LCD 顯示器保固費用預測模型: 一家 ODM 公司個案研究

產品或售後服務的價格，是客戶是否購買產品或服務首要的考量項目之一。IT OEM/ODM產業是一個非常競爭的環境，其淨獲利是相當的微薄。當客戶需要為產品買保固時，OEM/ODM能否提供一個具有競爭力的價格就顯得非常重要，其中價格包含產品物料價格以及平均單台產品保固費用等。本論文研究是與LCD顯示器保固策略相關，其四個目的如下: (一) 了解LCD顯示器OEM/ODM與客戶間一般的保固策略，(二) 對一般的保固策略與顧客興趣、期望做比較， (三) 從客戶的角度去看保固策略的趨勢， (四) 依不同區域建立保固服務費用的預測模型，以及其運用。本篇論文中提出三個LCD顯示器的保固服務費用預測模型，是經由分析一家ODM 公司超過四年的保固費用申請(或維修)資料所建立。三個保固服務費用預測模型是同樣運用多項式迴歸所建立，可以解釋54個月的累計保固返修的百分比的模型，這些模型也可以用來預測保固返修的數量。研究發現產品保固的總費用是與產品的返修數量有非常高的正相關。因為產品的使用時間資料收集不易，保固返修的時間是從產品的製造月份算起，直到產品返修的月份。另外，因為客戶不同和消費者行為的不同，LCD顯示器的保固申請資料得依歐洲、亞洲及美國等三區作分析。在個案研究中會提及T-公司所遭遇的保固相關議題及解決方法。

關鍵字

保固費用；預測模型；多項式回歸

並列摘要

Price of product or after sales service is always one of the top considerations for customer to buy the product or service. IT OEM/ODM industry is a very competitive environment that the net profit is extremely marginal. It is, therefore, crucial for OEM/ODM to offer a competitive price, including bill of material (BOM) price and average warranty cost per product, when customer wants to buy warranty from OEM/ODM. Efficient cost management on not only BOM price, but also average warranty cost per product is a main task for OEM/ODM. There are four purposes of the paper as follows: 1) to realize common warranty policies between OEMs/ODMs of LCD monitor and their buyers, 2) to compare the common warranty policies with customers’ interests whether they meet customers’ needs or expectations, 3) to find out a trend of change regarding to warranty policy from buyers’ perspective, 4) to build prediction models for warranty servicing costs per region, and their applications. Three prediction models are demonstrated for warranty service costs on LCD monitors based on analysis of warranty claim (repair) data over more than four years collected by an ODM company, said T-company. Polynomial regression on 54 months’ cumulative percentages of warranty returns, three prediction models are generated. These models can be applied for volume prediction of warranty returns. Total warranty costs are highly correlative to volume of warranty returns. Time period of warranty return is counted by the difference of returning and manufacturing month of product due to time usage data of product is difficult to collect. Warranty claim data of LCD monitors are analyzed to find return patterns for Europe, Asia Pacific, and United States of America when return patterns are different per different customer account and consumer behaviors in each region. Warranty related problems that T-company encountered are described with resolutions in a case study.

並列關鍵字

Warranty Cost ； Prediction Model ； Polynomial Regression

參考文獻

1. Akerlof, G. A., 1970, “The market for lemons: Quality, uncertainty and the market mechanism,” Quarterly Journal of Economics, Vol. 84, No. 3, pp. 488-500.

2. Attardi, L., Guida, M., and Pulcini, G., 2005, “A mixed-Weibull regression model for the analysis of automotive warranty data,” Reliability Engineering & System Safety, Vol. 87, No. 2, pp. 265-273.

3. Balcer, Yves and Sahin, Izzet, 1986, “Replacement costs under warranty: Cost moments and time variability,” Operations Research, Vol. 34, No.4, pp. 554-559.

4. Ben-Daya, M. and Noman, S.A., 2006, “Lot sizing, preventive maintenance, and warranty decisions for imperfect production systems,” Journal of Quality in Maintenance Engineering, Vol. 12, No. 1, pp. 68-80.

6. Dewally, M., and Ederington, L., 2006, “Reputation, certification, warranties, and information as remedies for seller-buyer information asymmetries: Lessons form the online comic book mark,” The Journal of Business, Vol. 79, pp. 693-729.

國際替代計量

LCD 顯示器保固費用預測模型: 一家 ODM 公司個案研究

全文下載

主題瀏覽