PurposeThe purpose of this study is to examine the satisfaction formation process under conditions of varying involvement for new, nontraditional, credence‐based service offerings.Design/methodology/approachHypotheses were tested using an on‐line simulation of a service encounter, beginning with perceptions of an advertisement and continuing through satisfaction with the completed service report. ElderCare, an assurance service provided by Certified Public Accountants to the children of an elderly parent, was the context of study.FindingsThe findings indicated expectations were not influential in the satisfaction formation process for these services, regardless of involvement. Under conditions of high involvement, performance evaluation was the dominant predictor of satisfaction. Low‐involvement subjects used disconfirmation to assess satisfaction.Research limitations/implicationsLimitations of the research stem from the use of a convenience sample in a simulated service encounter. Future research should examine the influence of individual characteristics, particularly innovation, on the results. Additional research examining the determinants of model variables, from disconfirmation to involvement, across subjects in a variety of situations would also be valuable.Practical implicationsThis research suggests a heightened need to guide the consumer experience where expectations are hard to develop and performance is hard to evaluate. Providers should carefully spell out key service dimensions, provide tangible information about performance outcomes for high‐involvement individuals, and focus on emotional appeals, provider qualities and sensory aspects of the service for low‐involvement individuals.Originality/valueThis paper helps researchers and practitioners better understand the applicability of expectancy disconfirmation theory and role of involvement in the context of new, nontraditional, credence‐based services.