Non-revenue water — the gap between water produced and water billed — gets attributed almost automatically to leaks and theft. Both are real contributors, but utilities that go looking only for physical losses often miss a third source hiding in the same numbers: systematic meter under-registration.
Large meters lose accuracy asymmetrically. Positive-displacement and turbine meters both tend to under-register at low flows well before they show any measurable error at design flow — which means a meter can pass a standard accuracy test at nominal conditions while quietly under-billing every low-flow period, the exact conditions most residential and light-commercial accounts actually produce.
That asymmetry compounds at the system level. A utility auditing individual meters one at a time can find each one 'within tolerance' at the flow rate it happened to be tested at, while the aggregate low-flow under-registration across thousands of meters adds up to a non-trivial share of reported non-revenue water — the loss shows up in the water balance, but it was never actually lost.
Distinguishing measurement-driven non-revenue water from true physical loss matters because the fixes are completely different: leak detection programs and pressure management address physical loss; meter replacement, testing frequency, and multi-point accuracy verification address the measurement component. Applying the wrong fix to the wrong source wastes budget without moving the number.
Multi-point testing under AWWA M6 guidance, tracked continuously rather than as an annual spot check, is what separates the two — and turns a chunk of 'non-revenue water' back into an accurate, billable number.