Every utility runs an annual water audit. And at some point, almost every utility finds the same thing: there is more water going into the distribution system than is showing up in customer billing. The gap has a name. It also has causes, and not all of them are obvious.
The industry calls this non-revenue water, or NRW for short. It has two components. The first is real losses: actual physical leaks where treated water escapes the system before it reaches anyone. The second is apparent losses: water that reaches customers and gets used, but is not accurately measured at the meter and therefore never billed. Meter underregistration, where a meter reads less water than actually passed through it, is one of the most common drivers of apparent losses and one of the most frequently overlooked.
Most utilities know they have an NRW problem. Fewer know how much of it is coming from their meter fleet versus their distribution infrastructure. That distinction matters a lot when you are deciding where to spend capital improvement dollars.
The Problem With Age-Based Replacement Schedules
The standard approach to managing meter accuracy is straightforward: set a replacement schedule based on age, usually somewhere between 15 and 25 years, and replace meters on a rolling cycle. It is clean, defensible, and easy to budget.
It is also, according to published research, an unreliable predictor of which meters actually need replacing.
A study published in the Journal AWWA, the peer-reviewed publication of the American Water Works Association (AWWA), tested 573 in-service residential meters pulled from utilities across the country. The researchers were specifically investigating water meter accuracy degradation: whether age, wear, or cumulative throughput could reliably predict when a meter stopped performing within AWWA standards. The answer was that none of those factors correlated consistently with accuracy. Some meters had failed almost immediately after installation. Others were still performing within standards after more than 30 years of service.
"A meter replacement program based solely on time may end up replacing some meters prematurely."
— Stoker, Barfuss & Johnson, Journal AWWA, 2012
What the study did find is that accuracy degradation is driven by a combination of factors: wear, water quality, debris in the distribution system, cumulative throughput, and installation conditions. These vary by system, by meter type, and by individual meter. They do not follow a predictable calendar.
A 20-year replacement schedule might be pulling meters that are still accurate while leaving genuinely underperforming meters in the ground because they are only 14 years old. Without testing, there is no way to know which is which.
Where the Revenue Is Actually Going
The study identified one pattern worth paying attention to. The most common residential meter type, the oscillating piston meter in which a piston moves in a circular path inside a chamber as water flows through, showed the weakest accuracy at low flow rates. Specifically at the minimum flow threshold, which represents slow drips, trickle flows, and small leaks on the customer side.
Low-flow underregistration is where a significant share of apparent loss accumulates in residential systems. A meter that reads correctly under normal household use may be missing a meaningful percentage of flow at low rates, and that gap rarely triggers anything in your billing system. Consumption looks plausible. The bills go out. The unbilled water shows up months later in your audit numbers.
For utilities with large residential fleets, even modest underregistration across thousands of meters represents material unbilled revenue. Unlike a main break, there is no visible event that flags it. The only way to surface it is through systematic water meter testing or a detailed audit.
What Water Meter Testing Actually Tells You
The Utah Water Research Laboratory study recommends that utilities establish a meter testing program: periodic bench testing of a subset of meters, conducted alongside annual water audits, to monitor water meter accuracy degradation in real time. Bench testing means physically removing meters from service and running known volumes of water through them at controlled flow rates to measure how accurately they register. It is the only definitive way to know whether a specific meter is performing within AWWA standards.
What bench testing produces is not just a pass/fail result for individual meters. Done systematically on a sample across different age groups, meter types, and service areas, it tells you which segments of your fleet are underperforming and by how much. That is the data you need to make a replacement decision based on actual performance rather than assumptions about age.
AWWA's Manual of Water Supply Practices M6, which covers meter selection, installation, testing, and maintenance, recommends periodic bench testing of residential meters specifically to identify groups with accuracy degradation issues. The intent is that testing drives replacement decisions rather than age schedules driving them in the absence of performance data.
The practical problem for most utilities is capacity. Running a meaningful bench testing program requires lab equipment, trained staff, and a process for pulling, shipping, testing, and reinstalling meters at scale. Most small and mid-sized systems do not have that infrastructure. So they default to age-based schedules built on rules of thumb, which the research shows may not accurately predict which meters actually need replacing.
The result is a fleet of unknown water meter accuracy managed by assumption. Some problem meters stay in the ground because they are only 15 years old. Others get replaced at 20 years even though they are still performing accurately. The gap between what you think your fleet is doing and what it is actually doing shows up as apparent loss in your audit.
What We See Before Every Program
We have run meter changeout programs across the country. Before almost every one, the conversation with the utility starts the same way: they tell us how many meters they have, what sizes, and roughly how old they are. And then we do a line survey and find out the actual answer.
A line survey is a field verification phase conducted before installation begins. A dedicated two-person crew works address by address to verify meter size, location, access conditions, and pipe material on both the utility-side and customer-side of each connection. That pipe material documentation, delivered with photos and GPS coordinates in a utility-ready format, also satisfies a requirement most utilities are already managing under the federal Lead and Copper Rule Revisions (LCRR): maintaining an inventory of service line materials across their entire system. Running it as part of a pre-installation survey means you are not paying for that data collection twice.
What we find, consistently, is that the utility's database does not match the field. Not because anyone made a mistake, but because distribution systems change over decades and records do not always keep up. Meters get swapped out during service calls and the size changes. Addresses get renumbered. Boxes get buried. The inventory drifts.
In our experience, address database errors are the single biggest source of schedule variance in a changeout program. A utility that goes into a program with 5,000 meters on the books may have 5,300 in the ground, or 4,800, or a different mix of sizes than expected. Any of those scenarios affects the budget, the timeline, and the accuracy of the NRW case you built to justify the program in the first place.
Building a Case Your Finance Director Will Accept
If your water audit is showing non-revenue water losses and you suspect the meter fleet is contributing, the internal question is always the same: how do you make the case for a capital program when you cannot prove exactly how much the meters are costing you?
The answer is to replace the assumption with data. A line survey gives you a verified inventory. That inventory tells you the actual age distribution of your fleet, the mix of meter types, and the condition of the infrastructure around each meter. Combined with your audit data, that is enough to build a meaningful estimate of apparent losses and a defensible cost-benefit analysis for the replacement program.
AWWA's Manual of Water Supply Practices M36, which covers water audits and loss control, recommends that utilities conduct annual audits and use the results to inform meter replacement decisions rather than relying on age schedules alone. If you are going to a city council or a rate authority with a capital request, having the AWWA framework behind your methodology matters.
What You Should Get at the End of the Program
Replacing the meters is the visible part. What the program actually delivers, if it is run correctly, is a documented fleet you can trust.
Every meter installation should produce a verified record tied to a specific address: the old meter's final read, the new meter's serial number, the meter interface unit (MIU) ID for the radio or endpoint that transmits consumption data to your network, GPS coordinates, and pipe material documentation. That data should come back in a format that uploads directly into your billing system and your Advanced Metering Infrastructure (AMI) platform without manual re-entry.
Why does that matter for NRW? Because a program that installs accurate meters but delivers messy data has not solved the billing problem. An MIU ID recorded incorrectly means consumption data from that address is not reaching the right account. A meter installed in the wrong flow direction underregisters from day one. Water meter accuracy after a changeout program is only as good as the quality control behind the installation.
We maintain a sub-2% quality assurance and quality control (QA/QC) error rate across all programs, meaning fewer than 2 in 100 installations require any correction. That matters not because it is a badge, but because every error that makes it past installation is an apparent loss your next audit will find.
Not sure where your NRW is coming from?
A line survey is the right first step before committing to a full program. It tells you what is actually in the ground, gives you the data to build an accurate scope, and usually surfaces the information you need to make the internal case. We are happy to walk through what that looks like for a fleet your size.