運用優勢關係約略集合理論探討半導體資本設備競標案之關鍵成功因素

Abstract

最佳競標案的選擇通常涉及複雜且高風險的決策流程。市場的不確定性、資訊的不對稱、以及競爭對手的動態策略對競標專案的成功與否有著決定性的影響。然而對資本設備製造商而言，由於產業特性的關係，每個專案的競標過程往往都必須耗費大量的時間與資源投入，因此錯誤的競標案選擇，對企業的損失是相當巨大的。針對這項議題，本研究以一家國際半導體資本設備製造商為研究對象，運用優勢關係約略集合理論針對該公司過去四年的198個競標案作歷史資料分析，探討其各種決策因素與競標案成功率之間的因果關係。根據不同的競標結果，本研究歸納出數十筆決策規則 (Decision Rules) 供決策者做事前評估參考。研究發現價格的競爭力、顧客的資金狀況、顧客參與專案討論的層級與積極度、技術對顧客的商業影響力、競爭者與顧客的政商關係對多數競標案的成功與否有關鍵性的影響。但是在大多數的狀況下，這些因素的水準並不容易直接被察覺，因此組織高層的投入、業務員的情報蒐集與動態應變能力、以及即時的環境監控在競標案事前與事中階段扮演著舉足輕重的腳色。

The project bidding usually involves a complicated and high-risk decision making process. In the high face of a highly competitive and fast-changing semiconductor capital equipment market, capital equipment suppliers must not only provide high-quality products but react appropriately to changes in customer needs. Many resources in research and development (R&D) process before and after bidding have to be invested on a project. A wrong bidding decision would cause great loss to a capital equipment supplier, either in the perspective of time consuming or money wasting. Unlike the previous studies, we utilized a more intuitive data mining technique called dominance based rough set theory (DRST) for decision rule induction that could provide more straightforward aids for bid/ no bid decision. The 4-year 198 historical data of the sales department of a multinational semiconductor capital equipment supplier were collected. With this technique, the features between 118 successful-bidden projects and 80 unsuccessful-bidden projects in Asia-pacific region were compared and the specific decision rules for the two conditions were generalized. We found that the low level of the price position of the case company and the low level of customer’s funding status leaded to the primary failure of project bidding. The other relatively implicit factors such as the degree of customer’s involvement, the degree of executive sponsorship, influence to customer’s business impact and competitor political leverage influence should be carefully inspected and evaluated in advance.