Quality function deployment (QFD) has been adopted to improve product quality and development in many fields. Numerous studies have demonstrated that attribute importance and attribute performance have a causal relationship and the customer self-stated raw importance is not the actual importance of a customer attribute. These findings generate questions regarding the applicability of the conventional planning matrix (PM) of QFD. This study presents a revised PM that integrates a back-propagation neural network and three-factor theory to assist practitioners in determining the actual importance of customer attributes. An illustrative case demonstrates the effectiveness of the revised PM and identifies the shortcomings generated when applying the conventional PM.