Parasuraman. Zeithaml, and Berry (1988) have proposed the use of a 22-item SERVQUAL instrument for the measurement of perceived service quality. Since their important work, several authors have criticized the use of this instrument in applied settings suggesting that the number and type of dimensions may vary by service category, that there are problems in attempting to use the same wording, across different service categories and in dealing with services that provide multiple service functions (e.g., hospitals), that the analysis of difference scores between perceptions and expectations raises questions about the psychometric properties of such a scale, and that the SERVQUAL instrument confounds the measurement of service satisfaction with service quality. We present an alternative measurement scheme for the measurement of perceived service quality, based on conjoint analysis, that can be easily modified to any service category by expectancy confirmation/disconfirmation response. The advantages of the proposed procedure are, first, that we measure true perceptions, as opposed to perceptions confounded with expectations and satisfaction; second, that the number, type, and operationalization of the specific dimensions (vis-a-vis the wording) are completely flexible according to the specific usage scenario; third, that estimation can be performed in an efficient manner utilizing orthogonal designs and simple OLS; and last, that the proposed model can lead to interesting quality optimization models as welt as models"that explore segmentation. The proposed methodology is illustrated with service quality perceptions of banks and dental offices. We conclude by discussing directions for future research.