Understanding elements of non-revenue water (NRW) has become more critical in an era where urban water has a rapidly increasing value. Meter under- and over-registration of flows has received considerable research attention over the years, but there has been limited focus on non-registration, particularly for domestic water meters over their life cycle. This paper presents empirical estimates on non-registration for typically installed domestic water meters and emphasises that both under- and non-registration should be considered for any meter replacement assessments and associated policies. Economic modelling showed that domestic water meters should be replaced at lower throughput volumes when considering both of these meter inaccuracy categories. The key value of the paper is that it provides a methodology for determining nonregistration for a fleet of domestic meters.