Event-level monitoring of water-saving interventions in a living lab: device effectiveness and user experience

Journal article, 2026

Most field studies of residential water-saving devices rely on aggregate meter data that cannot isolate device-level effects. We present a tap-level monitoring framework applied at HSB Living Lab, a 29-apartment facility in Gothenburg, Sweden. Over 32 months, 205,097 water use events were recorded across 80 taps (31 treated, 49 control) for nine water-saving products installed in three rounds. Event-level non-parametric tests, a difference-in-differences (DiD) framework, and three user experience surveys (n = 26) were combined. Event-level tests found that 65–81% of treated taps showed statistically significant shifts in the distribution of individual use events (median volume change −7.1%); because an event here reflects a ∼10-min reporting window rather than a fully independent discrete use, we treat this as exploratory description rather than confirmatory evidence. The aggregate, tap-clustered difference-in-differences estimate—our confirmatory analysis—was not significant. Product-specific effects were heterogeneous; in this small, exploratory survey (n = 26 respondents, 3–10 per product), the least-liked product showed the largest measured reduction and the most-liked showed none, a suggestive pattern that would need confirmation in larger samples. Event-level monitoring can surface intervention effects that aggregate data may not resolve; we interpret this as a consequence of finer measurement resolution rather than evidence of savings hidden within aggregate figures. We emphasise, however, that statistically significant event-level shifts indicate changes in the distribution of individual water-use events, not necessarily reductions in total water use at the building scale.