This paper examines the impact of variability of product expectations on customer satisfaction within an expected utility framework. Typical models for customer satisfaction use a disconfirmations paradigm based on the gap between mean expectations of product performance and actual product performance. This may not adequately explain satisfaction if expectations are distributional. Actual performance within the range of expectations may not engender any satisfaction response while actual performance outside a range may disproportionately impact on satisfaction. Additionally, satisfaction may be impacted by the alternatives available. Using a designed choice experiment, simultaneously the impact on satisfaction of variability of expectations and the impact of characteristics of alternatives are examined. Results suggest both are relevant to the measurement of customer satisfaction.