In France, the performance indicators applied to drinking water supply systems tend to be serviceoriented, making no distinction between residential and non-residential users. In this paper, we seek to test our working hypothesis that these different groups of consumers each have their own sets of expectations, constraints, and vulnerabilities and would thus constitute distinct actors in case of a service failure. Three water utilities located in southwestern France serve as a case study. Results show that non-residential users’ perceptions of service performance can differ significantly from those of residential consumers. Our findings indicate that non-residential users tend to focus more intensely on certain subjects, i.e. the balance of remaining comfortable while not wasting time, trade-offs between restrictions and profitability, etc. Furthermore, non-residential users do not form a homogeneous category. Within non-residential users, three rationales can be distinguished: 'productive', which relates to users who are highly dependant on the current model of drinking water supply; 'routine', in which use of water from the tap seems to continue out of habit, convenience, and/or safety reasons; and 'economic optimization' or 'moderation'. This additional performance-related knowledge could prove invaluable in designing effective strategies for water infrastructure asset management as it allows utilities to prioritise sectors for improvement and be more efficient. It helps utilities better serve their customers by addressing their specific needs. It also helps target communication on less familiar or understood topics. Finally, our work contributes to the debate on management through indicators as it questions their meaning and scope.