基于Internet的供应链选择

基于Internet的供应链选择Serguei Netessine费城，宾夕法尼亚州19104号，宾夕法尼亚大学，Nils Rudi欧洲工商管理学院，法国枫丹白露77305号，nils.Internet companies extensively use the practice of Drop-shipping, where the wholesaler stocks and owns the inventory and ships products directly to customers at retailers request. Under the Drop-shipping arrangement, the supply chain benets from risk pooling because the inventory for multiple retailers is stocked at the same location, the wholesaler s. Another more traditional channel alternative on the Internet is one in which retailers stock and own the inventory. These two supply chain structures, which predominate on the Internet,result in different inventory risk allocation, stocking decisions, and prots for channel members. Moreover, the two channel alternatives can be combined into a dual strategy whereby the retailer uses local inventory as a primary source and relies on Drop-shipping as a backup.We model the dual strategy as a noncooperative game among the retailers and the wholesaler, analyze it, and obtain insights into the structural properties of the equilibrium solution to facilitate development of recommendations for practicing managers. Finally, we characterize situations in which each of three channels is preferable by specifying appropriate ranges of critical parameters, including demand variability, the number of retailers in the channel, wholesale prices, and transportation costs. Key words: supply chain; Drop-shipping; e-commerce; fulllment; equilibrium; competitionHistory: Accepted by Fangruo Chen, supply chain management; received December 11, 2001. This paper was with the authors 1 year and 10 months for 4 revisions.Alliance Entertainment Corporation, Ingram Entertainment, and Baker and Taylor are major players in the business of distributing home entertainment productsincluding CDs, videos, games, and booksto Internet, as well as to brick-and-mortar, retailers. One of the services they offer is Drop-shipping, or consumer-direct fulllment, the practice whereby the wholesaler stocks and owns the product and ships it directly to customers at the retailers request. Under a Drop-shipping arrangement, the retailer serves as a middleman who acquires customers and accepts orders, while the wholesaler owns and holds inventory and also fullls orders. Drop-shipping offers two obvious benets to the retailer: No investment into fulllment capabilities is required, and the retailer takes no inventory risk (see the example of S, Forbes 2000). Drop-shipping also benets the wholesaler in that it allows her to charge a higher wholesale price and may provide access to a wider customer base (see Scheel 1990, p. 42). Finally, one would expect that the supply chain as a whole would benet due to risk pooling and possible economies of scale in transportation in cases in which the wholesaler serves multiple retailers.The invention of the Internet, a relatively cheap electronic medium, greatly reduced the transaction costs of Drop-shipping by allowing a seamless integration of retailers and wholesalers. The Internet allows in-stock availability and prices to be communicated from the wholesaler to the retailer, and orders can be placed from the retailer to the wholesaler, all in real time. Before the Internet, such integration would have required implementing a rather expensive EDI system and hence was not economically viable in most channels. As a result of this integration of retailers and wholesalers, in a recent survey 30.6% of pureplay Internet retailers cited Drop-shipping as their primary way of fullling orders. At the same time, 44.5% of Internet retailers relied primarily on stocking inventory internally (see eRetailing World 2000), whereas the rest either outsourced fulllment to third parties or used the fulllment capabilities of strategic partners.Supply chain arrangements in which the retailers stock and own inventory have been extensively analyzed in the operations literature. However, the practice of Drop-shipping, although almost as popular among Internet retailers, has attracted only limited attention. The marketing literature offers some references to Drop-shipping: Although Scheel (1990) provides extensive qualitative analysis of Drop-shipping in the catalog business, he also notes that the practice is generally perceived as having somewhat limited potential. The goal of this paper is to analyze comparative advantages of inventory ownership in the traditional channel and risk pooling under Drop-shipping.企业在互联网环境下已广泛的使用Drop-shipping模式，这是一种由供应商负责客户订单，拥有库存并由他们代发货给零售商指定客户的模式。在Drop-shipping模式下，供应商管理各零售商的库存，有利于分散供应链上的风险。另一种更为传统的互联网渠道选择是零售商制定库存决策并持有库存的模式。这两种基于Internet的供应链模式导致了渠道成员之间的不同库存风险分担，库存决策，以及利润分配。此外,这两种渠道可以合并为一种双重渠道模式,即以传统零售库存模式作为主体，以Drop-shipping库存的模式作为备选。我们在零售商和供应商非合作博弈想讨论这种双重渠道模式，分析并获得了此模型均衡解决过程中的结构特征并，用以提高该模式下的管理水平。最后，我们指定三种渠道模式下各模式的需求变化, 渠道中零售商的数量,，批发价格和运输成本等的临界参数范围，以确定哪种模式适合哪种情况。关键词：供应链 Drop-shipping 电子商务 履行 均衡点 竞争历史: 于2001年12月11日，供应链管理中被陈芳若推崇。本文是作者耗时一年零十个月经四次修改完成的。1、 简介及研究综述联盟娱乐公司,英格拉姆娱乐,以及贝克和泰勒是在互联网和实体环境中主要从事家庭娱乐产品分销业务（包括 CD、 视频、 游戏和图书）的主要零售商。他们提供的服务之一是Drop-shipping，或者叫做消费者直接实现，这是一种供应商负责客户订单，拥有库存并由他们直接发货给零售商指定的客户的模式。在一个Drop-shipping模式下，零售商作为中间人负责取得客户，并接受订单，而供应商则拥有库存，并履行订单。Drop-shipping对于零售商来说有两个明显的优势：一是不需要对仓库进行投资，二是没有库存风险（见2000年S ，福布斯 的例子）。Drop-shipping渠道对供应商也有好处，即它允许供应商抬高批发价格，并且可能使供应商接触到更广泛的消费者群体。（见Scheel，1990年, 第42页）。最后，我们预期供应链作为一个整体将从风险共担和供应商为多个零售商服务以在运输中实现规模经济中获益。Internet的发明大大降低了Drop-shipping下零售商和供应商无缝整合的交易成本。在互联网环境下实现了从供应商到零售商实时进行入库管理和价格谈判，并且可以实现订单从零售商到供应商的实时转移。在互联网出现之前，这种整合必须使用一个相当昂贵的EDI系统，因此从经济角度来说，这在大多数渠道上都是不可行的。正是由于零售商和供应商基于Internet的整合，最近的一项调查显示，30.6的Internet零售商使用Drop-shipping作为他们的订单主要履行方式。同时，44.5的Internet零售商主要依靠自有库存（见电子零售世界，2000年），而其余的要么采用第三方外包的库存要么采用战略合作伙伴的库存。对于零售商自己持有库存的传统供应链模式，已经有很多文献进行了研究。虽然目前Drop-shipping在Internet零售业日益流行，但是还少有学者涉足基于Internet的双重渠道的研究。市场营销文献提供了一些关于Drop-shipping的参考：虽然1990年谢尔提供了广泛的Drop-shipping业务的定性分析，然而他同时指出，这种做法通常被认为只有有限的潜力。本文的目的是分析比较传统渠道模式零售商拥有库存和Drop-shipping模式下风险共担的优势。1.1. Summary of ResultsWe begin by presenting two alternative models: the traditional channel, in which each retailer stocks inventory locally, and the Drop-shipping channel, in which the wholesaler stocks inventory centrally and ships directly to customers. We model both channels as two-stage supply chains with one wholesaler and multiple identical retailers. In both models the wholesaler charges the retailers a xed wholesale price (not necessarily the same). Additionally, based on our conversations with Internet companies, we assume that different channels may have different transportation costs due to economies of scale and possibly other cost differences.When faced with a choice between the traditional and Drop-shipping channels, the retailer trades off a higher wholesale price against inventory risk, and the wholesaler trades off the lower margin against inventory risk. Moreover, transportation cost differences play a role. By combining the two strategies, the retailer can rst satisfy demand with the internally stocked inventory (purchased at a lower wholesale price) and then request the wholesaler to drop-ship the rest. For example, BlueL, a major Internet retailer, uses the dual strategy for CDs: It stocks certain quantities of the top 40 CDs to increase prot margins and relies on Drop-shipping in case of a stockout (see Apparel Industry Magazine 2000) while utilizing pure Drop-shipping for the remaining CD titles. Also, several major retailers have recently installed Internet kiosks inside physical stores (see Computerworld 2001); if a customer does not nd a product on the store shelves, she can order it in store over the Internet and the product will be shipped directly to her from a central warehouse. Such an arrangement resembles the dual strategy we analyze.The dual strategy is somewhat similar to the combination of make-to-stock and assemble-to-order practices discussed in the operations literature () with one major difference: Under the make-to-stock/assemble-to-order practice, the same company decides both how much to stock and how much to assemble to order, whereas under the dual strategy considered in this paper, the retailers and the wholesaler manage their inventories competitively, such that a game theoretic situation arises. We model the dual-strategy problem as a noncooperative game in which retailers compete with the wholesaler for demand and, moreover, they compete with each other for the wholesalers inventory allocation. We analyze equilibria of this game, derive the structural properties of players best-response functions, and use them to show the uniqueness of the symmetric equilibrium. One of our results is that in equilibrium under the dual strategy it may be optimal for the wholesaler or for retailers to stock nothing, and thus the dual channel converts into a pure traditional or a pure Drop-shipping channel. We are able to identify simple conditions that lead to such boundary equilibria. Using properties of supermodular games and derived properties of the best-response functions, we conduct sensitivity analysis of the equilibrium to the changes in the wholesaleprice, the Drop-shipping markup, and transportation cost parameters.We further compare the traditional, Drop-shipping, and dual channels using numerical experiments. We nd that when the number of retailers is large, the benets of risk pooling make either the Drop-shipping or the dual channel more attractive than the traditional channel to both retailers and the wholesaler. This nding is consistent with the practical observation that Drop-shipping is often used for books and electronic products, two segments in which there are many retailers, most served by a few large wholesalers. In the dual channel we nd that changes in any cost-related parameters result in the reallocation of inventory between retailers and the wholesaler.For example, in this channel, increasing the markup on drop-shipped products leads to higher retail inventory and lower wholesale inventory, whereas in the Drop-shipping channel the same action would lead to higher wholesale inventory. Surprisingly, the impact of the transportation cost differential on prots in the dual channel is not always intuitive. In particular, as the transportation cost structure becomes more favorable for Drop-shipping, the wholesalers prot initially drops and the retailers prot increases.请键入文字或网站地址，或者上传文档。We nd that the Drop-shipping markup and the transportation cost differential are the main drivers of the choice of channel. We observe that both the Drop-shipping and the dual channels have the potential to be Pareto-optimal choices, but the traditional channel does not. This nding indicates that most Internet retailers should strive to utilize Drop-shipping at least to augment their internal inventory. However, in some industries wholesalers prefer the traditional channel, and hence do not want to stock any inventory for Drop-shipping, so this option is simply not available. In particular, this scenario arises when the Drop-shipping markup is relatively small but transportation costs favor Drop-shipping, so that the wholesaler does not make any money on shipping directly to customers even though the retailer prefers this option. In situations with a large set of problem parameters the dual channel is the Pareto-optimal choice, which is consistent with the proliferation of Internet kiosks and the increasing popularity of this arrangement. Due to the nonintuitive impact of transportation costs, the dual channel becomes the Pareto-optimal choice when the transportation costs are much higher in the Drop-shipping channel than in the traditional channel. As a result, retailers carry a lot of inventory and rely on Drop-shipping for only a small proportion of demand. Again, this nding is consistent with the existence of Internet kiosks in retail stores that already carry a lot of inventory. Overall, our results conform with the empirical ndings of Randall et al. (2006) that no channel option is uniformly preferred over the others and that channel choice depends on companies environments.1.1结论总结我们首先提出了两个可供选择的模式：各零售商自有库存的传统渠道模式和供应商拥有库存并直接发货给零售商指定客户的Drop-shipping渠道模式。我们将这两种渠道看作是由一个供应商对多个零售商的两阶段供应链。在这两种模式下供应商向零售商协定一个固定的批发价格，两种渠道的批发价格并不一定相同。此外，在对互联网公司了解的基础上，我们认为不同的渠道可能由于规模经济和其他成本的差异会导致不同运输渠道运输成本不同。当在传统渠道模式和Drop-shipping渠道模式之间进行选择时，零售商库存减少了风险从而导致批发价格被抬高了，供应商则抬高了批发价格以抵御增加的库存风险。此外，运输成本的差异发挥了作用。结合两个战略，零售商可以先以自有库存（以较低的批发价格购买）满足顾客需求，而其余的需求要求供应商以drop-ship模式满足。例如，大型的互联网零售商BlueL使用双重战略进行CD销售：它为增加边际收入，维持一定数量销量在前四十位的CD的自有库存，而无库存时则通过Drop-shipping模式来供货（见服装行业杂志，2000年）同时采用纯Drop-shipping模式维持CD类目的多样性。此外，一些大型零售商最近在实体店铺安装了互联网服务亭（见“计算机世界”，2001年），如果顾客在商店货架上没有找到他们想要的产品，她可以在互联网商店订购，并且该产品将直接从中央仓库运送至她手上。我们分析的这种模式类似于双重战略模式。双重战略模式有点类似于运营相关文献（见Rudi and Zheng，1997年 和 Cattani等，2000年）中讨论的按库存生产方式和按订单生产方式的结合的模式，这两种模式主要的不同是：后者的库存量和生产量由同一公司管理，而在本文所讨论的双重战略下，零售商和供应商以竞争博弈的方式对库存进行管理。我们将双战略问题作为一个非合作博弈模型，零售商和供应商在市场需求甚至库存分配之间进行竞争博弈。我们分析了这个博弈的均衡点，给出供应商的最佳响应函数，用它们来分析这个对称均衡点的唯一性。我们的结论之一是在双重战略均衡问题中可能在供应商或零售商库存为零使达到最优，从而双渠道转换成一个纯传统渠道或一个纯Drop-shipping渠道模式。我们能够确定维持这种临界均衡点的简单条件。使用超模博弈和最佳响应函数的派生性质，对批发价格、Drop-shipping利润及运输成本参数进行均衡点的灵敏性分析。我们进一步通过数值实验结果比较了传统模式、Drop-shipping模式和双渠道模式。我们发现，当零售商很多，零售商和供应商都能从Drop-shipping或双渠道实现风险共担中获益，这使得这两种模式比传统渠道模式更具吸引力。这一发现和实际观察一致， Drop-shipping经常使用于书籍和电子产品等通常采用由几个供应商为许多零售商提供服务的二级供应链模式的产品。在双渠道模式中我们发现任何费用相关的参数的变化将导致供应商和零售商之间库存的再分配。例如，在这个渠道，增加代发货渠道产品利润增加将导致零售商库存量较高和供应商库存量较低，而在Drop-shipping渠道，相同的行为将导致供应商库存量较高。而奇怪的是，在双渠道模式下运输成本差异对利润的影响并不总是直观的。特别是随着运输成本结构变得更加适用于Drop-shipping模式，供应商的利润开始下降，零售商的利润开始增加。我们发现，Drop-shipping的利润和运输成本差异是进行渠道选择的主要驱动力。我们观察到Drop-shipping和双渠道有同时实现供应商和零售商收益帕累托改进的潜能。但传统的渠道没有这种可能。这一发现表明，大多数Internet零售商应努力利用Drop-shipping至少可以增加其自有库存。然而，在某些行业，供应商更倾向于传统渠道，因此根本不愿以Drop-shipping模式进行库存控制，所以这些企业根本就没有提供这个模式。尤其是Drop-shipping的利润相对较小但运输成本方面更适合于采用Drop-shipping时这种情况会出现，这是因为虽然零售商倾向这个模式，但是供应商直接发货给顾客没有任何利润，供应商是不会愿意采取此模式的。在未知参数集很大的情况下，双渠道实现帕累托改进，这和互联网站的发展和日益普及是密不可分的。由于运输成本不直观的影响，在Drop-shipping的运输成本远远高于传统渠道时，双渠道实现了帕累托改进。因此，用传统零售渠道实现大需求产品的供应，用Drop-shipping实现小需求产品的供应。再次，这一发现和现在网上零售店已实现大量库存的情况是一致的。总的来说，我们的结果符合2006年Randall等人的实证研究即没有任何一个渠道选择永远优于其他的渠道，是否具有优势取决于其所处企业环境。1.2. Relation to the Literature1.2 相关文献Our paper addresses two key issues: (1) inventory ownership and stocking decision rights and (2) risk pooling. From the inventory ownership/stocking decision rights point of view, the traditional channel in our paper is similar to that in most of the supply chain literature because the retailer both owns inventory and decides on stocking quantity (see, for example, Zipkin 2000). In contrast, in the Drop-shipping channel both inventory ownership and stocking decisions rest with the wholesaler. The only related arrangement we are aware of is vendor-managed inventory (see, e.g., Bernstein et al. 2002), in which the wholesaler is endowed with stocking decision rights, but inventory is still owned by the retailer. A combination of vendor-managed inventory with consignment transfers both inventory ownership and stocking decision rights to the wholesaler, a scenario somewhat similar to our Drop-shipping model.The important differences are that under consignment, inventory is stored at each retailers location so there are no benets to risk pooling. Moreover, the structure of transportation costs is different. In the dual-channel model we consider, each echelon of the supply chain stocks inventory, a setting that parallels the multiechelon literature.Typically, though, in this stream of literature all echelons are managed centrally (whereas in our paper retailers and the wholesaler are separate entities) and, moreover, sales in each period are limited by inventory at the lowest (retail) echelon (whereas in our model the wholesalers inventory is used as a backup). Cachon and Zipkin (1999) and Lee and Whang (1999) relax the former assumption by considering a supply chain with one retailer and each echelon locally managed.Bernstein and Federgruen (2003), Cachon (2001), and Chen et al. (2001) further consider decentralized multiechelon supply chains with multiple retailers.Lawson and Porteus (2000) relax the second assumption by adding an option to expedite the product between stages but still assume that inventory is managed centrally.We are not aware of any papers that relax both of these assumptions. The dual strategy is also similar to the practice of subcontracting in case demand exceeds capacity (see Van Mieghem 1999 and Plambeck and Taylor 2005).In another related paper, Cachon (2002) models advance purchase discounts, a practice whereby the retailer can prebook inventory before the season at a lower per-unit price. His work is similar to ours in that both the retailer and the wholesaler are allowed to hold inventory.Cachon focuses on identifying Pareto-optimal price-only contracts and studies supply chain efciency under such contracts. However, because there is only one retailer， the benets of risk pooling are not reected, and issues of inventory allocation among retailers do not arise. Traditionally, research on risk pooling focused on comparing the costs and benets of building one central, versus many regional, warehouses for a single company (see Eppen 1979). Van Mieghem (1998) and Rudi and Zheng (1997) consider dual strategies under risk pooling by combining risk pooling at an extra cost per unit with the cheaper local inventory. All of these papers focus on the performance of a single rm and, therefore, do not compare the benets of risk pooling for both the retailers and the wholesaler, whereas we do. Anupindi and Bassok (1999) compare centralized and decentralized inventory management in a model with two retailers and one supplier while focusing on the effect of market search by customers in case of a stockout at one retailer. They compare the decentralized model with the model in which retailers form an alliance to centralize their inventories. Their work is similar to ours in that they consider the