MANAGEMENT SCIENCE Vol. 51, No. 10, October 2005, pp. 1505–1518 issn 0025-1909  eissn 1526-5501  05  5110  1505 inf orms ® doi 10.1287/mnsc.1050.0371 ©2005 INFORMS The Impact of Duplicate Orders on Demand Estimation and Capacity Investment Mor Armony Stern School of Business, New York University, New York, New York 10012, marmony@stern.nyu.edu Erica L. Plambeck Graduate School of Business, Stanford University, Stanford, California 94305, elp@stanford.edu M otivated by a $2.2 billion inventory write-off by Cisco Systems, we investigate how duplicate orders can lead a manufacturer to err in estimating the demand rate and customers’ sensitivity to delay, and to make faulty decisions about capacity investment. We consider a manufacturer that sells through two distributors. If a customer finds that his distributor is out of stock, then he will sometimes seek to make a purchase from the other distributor; if the latter is also out of stock, the customer will order from both distributors. When his order is filled by one of the distributors, the customer cancels any duplicate orders. Furthermore, the customer cancels all of his outstanding orders after a random period of time. Assuming that the manufacturer is unaware of duplicate orders, we prove that she will overestimate both the demand rate and the cancellation rate. Surprisingly, failure to account for duplicate orders can cause short- term underinvestment in capacity. However, in long-term equilibrium under stable demand conditions the manufacturer overinvests in capacity. Our results suggest that Cisco’s write-off was caused by estimation errors and cannot be blamed entirely on the economic downturn. Finally, we provide some guidance on estimation in the presence of double orders. Key words : maximum-likelihood estimation; duplicate ordering; distribution channels; queueing systems; reneging History : Accepted by William S. Lovejoy, operations and supply chain management; received April 19, 2002. This paper was with the authors 3 1 2 months for 2 revisions. 1. Introduction Amid a general economic downturn, networking titan Cisco Systems experienced a spectacular fall in mar- ket value from $430 billion in March of 2000 to $180 billion in March of 2001. Net income dropped from $0.8 billion in the first quarter of 2001 to −$27 billion in the third quarter of that year as Cisco wrote off $2.2 billion worth of component inven- tory and laid off 8,500 workers (Business Week 2002a). According to the Wall Street Journal, “Cisco executives ignored or misread crucial warning signs that their sales forecasts were too ambitious.” Because of dupli- cate orders, Cisco executives overestimated demand and therefore “continued to expand capacity aggres- sively, even after business slowed” (Thurm 2001a, p. A1). Cisco is certainly not the only technology company to have difficulties in forecasting because of duplicate orders. Intel and other semiconductor man- ufacturers believe that their bookings data is “irrele- vant and potentially misleading” because of duplicate orders (Business Week 2002b, p. 28). This paper shows that even in a stable business environment, a man- ufacturer that fails to account for double orders will carry excess capacity. Cisco’s policy of outsourcing all of its manufac- turing has been lauded in the business press. Less widely recognized is that, since 1998 (when Cisco achieved 65% of its revenues through direct sales), Cisco has sought to outsource sales and distribu- tion. Cisco sells networking hardware to distributors (e.g., Ingram Micro) that sell to systems integrators (including IBM and a host of smaller firms) that in turn sell to Cisco’s end customers and provide ongo- ing support and maintenance. By 2001, the number of Cisco-qualified distributors and resellers (systems integrators) had increased to 20,000 for the United States alone (Kothari 2001). Only 14% of Cisco’s sales were direct; 86% were through channels. Cisco was using the Internet to share real-time informa- tion about inventory and production schedules with its component suppliers and contact manufacturers (Business Week 2001). However, on the demand side, Cisco’s information systems were relatively weak. In particular, Cisco had limited visibility of distributors’ inventory and order backlog (Kothari 2001). In the summer of 2000, Cisco experienced short- ages of several key components. Customers had to wait for two and even three months for some of Cisco’s most popular products. Some frustrated cus- 1505