Factors Influencing User Feedback on Predicted Satisfaction with Software Systems

[Context and motivation] Requirements engineers need feedback from users on planned system features. The simplest way is to present feature descriptions to the users and ask for their opinion. [Problem/question] The feedback users can give in such a situation is not always accurate. The mechanisms which cause a mismatch between actual and predicted user satisfaction are currently not well understood. [Method/results] We used the results from a previous study we conducted, together with insights on consumer satisfaction from marketing, to create a working model of predicted user satisfaction. We validated the model with a new, more extensive empirical study. [Contribution] We present a model of predicted user satisfaction. Unlike the existing models of user satisfaction for software systems, it can be used for gathering feedback before a user has had experience with a software system. Our study shows that measuring predicted satisfaction can deliver a good approximation of actual satisfaction, although there is some prediction discrepancy which could be reduced by choosing the right combination of influence factors.