWh91; HeKe94; KaVh96;KoWa89;Lee96;MaYa87;Schm93;VhRi96;VhVd97].

2.1 Auctions

Milgrom discusses the characteristics of auction theory [Milg89]. Hendricks and Porter examines federal auctions for drainage leases and find that their data suggest that neighbor firms are better informed about the value of a lease than non-neighbor firms, that neighbor firms coordinate their bidding decisions, and that both types of firms bid strategically in accordance with the Bayesian-Nash equilibrium [HePo88]. Cramton described the full history of bidding dealing with the successful Nationwide Narrowband PCS Auction [Cram95]. Rothkopf and Harstad critically analyzed the gaps between the existing theory and the reality of auctions [RoHa94]. For theorists, Rothkopf and Harstad advise attention to the particulars of how auctions are modeled. They argue that modelling improvements will be of more direct value than use of equilibrium concepts more subtle than Nash's or new theorems that apply only to oversimplified models. Research presented in this article follows their advice.

2.2 Electronic Markets

Electronic markets are defined as interorganizational information systems that allow the participating buyers and sellers to exchange information about prices and product offerings [Bako91]. Research on the effects of IT on exchange organizations and processes is relatively new. Early research applied transaction costs and agency theory to predict shifts from hierarchies to market form of organizations [MaYa87;GuWh91]. The central argument of this line of research was that IT would improve communication, search, monitoring and information sorting capabilities, thereby reducing transaction costs and enabling purchasers to take advantage of production economies available in markets. A critical drawback inherent in this analysis was the definition and treatment of markets in abstract economic terms (i.e., markets coordinate economic activity through a price mechanism). In reality, different market structures exist, e.g. direct search, brokered, dealer and auction markets. Each of these structures organizes the trading process and related information processing activities in different ways. Thus the role and impact of IT can vary across types. The literature provides some examples. Konsynski et al. provided a descriptive case study of an electronic market in used cars [KoWa89]. Clemons and Weber examined the effects of computerization on the London Stock Ex change [ClWe90]. Clemons, Reddi, and Row examined the impact of IT on the organization of production; they presented the 'Move to the Middle' hypothesis [ClRe93]. Hess and Kemerer tested the Electronic Market Hypothesis (EMH) against the empirical results of five case studies in the home mortgage market [HeKe94]. Kambil and Van Heck showed the role and impact of IT on the Dutch flower auction markets [KaVh96]. Lee examined two types of electronic markets: electronic brokerage, and electronic auction [Lee96]. Van Heck et al. discussed the successful entrance of Tele Flower Auction (TFA) into the Dutch flower industry [VhVd97].

Due to the convergence of IT and telecommunication, and the proliferation and availability of bandwidth, the impact of electronic markets is expected to grow rapidly. Their effectiveness, however, is dependent on their design. Existing research in this new area provides examples of relevant issues supporting an effective design. What is lacking, however, is a systematic classification of various complex economic issues that arise when designing and implementing electronic markets. This article proposes, through an indepth analysis of two cases, relevant elements for describing electronic auction systems.

2.3 Descriptive Framework

Based on the analysis of these studies, we develop a framework to describe electronic auction systems and apply this framework to formulate relevant lessons with regard to successes or failures of electronic auction systems. This descriptive framework consists of eight elements, which will be discussed next; see Figure 1 (adopted from [VhVd97]). The framework is derived from an analysis of auction theory and electronic markets theory.

Figure 1: Descriptive Framework

1. Buyers

Buyers can be characterized by their decision-making behaviors (bidding strategies). In a Dutch flower auction there are many buyers bidding on each object; for example, in the TFA case there are 160 buyers.

2. Objects

Auctioned objects deal with the characterization of the exchanged products in terms of product characteristics, quality, length of stems, number of units, and stems per unit.

3. Sellers

Sellers can be characterized by their decision-making behaviors (selling strategies). In a Dutch flower auction each object is sold by a single seller; usually the seller is the producer of the object.

4. IT Innovations

New IT innovations may be adopted and used by sellers and buyers, in different ways. We shall pay special interest to the way these innovations influence trust among sellers and buyers. We shall also study which (new) information can be stored, sent, and retrieved by IT.

5. Auction Rules

Auction rules deal with statements of what can, should or must be done in certain auction circum stances. For example, specific auction rules may concern quality control. There are also auction rules concerning the role of the auctioneer.

6. Trust

Trust deals with the belief, or willingness to believe, that one can rely on the goodness, strength, and ability of somebody (the seller or the buyer) or something (for example, IT innovations). Trust is the expectation that arises within a community of regular, honest, and cooperative behavior, based on commonly shared norms, on the part of other members of that community [Fuku95]. Fukuyama argues that [Fuku95, 27]: 'people who do not trust one another will end up cooperating only under a system of formal rules and regulations, which have to be negotiated, agreed to, litigated, and enforced, sometimes by coercive means. This legal apparatus, serving as a substitute for trust, entails what economists call "transaction costs"'.

7. Auction Results

Results or effects of the auction can be distinguished into results for (i) sellers, (ii) buyers, and (iii) intermediaries (auction house) [VhRi96]. Examples are prices of auctioned objects, transaction volumes, transaction speed, price volatility, and feasibility.

8. Competition with other markets

The literature shows that most research on auctions has a single isolated auction point of view [RoHa94]. In Dutch flower auctions this viewpoint is too narrow. Dutch flower auctions are market places in an international flower market. There are other market channels between sellers and buy ers, such as the mediation offices. Therefore we introduce an extra element, namely competition with other markets (auctions, and market channels). We shall not primarily focus on this element, but shall consider it to be an important part of the environment.

In section 4 and 5 we shall study two practical cases, namely the SBA case, and the TFA case, to explore relevant lessons dealing with the success or failure of electronic auction systems. The nature of this type of research is explorative. Such a case can be used for so-called analytical generalization, not statistical generalization [Yin89]. We use the case study method, because it enables 'reality' to be captured in considerably greater detail than other methods, and it also allows the analysis of a con siderably greater number of variables. We held interviews, analyzed relevant reports, and obtained archival data.

3. The Dutch Flower Industry

3.1 Industry Background

The Netherlands is the world's leading producer and distributor of cut flowers. The Dutch dominated the world export market for cut flowers in 1995 with a 59 per cent share. The world's two biggest flower auctions are in Aalsmeer (Flower Auction Aalsmeer) and Naaldwijk/Bleiswijk (Flower Auction Holland); every day on average 30 million flowers - originating not only from the Nether lands but also from countries such as Israel, Kenya and Zimbabwe - are traded in 100,000 transac tions. There are seven Dutch flower auctions, namely in the villages of Aalsmeer, Naaldwijk/ Bleiswijk, Rijnsburg, Grubbenvorst, Eelde, Bemmel, and Vleuten. Their sales of cut flowers in 1995 were 1.44, 1.26, 0.54, 0.72, 0.045, 0.036, 0.027 billion DM (1 billion = 1,000,000,000.00). The Dutch flower auctions play a vital role in Holland's leadership of this industry, by providing efficient centers for price discovery and transactions of flowers between buyers and sellers. These auctions traditionally use the 'Dutch auction' as the mechanism for price discovery. They are established as cooperatives by the Dutch growers.

3.2 The Dutch Flower Auction Concept

SBA and TFA still use the Dutch flower auction as the underlying price discovery concept. Surpris ingly, the economics literature does not pay specific attention to the Dutch flower auction. Davis and Holt [DaHo93], among others, do mention the Dutch flower auction, but do not discuss it in great de tail. Kambil and Van Heck [KaVh96], and Van Heck and Ribbers [VhRi96] describe the function ing of Dutch flower auctions in detail, but not from an economics perspective. In this section we shall describe the auction rules of the Dutch flower auction concept. We illustrate characteristics and results of the Dutch flower auction concept through empirical data of Flower Auction Aalsmeer.

There are approximately 3500 varieties of cut flowers. These varieties are classified into 120 auction groups, according to the variety, size of the lot, and quality of the flowers. Dutch flower auctions use a clock for price discovery, as follows. The computerized auction clock in the room provides the buyers with information on producer, product, unit of currency, quality, and minimum purchase quantity. The flowers are transported through the auction room, and are shown to the buyers. The clock hand starts at a high price determined by the auctioneer, and drops until a buyer stops the clock by pushing a button. The auctioneer asks the buyer by intercom, how many units of the lot he or she will buy. The buyer provides the number of units. The clock is then reset, and the process begins for the left-over flowers, sometimes introducing a new minimum purchase quantity, until all units of the lot are sold. Table 1 illustrates the auction process by an example with empirical auction data. The first rows deal with producer 1234 (column 2), who is responsible for transactions 408 to 420 (column 1). On January 4, 1996 this producer delivered roses (product group 52), or more specifically the brown rose 'Leonidas' (product number 10288). He delivered four lots of that type of rose (column 4). These lots had the same type of quality (A1), but were different in length (70, 60, 50, and 80 centimeters respectively) and in amounts of 9, 5, 3, and 12 units respectively. The first lot was auctioned, and buyer 3782 took 1 unit (out of 9) for a price of 93 cents per stem. The rest of the lot was auctioned again, and buyer 1854 bought 2 units for 95 cents. The remainder of the lot (6 units) was auctioned, and buyer 727 bought 3 units for 96 cents. Finally, the rest of the lot was bought by buyer 42 for 97 cents. The table shows that the price may increase during the auctioning of a lot (see, for example, transactions 408 through 411) or may decrease within a lot (see, for exam ple, transactions 729 through 731). So the price is very volatile, considering different lots of the same producer or different lots of different producers.

Table 1: Auction Data Illustrating the Dutch Flower Auction Concept (Source: VBA, 1996).

Buyers must be physically present in the auction room. In practice, it turns out that the Dutch flower auction is an extremely efficient auction mechanism: it can handle a transaction every four seconds; for example, Flower Auction Aalsmeer handles via 13 clocks 30,000 transactions a day. It also reduces the amount of time that growers must spend on price discovery and bidding; hence they can focus on production. The auction provides a central location for the meeting of buyers, creating effi cient quality control and logistics of product redistribution. This auction has "backtracking" possibili ties: though the price movements are decreasing per sub-lot, the price can be multidirectional (up or down) within the whole lot. Buyers can withdraw their willingness to buy: they can indicate to the auc tioneer fewer or more units then they originally intended to at the time they pushed the button. During the auctioning of the lot, buyers produce information on the value of the lot; this information is available to all buyers. Given these characteristics, we call the Dutch flower auction a multi-unit, multiform price Dutch auction.

4. Sample Based Auction (SBA) Case Description

4.1 The SBA Concept

Flower Auction Aalsmeer began a sample-based auction for trading potted plants in 1994. In this concept, growers send a sample of the product to the auction house along with information on available inventory. During the auction the sample represents the entire inventory available to buyers who can bid for the product and specify product packaging and delivery requirements. Growers then package the product as specified and deliver it the next day to the buyer location in the auction complex or to other buyer warehouses. Buyers have to be physical in an auction room. Growers, buyers, and the auction used electronic data interchange (EDI) to share all information required in this process. This trading model reduces the number of times a product is handled, reducing overall packaging costs and damage.

4.2 Description of the SBA Case

1. Buyers

Exporters, wholesalers and retailers are the main buyers at the traditional Dutch auctions and SBA.

2. Objects

SBA auctions potted plants, such as ficus, dracaena, kalanchoe, dendranthema, and begonia.

3. Sellers

SBA started with 30 growers. After the introduction, the number of growers decreased rapidly to around 10 growers. SBA was restricted for growers, who are a member of FAA.

4. IT Innovations

In the SBA concept, an important element was the use of electronic data interchange (EDI) between computers of the growers, FAA, and the buyers. Growers could send their supply data by electronic supply letter to the auction computer. The data were stored in a supply information system. Buyers could use that system in the process of negotiating with their clients. After the auction process, growers received electronically the orders including packaging and delivery practices.

5. Auction Rules

SBA uses the same auction rules as in traditional Dutch flower auctions. The role of the auctioneer is the same. He identifies buyers, and determines the minimum amount of units. By setting this amount, the auctioneer can try to influence the price level.

6. Trust

It soon became clear that one of the main critical factors of SBA was that the quality of the sample, as a representation of the lot, determines the buyers' trust in the SBA concept. The trustworthiness of the sample, as a representation of the lot, was questioned by the buyers. Buyers had the feeling the best potted plants of the lot were used as a sample. Reliable product information and stable quality control are essential. Buyers trust the IT innovations in SBA. One of the reasons seems to be that the Dutch auction clock is still the price discovery mechanism; buyers are used to that mecha nism.

7. Auction Results

The different actors, the growers, the buyers, and the auction, expected a number of different benefits. First, by uncoupling logistics and price determination, the auction and growers expected the number of transactions per hour to increase. In reality the number of transactions per hour decreased as buyers had to specify terms of delivery. Second, while the auction expected 45% of the supply of potted plants to be transacted in the sample-based auction, only 10% of the product was transacted this way. Thus, SBA also did not effectively reduce storage requirements at the auction. After numerous attempts to increase the volume of sample-based auctions they were discontinued in late 1994. The system had a negative effect on the functioning of growers, the auction house and buyers. The sample-based auction system ended up in a complete failure.

8. Competition with other markets

SBA's main competitors were the traditional Dutch auctions of FAA. Growers, and buyers considered the traditional auctions as an alternative for selling and buying potted plants.

5. Tele Flower Auction (TFA) Case Description

5.1. The TFA Concept

An important effect of the import restrictions imposed by the Dutch flower auctions, was the creation of TFA by EAF [Vvli94b]. EAF is one of the biggest importers of cut flowers; they specialize in supply from East Africa (Kenya, Tanzania, and Uganda). EAF was established in 1984; it is located in the Aalsmeer area. For EAF, the effect of the import restrictions was that 30% of their imports could no longer be traded via the Dutch auction clocks during the traditional import season; in the summer season 100% of their imports could not be traded at all. EAF announced the creation of TFA in December 1994. In January 1995 the system was tested. On March 24, 1995 TFA was launched with 70 buyers. In the beginning, TFA was restricted to 15 growers who were the main EAF suppliers. In March 1995 TFA and Flower Auction Aalsmeer agreed to use the same type of carts for transport, and they agreed that TFA could deliver its products to the buyers who had their facilities in the auction hall of Flower Auction Aalsmeer. After some months, EAF decided that grow ers from other countries (for example, Spain, Colombia, France, India, and Israel) were allowed to use TFA. After one year, approximately 150 buyers were connected to TFA. In October 1995, EAF decided that TFA would become a permanent electronic auction market. EAF expected a turnover of 100 million Dutch guilders for the growing season 1995/1996.

In the TFA, buyers can bid via their personal computer (PC) screens [Bos95;Eras95;Vvli94b]. Each PC is connected to a fully computerized auction clock. Logistics and price discovery are uncoupled. Flowers are no longer visible for buyers, and buyers are no longer physical in an auction room. The PC provides the buyer with information on the next flower lots. On his PC the buyer can earmark interesting lots, so at the time those lots will be auctioned, the PC will warn the buyer. The PC provides information on the producer, product, unit of currency, quality, and minimum purchase quantity. For each lot two images are presented on the PC screen. The underlying auction concept remains the same: Dutch flower auction. On the PC screen the buyer sees the Dutch auction clock. The clock hand starts at a high price, and drops until a buyer stops the clock by pushing the space bar at the key board of the PC. The auctioneer asks the buyer, via an open telephone connection, how many flowers of the lot he or she will buy. The buyer provides the amount. The clock is then reset, and the process begins for the next units, until the remainder of the lot is sold.

Growers send the flowers to EAF, and EAF stores these flowers in Amstelveen. Logistics and price discovery are uncoupled within the auction hall. The distribution of the flowers from the Amstelveen area to the buyer's addresses (nearby the traditional auctions of Aalsmeer, Naaldwijk, and Rijnsburg) is done by transporters of EAF. Transport costs are paid by EAF.

5.2. Description of the TFA Case

1. Buyers

TFA started with 70 buyers. In March 1995 there were 125 buyers, mainly wholesalers/exporters, attached to TFA; in May 1995 there were 160 buyers (50% located in the Aalsmeer area, the other 50% distributed over the Netherlands). Buyers pay 430 guilders per month (VAT excluded) to lease the equipment. For each transacted flower stem, buyers have to pay 7 cents per stem.

2. Objects

TFA auctions imported flowers. TFA has strict quality control norms, and provides the standard information on flowers; moreover, TFA also provides positive quality remarks, and images of the lot.

3. Sellers

TFA started with 2 African growers. After the decision to open TFA for others, the number of growers increased rapidly. TFA operates now with 40 growers. Growers from Southern Europe, Africa (Kenia, Uganda, Malawi, and South Africa), Equador, and India use it as their marketing channel. Sellers pay for commission.

4. IT Innovations

Compared with traditional auctions and SBA, buyers can trade at a distance. TFA provides better and more frequently updated supply information. The speed of the TFA system is amazing. Not only the auctioning process, but also the after-sales process is very fast; sometimes within half an hour products are delivered at the buyer's address. The IT architecture consists of several components. Buyers have to lease a Personal Computer (PC), earphone- and microphone-equipment, and a printer from TFA. The PC's are connected via ISDN connections to a workstation at TFA in Amstelveen. The TFA workstation includes technology for multi-modem connectivity, and computerized auction clock technology. The ISDN connections are leased by TFA from the Dutch PTT Telecom. The workstation is connected to an auction PC, which is handled by the auctioneer. At the buyer's PC screen, for each of the lots two digital images are available. One image gives an overview of the lot, especially to inspect the ripeness of the flowers. The other image presents the details of the flower bud, especially to analyze the size of the flower bud, and to diagnose a disease.

5. Auction Rules

TFA uses the same auction rules as in traditional Dutch flower auctions. The role of the auctioneer is the same. He identifies buyers, and determines the minimum amount of units. By setting this amount, the auctioneer can try to influence the price level.

6. Trust

It soon became clear that one of the main propositions of TFA was that the quality of the flowers determines the buyers' trust in the TFA concept. TFA's motto is: 'Buyers have to trust the quality blindfold', because buyers cannot physically see the product anymore. Still, buyers who are nearby TFA, can inspect the imported flowers; 30% of the buyers do so regularly. Reliable product infor mation and stable quality control are essential. Quality control is done by TFA's quality inspectors at the grower's place, at the distribution point in Nairobi (Africa), and at TFA in Amstelveen. Buyers also trust the IT innovations. One of the reasons seems to be that the Dutch auction clock is still the price discovery mechanism; buyers are used to that mechanism.

7. Auction Results

In June 1995, the Chief Executive Officer (CEO) of EAF/TFA, Mr. Simon van der Burg, stated that results of TFA were better than expected [Vdme95]. Buyers were enthusiastic about the quality and the delivery time of the auctioned products, and about the service level of TFA. The prices were on average not higher or lower than in the traditional Dutch flower auctions. Growers were also enthusiastic; EAF decided that growers not related to EAF could use TFA as their selling point. Every day an auction lasts for two hours. Every day, approximately 2 million stems are auctioned [Eras95].

8. Competition with other markets

The Dutch growers perceived TFA positively. In June 1995, 68 % of 378 Dutch growers who were interviewed, were in favor of the import restrictions; 55 % answered that TFA was the best alternative for imported flowers; 26% would like to re-install the liberal import policy [Vvli95]. Another interesting aspect is that 10% of the Dutch growers is thinking about choosing TFA as their marketing channel. In 1995 229 growers (in 1994: 269) ranked foreign production as the most important threat, and 92 (in 1994: 144) growers ranked auctioning imports as such [Vvli95]. Overall, Dutch growers still seem to believe that their position improved, although prices did not increase. The development of TFA had an impact on other auctions. For example, the imports of Flower Auction Aalsmeer decreased with 23 %, whereas the average was 10% for all traditional Dutch auctions [Bos96]. Other auctions can have an impact on TFA. For example, Flower Auction Holland decided to copy the TFA concept; they introduced their own Tele Auction System in 1996. The auctions in Vleuten and Eelde also introduced the Tele Auction principle; approximately five wholesalers/retailers use these two systems.

6. Lessons Learned

What can we learn from these case studies? We think that the SBA case, and the TFA case tell the following important and interesting story.

6.1 Dominant Players' Failure and New Entrants' Success

Based on our analysis of the SBA and TFA case, we propose the following lessons.

Lesson 1: Market dominance by established dominant players is a precursor of elec tronic auction system failure.

Lesson 2: New entrants, facing established dominant players, can quickly build competitive advantage with an innovative electronic auction system concept.

Sample Based Auction (SBA)

The SBA case shows the weaknesses of the established, dominant player Flower Auction Aalsmeer. In the sample-based auction system the trustworthiness of the sample, as a representation of the lot, was questioned by the buyers. Buyers had the feeling the best potted plants of the lot were used as a sample.

The SBA failed to meet expectations for many reasons [KaVh96]. First, the incentives and benefits to buyers and growers (in particular) did not change substantially to encourage their participation in this market. Specifically, growers received no extra compensation for modifying packaging and delivery practices to suit the customer. Second, the growers perceived they got lower prices in a slower auction. To overcome this disadvantage growers would break the same product into different sample lots so that it would be priced multiple times during the auction hoping it would lead to higher prices. Third, the auction rules initially did not provide incentives to buyers by supporting transactions on large lots. Instead, the auction maintained rules to favor transactions in small lots. Thus, an insufficient number of buyers and sellers initially adopted this new form of trading [VhRi96].

Tele Flower Auction (TFA)

The efforts to reduce foreign access to the traditional Dutch auctions, led buyer organizations and foreign growers to announce the creation of competing auctions. Indeed, EAF's development and introduction of TFA is one of the initiatives in response to these import restrictions by the traditional Dutch flower auctions. Surprisingly, from the beginning the CEO of EAF/TFA (Mr. Simon van der Burg) made clear that TFA was for EAF the second best solution. EAF preferred no import re strictions; EAF had to create a new market channel. Therefore, EAF moved forward into the flower chain. In the new situation, EAF imports flowers, and coordinates and organizes the price discovery process for imported flowers. Other options were to set up a traditional auction or to set up a mediation office (trade by telephone). The first option was not chosen due to the high initial costs related to build up a new auction hall. The second option was not chosen because in that option no objective price forming mechanism would be available. Therefore, the development of an electronic auction was chosen. It was the first time in Dutch history that an importer organization performs this function. Traditionally, the Dutch flower auctions are established as cooperatives by the Dutch growers. Another interesting point was the high speed of entrance. The import restrictions were valid in October 1994; At that time EAF developed the first ideas about TFA. TFA started in March 1995. So EAF developed and implemented TFA in a few months.

The entrance of TFA was successful for all stakeholders. On average it resulted in the same prices as similar products made at the other Dutch flower auctions. Buyers are impressed by TFA's service, quality of products, and speed of delivery. One buyer stated that 'their logistical process is excellent; although we are very close to Flower Auction Aalsmeer, TFA delivers much faster than Flower Auction Aalsmeer' [Bos96]. The reason is much simpler internal logistical streams, compared with the traditional Dutch auction. After some months TFA attracted more buyers and non-EFA sellers (from, for example, France and India). Some buyers say they miss the atmosphere of the traditional auction halls, where they could 'feel' the tension in the market. Other buyers prefer working behind the PC screen: they mentioned better concentration and a better overview of the market, [Bos95]. All buyers are very impressed by TFA's service, quality of products, and speed of delivery to their addresses. Buyers are very positive about the supply data presented by TFA. Buyers have a better overview of what is auctioned (compared with the traditional Dutch auctions), due to TFA's supply database.er overview and had to perform fewer administrative efforts. TFA expects a turnover of 100 million guilders for the growing season 1995/1996. Compared with the seven Dutch flower auctions, TFA ranks fourth.

The new entrants' success in general may be defined as the ability of the new entrants to build up significant sales volumes or turnovers, with fair prices for the traded products, in a short time frame. Usually, for the traditional competitors in the industry this will lead to lower sales volumes, turnovers, or price levels.

Besides the strengths of TFA, the weaknesses of the traditional Dutch auctions partly explain the success of TFA. The cooperative structure of the Dutch auctions (every single grower has one vote), the complexity of the after-sale logistics (due to the coupling of the logistics with the price discovery process), their inability to implement IT innovations quickly further decreased the market share of TFA's competitors.

6.2 Quality of Information

Based on our analysis of the SBA and TFA case, we propose the following lesson.

Lesson 3: IT innovations lead to improved quality of information; improved quality of information leads to a more efficient market, and contribute to electronic auction systems' success.

In the SBA case the system provided growers with detailed information on delivery and packaging requirements of the buyers. One of the objectives of the system was to make market information more symmetric for growers. Therefore, growers would become more consumer oriented. However, what was lacking was an adequate incentive system for the growers, because they did not get paid for fulfilling these package requirements.

In the TFA case, the information and the images on the screens, that represent the quality of the flowers, are reliable and precise. The IT innovation provides up-to-date supply information, which gives buyers a better overview. In the TFA case the supply information becomes more distributed among sellers and buyers, which will lead to more efficient markets (see [HePo88]). Kleijnen [Klei80] presents the many attributes or characteristics that determine the quality of information, such as timeliness, accuracy, aggregation, report mode, retention time, privacy and security, reliabil ity and recovery. IT innovations do not automatically lead to a higher quality of information. However, in the design of an electronic auction system one has to specify the desired level of quality of the supply information, the demand information, and the product information, for each of the stakeholders involved.

6.3 Logistical Performance

We formulate the fourth lesson.

Lesson 4: IT innovations lead to the uncoupling of logistics from price discovery pro cesses; this leads to better logistical performance, and contributes to electronic auction systems' success.

In the SBA case no full decoupling of logistics and price discovery in the auction hall was established. The coupling, by using a sample of potted plants, caused logistical performance problems.

TFA uncouples logistics and price discovery in the auction hall. Therefore, the internal logistics of the auction hall is much simpler, compared with the traditional auction system. This fact explains why TFA has a much better logistical performance and service level, in the opinion of the buyers. EAF paid much attention to the after-sales program (providing transport to the buyers).

6.4 Convergent Motives

We formulate the next lesson.

Lesson 5: Convergent motives of participants of an electronic auction system con tribute to a successful electronic auction system.

In the SBA case, the different stakeholders expected a number of different benefits. By uncoupling logistics and price discovery, the growers and the auction expected the number of transaction per hour to increase. Growers would like to obtain a better price for their products. Buyers would like to specify requirements (package material) for their customers prepared by the growers. In this case, divergent motives lead to negative effects for all stakeholders.

In the TFA case, sellers, intermediary, and buyers have convergent motives. Sellers (foreign growers) would like to have a reliable market channel for their products. They would like to obtain a fare price for their products. Sellers and TFA concentrate on the quality of their products. Buyers would like to get reliable, high-quality products, even if they have to pay a fair price for it.

6.5 Trust

We formulate the next lessons.

Lesson 6: Conformance of the actual and the perceived quality of the product, logisti cal performance, and IT performance result in high trust; High trust contributes to a successful electronic auction system.

In the SBA the buyers chose to discount the prices bid for non-sample lots by nearly ten percent because they could no longer authenticate quality by visual inspection. Logistical performance was questioned by growers, and buyers. No problems were reported about the IT performance as such.

In the TFA case, sellers and buyers find that TFA keeps their promises concerning quality of products, delivery time of products, and reliability of IT performance. Buyers trust the TFA products. Usually, they get better products than expected from the data and images provided on the PC screen, due to a centralized quality control program. Buyers also trust TFA, because the underly ing auction concept is the same: Dutch flower auction. Buyers trust the IT innovation: if a buyer is the first buyer to push the space bar at the key board of the PC, he or she is certain that the computer network transfers this signal fast and reliable, so he or she will be able to buy (part of or all of) the lot.

Lesson 7: Entry barriers are functional to maintain high trust beween sellers and buyers; high trust contributes to a successful electronic auction system.

Every member of the FAA could trade at the SBA, because of its cooperative structure. Low entry barriers did not guarantee thrustworthy business partners.

In the TFA case, EAF was strict in quality control and selecting thrustworthy growers. Therefore, higher entry barriers result in higher trust.

6.6 Learning costs

We formulate the next lesson.

Lesson 8: Applying IT innovations to unchanged auction rules results in low learning costs, and contributes to electronic auction systems' success.

One of the strengths of SBA and TFA is the use of the Dutch flower auction, as the underlying price discovery concept. The buyers are end-users of the system; they are used to work with this concept. They had to learn how to trade via the PC-screen, but these learning costs were relatively low. EAF paid much attention to the design of the user-interface of TFA. However, in the SBA case products ordered by buyers arrive one day later at the buyers box in the auction complex or at other buyer warehouses. Higher learning costs were introduced because buyers are not used to a two-day transaction cycle, in stead of a one-day transaction cycle.

6.7 Competition with other markets

The entrance of TFA into the Dutch floristry industry definitely had a profound impact on the competition in this industry. By opening TFA to non-EAF growers and buyers, TFA built critical mass in a short period. Imports of Flower Auction Aalsmeer decreased with 23 %, whereas the average was 10% for all traditional Dutch auctions [Bos96]. Recently (May 1996), Flower Auction Aalsmeer decided to re-install the liberal import policy, and to implement an electronic clearinghouse for cut flowers and potted plants. The main reason is that they would like to remain the central market place for cut flowers and potted plants.

Other auctions can have an impact on TFA. For example, Flower Auction Holland decided to copy the TFA concept; recently (August 1996), they introduced their own Tele Auction System. The auctions in Vleuten and Eelde also introduced the Tele Auction principle.

It is difficult to predict the near future of TFA. A complete uncoupling of logistics and price discovery, by sending data in stead of flowers to Amstelveen, might be possible in the near future.

7. Conclusions

The TFA case demonstrates the way a new entrant may use IT in an innovative way, in order to enter a market and compete with dominant players in that market. This case shows that new entrants can quickly build a competitive advantage. It illustrates the conclusion derived by Clemons, Croson, and Weber concerning the strengths of new entrants in a competitive market [ClCr96]. The SBA case demonstrates the weakness of the established, dominant player in implementing an electronic auction system. It illustrates that market dominance is a precursor of market failure [ClCr96].

We proposed a descriptive framework that identifies relevant elements for describing electronic auctions. The framework turned out to be useful to describe the Sample Based Auction case and the Tele Flower Auction case. In the analysis of these two case studies, we identified several lessons learned relevant to the success of electronic auction systems; these lessons are related to the nature of the innovator (new entrant versus established, dominant player), the quality of information, the logistical performance, the convergent motives of stakeholders, trust between stakeholders (also related to entry barriers), and the learning costs.

We suggest that the next step will be to refine these lessons into measurable variables and hypothe ses. With the help of multiple-case analysis, it will be possible to empirically test these hypotheses outside the flower industry. We think that in other industries IT also affects new entrants' success; for example, Clemons, Croson, and Weber [ClCr96] identify cases in the telephone and telecommunica tions industry, the retail banking industry, and the stock exchange markets. Testing in other industry settings will further generalize the results presented in this article.

8. Acknowledgement

We thank Prof. Dr. Hans Ulrich Buhl (Universitaet Augsburg) for his comments on a previous version of this paper, and the three anonymous WI reviewers for their comments and suggestions.

9. References

[Bako91] Bakos, J.A., A Strategic Analysis of Electronic Marketplaces, MIS Quarterly, 15, 3, p. 295-310, September 1991.

[Bos95] Bos, J. Th., Klok met afstandsbediening, Groot Handelsblad, p. 22-23, November 1995. (in Dutch)

[Bos96] Bos, J. Th., Een jaar TFA: de innovatieve noodsprong, Groot Handelsblad, p. 8-11, Maart 1996. (in Dutch)

[ClCr96] Clemons, E.K., D.C. Croson, and B.W. Weber, Market Dominance as a Precursor of Market Failure: Emerging Technologies and the Competitive Advantage of New Entrants, in: J.F. Nunamaker and R.H Sprague (eds.), Information Systems - Organizational Systems and Technology, Proceedings of the Twenty-Ninth Annual Hawaii International Conference on System Sciences, volume IV, IEEE Computer Society Press, Los Alamitos, p. 262-270, 1996.

[ClRe93] Clemons, E.K., Reddi, S.P., and Row, M.C., The Impact of Information Technology on the Organization of Production: The 'Move to the Middle' Hypothesis, Journal of Management Information Systems, 10, 2, p. 9-35, 1993.

[ClWe90] Clemons, E. and B. Weber, London's Big Bang: A Case Study of Information Technolo gy, Competitive Impact and Organizational Change, Journal of Management Information Sys tems, 6, p. 41-60, 1990.

[Cram95] Cramton, P.C., Money Out of Thin Air: The Nationwide Narrowband PCS Auction, Journal of Economics & Management Strategy, vol. 4, no. 2, p. 267-343, 1995.

[DaHo93] Davis, D.D. and C.A. Holt, Experimental Economics, Princeton University Press, Princeton, 1993.

[Eras95] Eras, P., Tele Flower Auction veilt via telefoon en beeldscherm, Telecommagazine, jrg. 10, nr. 10, p. 40-43, December 1995. (in Dutch)

[Fuku95] Fukuyama, F., Trust - The Social Virtues and the Creation of Prosperity, New York, The Free Press, 1995.

[Grif94a] Griffioen, A., Goede start informatieveilen in Aalsmeer, Vakblad voor de Bloemisterij, nr.2, pp 9, 1994.

[Grif94b] Griffioen, A., Informatieveilen moet de stagnatie overwinnen, Vakblad voor de Bloemis terij, nr.7, pp. 56-57, 1994.

[GuWh91] Gurbaxani, V. and S. Whang, The Impact of Information Systems on Organizations and Markets, Communications of the ACM, 34, 1, p. 59-73, 1991.

[HaTa92] Haak, M., H. Tap and A.M.A. Heybroek, A View of International Competitiveness in the Floristry Industry, Rabobank Nederland i.c.w. LEI-DLO, Eindhoven/The Hague, 1992.

[HeKe94] Hess, C.M. and C.F. Kemerer, Computerized Loan Origination Systems: An Industry Case Study of the Electronic Markets Hypothesis, MIS Quarterly, 251-275, september 1994.

[HePo88] Hendricks K. and R.H. Porter, An Empirical Study of an Auction with Asymmetric Information, The American Economic Review, 78, p. 865-883, 1988.

[KaVh96] Kambil, A. and E. van Heck, Re-engineering the Dutch Flower Auctions: A Framework for Analyzing Exchange Organizations, New York University, Department of Information Systems, Working Paper Series, Stern IS-96-24, New York City, 1996.

[Klei80] Kleijnen, J.P.C., Computers and Profits: Quantifying Financial Benefits of Information, Addison-Wesley, Reading, 1980.

[KoWa89] Konsynski, B., A. Warbelow, J. Kokuryo, Aucnet: TV Auction Network System, Har vard Case Study 9-90-001, Harvard Business School, Boston, 1989.

[Lee96] Lee, H.G., Electronic Brokerage and Electronic Auction: The Impact of IT on Market Structures, in: J.F. Nunamaker and R.H Sprague (eds.), Information Systems - Organizational Systems and Technology, Proceedings of the Twenty-Ninth Annual Hawaii International Conference on System Sciences, volume IV, IEEE Computer Society Press, Los Alamitos, p. 397-406, 1996.

[MaYa87] Malone, T.W., J. Yates and R.I. Benjamin, Electronic Markets and Electronic Hierar chies, Communications of the ACM, 30, 6, p. 484-497, 1987.

[Milg89] Milgrom, P., Auctions and Bidding: A Primer, Journal of Economic Perspectives, vol.3, no.3, p. 3-22, Summer 1989.

[RoHa94] Rothkopf, M.H. and R.M. Harstad, Modeling Competitive Bidding: A Critical Essay, Management Science, vol.40, no.3, p. 364-384, March 1994.

[Schm93] Schmid, B., Elektronisch Märkte, Wirtschaftsinformatik, 35, 5, pp. 465-480, 1993.

[Vdme95] Van der Meer, M., Effecten strak importbeleid heropenen discussie, Vakblad voor de Bloemisterij, nr. 22, p. 26-29, 1995. (in Dutch)

[VhGr94] Van Heck, E. and T.P. Groen, Towards an Electronic Flower Auction?, in: G. Hagelaar (ed.), Management Studies and the Agri-business 1994: Management of Agri-Chains, Wageningen Agricultural University, Wageningen, p. 218-233, 1994.

[VhRi96] Van Heck, E. and P.M.A. Ribbers, Economic Effects of Electronic Markets, Discussion Paper No. 9669, Center for Economic Research, Tilburg University, Tilburg, July 1996.

[VhVd97] Van Heck, E., E. van Damme, J. Kleijnen, P. Ribbers, New Entrants and the Role of Information Technology, Case-Study: the Tele Flower Auction in the Netherlands, in: J.F. Nunamaker and R.H Sprague (eds.), Information Systems - Organizational Systems and Technology, Proceedings of the Thirtieth Annual Hawaii International Conference on System Sciences, volume III, IEEE Computer Society Press, Los Alamitos, pp. 228-237, 1997.

[Vvli94a] Van Vliet, C., Telers wijzen importbeleid veilingen massaal af, Vakblad voor de Bloemis terij, nr. 37, p. 22-27, 1994a. (in Dutch)

[Vvli94b] Van Vliet, C., EAF wil met importveiling centrum bloemenhandel in Nederland houden, Vakblad voor de Bloemisterij, nr. 49, p. 40-41, 1994b. (in Dutch)

[Vvli95] Van Vliet, C., Merendeel telers wil huidige importbeleid continueren, Vakblad voor de Bloemisterij, nr. 22, p. 30-31, 1995. (in Dutch)

[Vvli96] Van Vliet, C., Gewijzigd importbeleid verzwakt Nederlandse bloemisterij, Vakblad voor de Bloemisterij, nr. 4, p. 26-29, Januari 1996. (in Dutch)

[Yin89] Yin, R.K., Research Design Issues in Using Case Study Method to Study Manage ment Information Systems, in: J.I. Cash and P.R. Lawrence (eds.), The Information Systems Re search Challenge: Qualitative Research Methods, volume 1, Harvard Business School, Boston, 1989.

About the authors

Eric van Heck is an Assistant Professor at Erasmus University Rotterdam in the Department of Decision & Information Sciences. His research focuses on interorganizational strategies and information systems (EDI, electronic markets, and electronic commerce).

Pieter M. Ribbers is Professor of Information Systems at Tilburg University, School of Economics, the Netherlands, where he chairs the Postgraduate School of Information Management. His research interests span information economics, interorganizational systems, and the strategic and organizational consequences of the use of Information Technology.