電子產品是由許多不同類別零組件組裝而成。多數電子產品製造商對不同類別零組件的供應商，並不分類評估其績效，也不瞭解不同分類的供應商評量模式對績效評量的影響。另外，雖然多數電子產品製造商有其供應商評量模式，但因非常（甚至完全）仰賴人類專家，而常有穩定性與公正性不佳的問題。本研究以台灣電源供應器製造商與其供應商為研究對象，對本研究所提出之供應商評量模式進行檢驗，並探討不同分類程度的供應商評量模式對績效評量效果的影響。本研究提出一套結合平衡計分卡（Balanced Scorecard，BSC）與Kanji 企業計分卡（Kanji Business Scorecard，KBS）的供應商評量模式（吾人稱之為「BSC 評量模式」），以解決電源供應器製造商非常（甚至完全）仰賴人類專家的缺點，並補救傳統評量模式不足之處。本研究的結果發現：(1)將供應商分為四個次分類之「BSC 評量模式」，其評量績效優於將供應商分為
兩個分類之「BSC 評量模式」，且更優於不對供應商分類之通用型「BSC 評量模式」；(2)次分類之「BSC 評量模式」評量績效不僅是最優的，也最接近人類專家的評量績效，此結果也證明本研究所發展的「BSC 評量模式」之正確性與有效性；(3)最後，本研究也發現採用與供應商之所屬分類別相應之「BSC 評量模式」，其評量效果是優於非相應之「BSC 評量模式」。

Power supplies are made up of many components that are supplied by many suppliers. Thus, the success of a power supply manufacturer replies very much on the collaboration and cooperation of its suppliers. As a result, one important issue that arises is how to select good suppliers that a power supply manufacturer can depend on. In practice, many power
supply manufacturers have their own evaluation models or methods for this purpose.However, their supplier-evaluation models or methods may not work very well for variousreasons, e.g. relying too much on human experts. Furthermore, many power supplymanufacturers often apply one evaluation model or method for all categories of suppliers.In other words, the possible benefits of applying different evaluation models or methods to
different categories of suppliers have never been investigated. In this paper, we propose asupplier evaluation model that combines the structure and dimensions of BalancedScorecard (BSC) and the concept of Kanji Business Scorecard (KBS) and we adopt theAHP method to help us build up the evaluation model. We then test the model’s validity and performance on Taiwanese power supply manufacturers. The test result shows that the
proposed evaluation model can evaluate suppliers very correctly. We also find that better evaluation results can be obtained if one categorizes the suppliers into smaller categories,then applies evaluation models specifically for these categories. And, the smaller the categories, the better the evaluation results of the evaluation models.