This study explores dissatisfaction and neutrality metrics from 12 years of a national-level undergraduate student survey. The notion of dissatisfaction is much less prevalent in the narratives surrounding student survey outcomes, and the underpinning metrics are seldom considered. This is despite an increasingly vociferous debate about "value for money" of higher education and the positioning of students as consumers in a marketised sector. We used machine learning methods to explore over 2.7 million national survey outcomes from 154 institutions to describe year-on-year stability in the survey items that best predicted dissatisfaction and neutrality, together with their similarity to known metric predictors of satisfaction. The widely publicised annual increases in student 'satisfaction' are shown to be the result of complex reductions in the proportions of disagreement and neutrality across different survey dimensions. Due to the widespread use of survey metrics in university league tables, we create an anonymised, illustrative table to demonstrate how UK institutional rankings would have differed if dissatisfaction metrics had been the preferred focus for reporting. We conclude by debating the tensions of balancing the provision of valuable information about dissatisfaction, with perpetuating negative impacts that derive from this important subset of the survey population.