ONLINE, REVERSE AUCTIONS: ISSUES, THEMES, AND PROSPECTS FOR THE FUTURE*

Invited manuscript forthcoming in the MSI/JAMS Special Issue on Marketing to and Serving Customers Through the Internet: Conceptual Frameworks, Practical Insights and Research Directions

Sandy D. Jap**

May 2002

* This research effort was supported by a financial grant from the Center for EBusiness@MIT at the Massachusetts Institute of Technology. The participant firms and B2eMarkets provided data support. The manuscript has benefited from the insightful comments of Ajay Kohli, an anonymous reviewer, and the Special Issue editors. ** Sandy D. Jap is Associate Professor of Marketing at the Goizueta Business School at Emory University, 1300 Clifton Road, Atlanta, GA 30322-2710. Phone 404.727.7056, fax 404.727.3552, www.bus.emory.edu/sdjap, [email protected]

ONLINE, REVERSE AUCTIONS: ISSUES, THEMES, AND PROSPECTS FOR THE FUTURE ABSTRACT Online, reverse auctions are increasingly being utilized in industrial sourcing activities. This phenomenon represents a novel, emerging area of inquiry with significant implications for sourcing strategies. However, there is little systematic thinking or empirical evidence on the topic. In this paper, the use of these auctions in sourcing activities is reviewed and four key aspects are highlighted: (i) the differences from physical auctions or those of the theoretical literature, (ii) the conditions for using online, reverse auctions, (iii) methods for structuring the auctions, and (iv) evaluations of auction performance. Some empirical evidence on these issues is also provided.

ONLINE, REVERSE AUCTIONS: ISSUES, THEMES, AND PROSPECTS FOR THE FUTURE INTRODUCTION For nearly the past decade, managers, analysts, researchers, and the business press have been remarking that, “The Internet will change everything.” And since the advent of the Internet, we have seen it challenge nearly every aspect of marketing practice. This raises the obligation to consider the consequences of the Internet to management practices, the theme of this special issue. Yet, it may take decades to fully understand the impact of the Internet on marketing practice, in general. This paper is one step in that direction. Specifically, I consider the impact of the Internet in a business-to-business context, the sourcing of direct and indirect materials from a supply base. It has been predicted that the Internet will bring about $1 trillion in efficiencies to the annual $7 trillion that is spent on the procurement of goods and services worldwide (USA Today, 2/7/00, B1). How and when this will happen remains an open question. However, one trend that is showing increasing promise is the use of online, reverse auctions. Virtually every major industry has begun to use and adopt these auctions on a regular basis (Smith 2002). During the late 1990s, slow-growth, manufacturing firms such as Boeing, SPX/Eaton, United Technologies, and branches of the United States military, utilized these auctions. Since then, consumer product companies such as Emerson Electronics, Nestle, and Quaker have followed suit. Even high-tech firms such as Dell, Hewlett-Packard, Intel, and Sun Microsystems have increased their usage of auctions in sourcing activities. And the intention and potential for the use of these auctions to continue to grow in the future is clear. In their annual survey of purchasing managers, Purchasing magazine found that 25% of its respondents expected to use reverse auctions in their sourcing efforts. Currently, the annual throughput in these auctions is estimated to be $40 billion; however, the addressable spend of the Global 500 firms is potentially $6.3 trillion.

1

The popularity of online, reverse auctions can be attributed to three key factors. The first is that online, reverse auctions create immediate financial savings. These auctions have been shown to produce anywhere from 5-40% cost savings (Tully 2000), with 15% being more typical (Cohn 2000). When you consider that a major manufacturer procures billions of dollars worth of goods on annual basis, a 15% savings can have a huge impact on the bottom line. Another way to think about this is to realize that if a firm has a 20% gross margin, every $1 saved in procurement is the equivalent of adding $5 in revenue. The second factor in their rapid growth is the process efficiencies that auctions create. In the past, buyers would spend approximately six weeks from the time a request for purchase (RFP) is created to the point at which a set of viable bids are available for consideration. These six weeks are spent mailing, faxing, and delivering papers back and forth, numerous voice mails and phone conversations, with multiple rounds of bidding shepherded throughout the process. By holding the reverse auction online and employing electronic communications, this process can be reduced to the span of several hours. The third factor in the growth of these auctions is the enabling capabilities of emerging technologies. The technology behind an auction format is not complex; hence, many firms can offer them as part of their sourcing solutions or software (e.g., B2eMarkets, CommerceOne, Oracle). The technology also enables unprecedented temporal and geographical efficiencies, making it easier than ever to organize and host an online, reverse auction. Clearly, the utilization of online, reverse auctions is here to stay and will continue to become an important aspect of sourcing activities in the future. However, this raises a myriad of issues regarding their role in procurement strategies. In virtually every industry, buyers are asking similar questions: What’s new and different about online, reverse auctions? When should they be used? How should I set up an online, reverse auction? What suppliers should be invited to bid? How will they react to the format? How do I evaluate whether the auction is worthwhile? Will I save money 2

if I run an online, reverse auction? What will the long-term impact be, if any? Essentially, these questions can be boiled down to four key issues: (i) Differences, from physical auctions or those studied in the theoretical literature, (ii) Conditions, or when to use online, reverse auctions, (iii)Structure, or how to use these auctions, and (iv) Evaluation, or why the auctions should be used. Together these aspects provide a perspective on the process of using online, reverse auctions in industrial sourcing activities. Buyers contemplate how online reverse auctions differ from past sourcing techniques, then consider when they should use the auctions, how the auctions should be used, and then evaluate whether the auction was worthwhile. To date, there is little systematic work on these issues. While there is a growing body of anecdotal stories and reports in the marketplace, the area remains ill defined and fragmented. In this paper, I reflect on the role of online, reverse auctions in industrial sourcing activities by offering some perspectives on when they should be used, how they should be used and why the auctions are useful. The intention is to examine the marketplace phenomenon and provide a structured overview of the area with the goal of highlighting the advantages and disadvantages of these auctions and opportunities for research. There is not, however, an attempt to be exhaustive of the possibilities within each aspect.1 The paper is organized as follows. In the next section, the various types of online, reverse auctions are discussed and differentiated from the physical auctions used in industrial purchasing and the auctions of the theoretical literature. This is followed by a description of the conditions for using auctions, considerations for structuring auctions, and evaluation issues. An overview of the discussion is provided in Figure 1. When appropriate, exploratory empirical results from a recent survey of industrial buyers are offered throughout the paper. Concluding remarks include additional long-term considerations and opportunities for additional research.

3

ONLINE, REVERSE AUCTIONS SOME BASIC DEFINITIONS Before beginning the discussion, it is useful to clarify some definitions and terms that will be used throughout the paper. To begin with, an auction is a market institution with an explicit set of rules determining resource allocation and prices on the basis of bids from market participants. In theoretical terms, auctions play a prominent role in the theory of exchange, as they remain one of the simplest and most familiar means of price determination in the absence of intermediate market makers. The auctions that are used in sourcing activities have been dubbed “reverse” to reflect the fact that sellers bid and the goal of these events is to push the price down. This is in contrast to the more common form of auction, the “forward” auction, in which buyers bid and the seller’s goal is to push the price up. In this paper, the discussion will be restricted to the use of one-sided, online, reverse auctions, with “one-sided” referring to the fact that there is a single buyer instead of multiple buyers. The format of online, auctions can be represented by two extremes: open vs. sealed bid formats. These formats vary along a price visibility continuum in which asynchronous, sealed bidding formats would anchor one end and simultaneous, open bidding formats would anchor the opposite end of the continuum. The term, “sealed” refers to the fact that only one supplier and the buyer has access to the details of a bid. The bid process is asynchronous in the sense that the buyer and supplier take turns viewing the bid. The buyer posts the RFP electronically, the supplier submits a bid, then the buyer views the submitted bid, and either makes a decision after viewing all bids or if multiple rounds of bidding are involved, the buyer may respond to the supplier who then resubmits a new bid. In contrast, an open bid event is one in which the buyer electronically posts an RFP and during the course of a designated time period (usually several hours), suppliers are invited

4

to bid simultaneously on the contract. In this situation, all suppliers and the buyer view the bids at the same time. This is why the process is referred to as “open.” Many open bid formats in industrial purchasing utilize a moving soft-close rule for ending the auction, instead of a hard-close rule. A hard-close rule, which is often used in forward auctions such as eBay.com, is one in which the end time is fixed. In contrast, a soft-close rule means that if there is a bid within a prespecified time period prior to the closing time of the event, then the real close time will automatically extend for a few minutes more to allow other bidders to respond. This will continue until a close time is reached with no bidding activity in the minute prior. The motivation behind this is to allow for price competition to run its full course, and to avoid “sniping,” which is the practice of waiting until the last minute before the auction ends and trying to submit a bid which just barely beats the high bid and gives rival bidders no time to respond. With a hardclose rule, the event becomes essentially a sealed bid event, as all bidders submit their best bids without being able to view or respond to competitive bids. OPEN BID ONLINE, REVERSE AUCTIONS Figure 2 provides an example of the format of an open bid auction. Bid amounts are plotted on the vertical axis and time is plotted on the horizontal axis. A diamond indicates a supplier’s bid, while a line connects the lowest bid at any point in time. The buyer’s posted reserve price is the price at which it would be willing to move the purchase contract from an incumbent to a new supplier. This is essentially the buyer’s switching cost, reflecting the risk that the buyer faces in moving the business. The pattern of bids in Figure 1 indicates that most activity occurs close to the scheduled closing time, and in this case, the buyer has chosen to use a soft close on this event. At this point, bidding becomes aggressive and the price falls dramatically over the next 30 minutes. It is also worth noting that bids are being made above the reserve line throughout most of the auction. These represent price concessions by suppliers who may not be willing to bid at the market 5

lead. In many industrial auctions, the buyer may not necessarily commit to the lowest or secondlowest price bidder. Instead, it commits to select one winner among the bidders, but selection may occur on any basis other than price. This is the typical approach in product categories where additional non-price considerations are important such as the procurement of direct materials. As such, suppliers in the open bid event may bid above the market lead because they hope to remain in the buyer’s consideration set by signaling a willingness to make some price concessions, while relying on their non-price value to justify their inability to offer the lowest prices in the market. Hence, while the open bid auction involves a dynamic competition on price, it is also an important signaling mechanism whereby all suppliers can make decisions regarding price concessions in real time. DIFFERENCES FROM PHYSICAL AUCTIONS The question often arises, “Are online, reverse auctions a new phenomenon or are they essentially the same type of phenomena as the manual auctions that have historically been used in sourcing activities?” The short answer is yes and no. Yes, there are many similarities; the goals of online and manual reverse auctions are to secure the best possible price on a particular good. Suppliers are often prequalified and invited to bid after viewing the RFP. However, the answer is also No, because there are many aspects of online, reverse auctions that are sufficiently different from manual auctions, such that they require a different type of management process from those in the past. These differences are classified as differences of degree and differences of kind. The former refers to aspects of the manual negotiation process that have been improved by emerging technologies, whereas the latter refers to aspects that are fundamentally different. Differences in degree. Online, reverse auctions are generally cheaper and easier for buyers and auction providers to organize than manual auctions. Online technologies have enabled rapid and efficient forms of communication that allow buyers to invite more suppliers to events while 6

requiring less time to notify and organize these suppliers. And suppliers from all around the world can easily participate in the event without having to travel to a remote, physical location and congregate with other suppliers in order to bid. These technologies also enable auctioneers to provide immediate feedback to the bidder if there is a problem on the bid, whereas bidders in a manual auction process would not receive the feedback until the event is over. Differences of kind. The negotiation format of manual versus online auctions is clearly different. In manual auctions, suppliers bid face-to-face, while in online auctions, suppliers use a computer mediated interface and the auctioneer provides the software that allow suppliers to make, modify, and view bids. Another significant difference is that the identity of suppliers in an online auction is typically anonymous, whereas in manual auctions, suppliers would be able to observe who their competitors are and what they are bidding. All of these characteristics combine to make the online, reverse auctions of today a different phenomenon than the manual auctions of the past. DIFFERENCES FROM AUCTIONS IN THE THEORETICAL LITERATURE The economics literature considers auctions from both a theoretical and empirical perspective, with a focus on the process by which individual actions translate into prices. Hence, this literature considers differences among the valuations that bidders may have for auction objects, characteristics of the bidder (e.g. risk averseness), psychological mistakes (e.g., the winner’s curse) and how these factors affect the prices in a variety of auction formats (e.g., sealed, open, English, Dutch, etc.) that vary in their allocation rules. For an overview of this literature, see Milgrom (1989), McAfee and McMillan (1987), and Kagel (1995). Are the online, reverse auctions that we observe in the marketplace similar to or different from the auctions described in this large body of literature? At first blush, there appear to be a number of similarities between the marketplace and theoretical auctions; however, there are also a number of significant differences.

7

The similarities between marketplace and theoretical auctions include a focus on price bidding among competing sellers and the possibility of multiple buyers. The auctions are a tool in the determination of resource allocation and the format provides a structured space and well-defined rules along which bids may be offered. And for industrial auction events that are similar in format and assumptions to those of the economics literature, one would expect that the findings of this literature would be applicable. For example, in the sourcing of indirect materials2, which include goods not used directly in the manufacturing process, many online, reverse auctions are similar to the theoretical literature. Hence, the results should generalize, although this has yet to be demonstrated (cf., Kagel 1995). There are, however, a number of significant differences that make it difficult, if not impossible, to generalize from the economics literature to many industrial settings. Differences in degree. Many of the auctions in the theoretical literature involve commoditized products, for which price essentially determines the complete value of the product. In contrast, many marketplace auctions may involve products that, while largely determined by price, may also be differentiated by quality or other non-price attributes. For example, many marketplace auctions involve the purchase of direct materials, production materials for which aspects such as supplier quality and delivery may be important factors. Differences in kind. One of the major differences between online, reverse auctions and those of the theoretical literature is that the vast majority of auctions used in the marketplace today do not determine a winner. The online, reverse auction is merely a mechanism for clearing suppliers on one basis, namely price, and the buyer may reserve the right to select the winner on any basis. In contrast, the auctions in the economics literature will explicitly specify a priori whether the winner will be determined on a first- or second-price basis. As a result, bidders in an open auction know exactly where they stand relative to their competitors and can use this information to determine how to respond to competitive bids. In online, industrial auctions where suppliers are not 8

told the buyer’s selection rules, the suppliers may not have a real understanding of how competitive their offer was or even why they lost the purchase contract. In fact, it is not uncommon for suppliers to wait as long as 4-6 weeks before a winner is announced. And then there is often little feedback given as to how or why the contract was allotted. A second major difference between online, reverse auctions in the marketplace and those of the theoretical literature is the fact that many marketplace auctions require suppliers to bid on a series of interdependent product lots. A lot is a sub grouping of parts; for example, a buyer may need to purchase 20 parts and will group these parts into 4-5 lots. By conducting an event with multiple lots, the buyer hopes to capitalize on supplier competencies and synergies in manufacturing or distribution while also creating a sizable contract value that motivates suppliers to participate. The division of parts into lots is a subjective decision of the buyer, who may divide them on the basis of the suppliers’ capabilities to bid on or produce each lot, similarities in manufacturing processes, delivery regions, etc. Suppliers who bid on these parts will then bid on each lot sequentially over a multi-hour time period. However, it is important to note that the use of lotting practices in an auction event creates interdependence among the bids for various lots; the bids that are placed in the first lot may determine or impact the supplier’s bids for the next lot, and the next and so forth. When this interdependence of bids is combined with an award rule in which the buyer determines the winner, then bidding in an online, reverse auction can become a very complex proposition. From the supplier’s perspective, it has a fixed capacity and cost structure that it attempts to account for in its bid in each lot. When the supplier doesn’t know for certain whether it has won a lot, this decision process becomes extremely difficult, because it creates uncertainty about how and whether the supplier bids on the next lot. Interdependence among auction items has

9

yet to be examined in the economics literature; the closest related research considers multiunit purchases of homogenous commodities (Kagel and Levin 2001; Swinkels 2001). To summarize, online, reverse auctions are differentiated from the manual auctions of the past in that the former enable geographic and temporal conveniences, and immediate bidder feedback that were not possible before. Additionally, suppliers in online auctions bid via a computer mediated environment and are typically anonymous to competing suppliers. Online reverse auctions are also differentiated from the auctions of the theoretical literature in that the former may involve the purchase of non-commoditized products, do not necessarily determine a winner, and is marked by an interdependent bidding process in which suppliers must allocate their fixed resources across heterogeneous product lots. Together, these differences would suggest that (i) our prior understanding of how to manage the physical auction process and (ii) the findings and theories associated with the economics literature may not necessarily generalize to the online, reverse auction sourcing activities currently being witnessed in the marketplace. These auctions appear to be a fundamentally different phenomenon, both in degree and kind, than the manual auctions of the past or the auctions of the economics literature. This is not to say that there is no overlap or similarities to manual auctions or the auctions of the theoretical literature. And this does not imply that the insights generated in these areas are irrelevant to industrial, online, reverse auctions. Instead, the critical point is to realize that the dynamics and use of online, industrial sourcing auctions are very different from those of the past and it is important to carefully consider these differences when investigating this phenomenon. It is also worth noting that the focus of my own interest in and subsequent comments on this topic also differs from that of the past. This paper is less concerned about explaining how bidders determine prices across a variety of auction formats and more concerned about the role of online,

10

reverse auctions within the broader organizational context. This is a critical direction of research that has been acknowledged and called for by researchers in the theoretical literature: The auction models are partial equilibrium models. The role of the price system in coordinating the actions of different people cannot be understood except within a general equilibrium system. How to embed bidding models in a general-equilibrium context remains an open question. Questions of the existence and social optimality of competitive equilibrium with informational asymmetries await the resolution of this question. – R. Preston McAfee and John McMillan (1987)

These experts point out that in practice, auctions do not occur in a vacuum. In fact, they are deployed into organizational settings that may systematically differ and these differences may interact with auction outcomes. Hence, more work is needed on understanding how auctions should be applied across specific organizational contexts, as such contexts could play a key role in the successful adoption and implementation of online auctions into industrial sourcing activities. EXPLORATORY EMPIRICAL EVIDENCE Systematic, empirical evidence on the role of online, reverse auctions in sourcing activities is scant. As a first step in better understanding this area, I recently conducted an exploratory survey of four Fortune 100 buying organizations in consumer foods, industrial chemicals, high-tech systems, and the automotive industries. Collectively, these four firms provided the name of 54 sourcing managers who were invited to complete an online survey. The survey directed the respondent to report on an online, reverse auction experience within the past two years and to complete all measures in regard to that specific auction event. Hence, the unit of analysis was a single auction event. The survey measured aspects of the conditions under which their auctions were held, how the auctions were structured, and the manager’s evaluations of the outcomes of the auction. The scales for multiple item measures were adapted from previous research or created specifically for the purposes of this study. Because of the limited sample size, a confirmatory factor analysis was not conducted, but an exploratory factor analysis was, to verify that the items did in

11

fact, load on a single factor. A complete listing of the items, constructs, and their coefficient alphas is displayed in Appendix 1. Thirty-eight of the managers completed the survey, yielding a 70% response rate. These managers had on average, 5.2 (sd=5, min=1 max=24) years of experience in sourcing and had completed an average of 2.3 (sd=2.4, min=1, max=12) online, reverse auctions. The products that they reported on included steel parts, raw materials, components, piping, and sheets; these were primarily used directly in their production activities. Appendix 2 contains a complete listing of the products that were purchased through these auctions. 20% of the lots in these events involved the sourcing of new contracts. 78% of the auctions were open bid auctions, while the remaining 22% were sealed bid auctions. Hence, the results are skewed to reflect open bid auctions, although differences across the two formats are assessed; any significant differences are noted throughout the manuscript. In the sections to follow, aspects of when, how, and why online reverse auctions are used in sourcing strategies are highlighted and the data is used to illuminate or illustrate the various concepts. The data set is small and primarily used for illustrative purposes; it should, however, be considered exploratory in nature. CONDITIONS FOR ONLINE, REVERSE AUCTIONS In recent years, Purchasing Magazine published an article entitled, “The E-Auction Playbook,” which provided a number of marketplace perspectives on the use of online, reverse auctions in sourcing activities. These perspectives came from a variety of buying executives (e.g., Eastman Kodak, Sprint, United Technologies), purchasing and auction intermediaries (e.g., Chem Connect, Freemarkets, Moai Technologies, etc.), suppliers (e.g., ImageX.com), and consultants (e.g., CGEandY), with each firm offering up suggestions on when, how and why auctions should be used in industrial sourcing activities (Purchasing 2001). On the basis of this and many other 12

articles that have appeared on this topic and my own field interviews, research, and experiences with auctions in a variety of product categories, I consider three conditions that are particularly critical: product characteristics, sourcing strategies, and supply base characteristics. Research ideas and themes are sprinkled throughout the ensuing discussion in each section. These ideas, along with additional related issues, are summarized in Table 1. PRODUCT CHARACTERISTICS Price-based products. Products for which the purchase price constitutes the largest component of its value are ideally suited for online, reverse auctions. Examples of such products from the survey data include: cables, drives, metal parts, software, and switches. Additional examples can be found in Appendix 2. When the value of a product is easily expressed quantitatively, online auctions provide an efficient mechanism by which to rapidly evaluate all suppliers and set the purchase price. Under these conditions, the theoretical literature does much to inform this phenomenon. In the exploratory survey, respondents report that the products they put through auctions are ones in which on average, 83% of the total cost of ownership of the product (which includes aspects such as duty, freight, inventory, and carrying costs) are accounted for by the purchase price alone. When the products are commoditized, the capabilities of Internet-based technologies are particularly valuable; they enable unprecedented temporal and geographical conveniences that allow both players to benefit. Buyers are able to expand their reach into distant markets and avail themselves of new, alternative suppliers, while suppliers are able to bid on more events and attract new customers. Consistent with this, a correlational analysis indicates that as the product becomes increasingly price-based, buyers evaluate alternative suppliers more positively (r=.53, p