Why does water meter accuracy matter to utilities?
Water meters are the financial backbone of any water utility. They track every gallon delivered, set every customer bill, and drive the revenue that keeps operations running. They are also expensive assets to buy, install, and maintain. And like all equipment, they degrade.
A 2023 study of about 1.6 million household meters at the Taipei Water Department found meters recorded, on average, 6% to 10% less water than customers actually used.¹ Research from the Utah Water Research Laboratory confirms that mechanical meter accuracy degrades over time, driven by wear, water quality, flow velocity, throughput, and installation.²
What is meter under-registration?
Meter under-registration is when a customer meter records less water than actually passes through it. The unrecorded water is an apparent loss: it reached the customer but was never billed. Apparent losses are one component of non-revenue water.
Unlike a main break, meter accuracy loss is silent. Utilities keep delivering water they are not fully billing for. That unbilled volume quietly erodes revenue, raises apparent water loss, and creates difficult customer conversations when corrections finally come. The question is not whether water meter accuracy matters. It is where the greatest risk sits, and where limited resources should go first.
How confident are utilities in their meter accuracy?
We ran a live poll at the 2026 Smart Water Summit (SWS26) in Westminster, Colorado: 75 participants gave 276 responses to three multi-select questions. The results confirm what many managers already suspect: confidence in meter billing accuracy is shaky, and the tools most utilities use to decide what to fix are not built for the job.
| Question | Key finding | Full breakdown (share of responses) |
| How confident are you that your meters are billing accurately? (n = 87) | Only 17% are very confident and test regularly | 62% somewhat confident, but know there are gaps 18% not confident, suspect losses but can’t quantify them 17% very confident, test regularly 2% no idea, don’t test them |
| Have you measured revenue lost to under-registering meters? (n = 89) | Only 22% track it closely | 60% have a rough sense 22% track it closely 18% no idea 0% not applicable (assume meters are accurate) |
| How does your utility currently decide which meters to replace? (n = 100) | 65% use an age-based schedule | 65% age-based schedule 17% customer complaints 10% data-driven or risk model 5% no formal process 3% random sampling |
Source: VODA.ai live poll, 2026 Smart Water Summit (SWS26), Westminster, Colorado, Aug 31 to Sept 2, 2026. 75 participants, 276 total responses. Multi-select: participants could choose more than one answer. Percentages are shares of all responses per question; totals may not equal 100% due to rounding.
Put simply: most respondents know they have a problem, most don’t know how big it is, and most are still solving it with the bluntest tool available: a calendar.
Why age-based meter replacement falls short
Utilities can’t test or replace every meter. Capital budgets are limited, and field crews are stretched thin. So replacement decisions usually follow meter age, customer complaints, regional replacement schedules, or usage patterns.
These rules are reasonable on the surface, but they are blunt instruments. A meter may be old and still performing well. Another may look normal but be quietly under-registering. Without a more precise way to prioritize meter testing, utilities spend time and budget on the wrong meters and leave revenue at risk elsewhere in the system.
The core question is prioritization: which meters are most likely measuring less water than is actually delivered, and how much revenue is at risk?
“Meter programs do not need more guesswork. They need a smarter way to see which meters are putting revenue at risk.”
Lowell Rust, VP of Implementation, VODA.ai
How does AI find under-registering water meters?
AI replaces broad rules and reactive triggers with a meter-by-meter view of risk. It works from data utilities already hold in meter-reading systems and utility records, such as usage history, meter age, size, and model. It works in four steps:
- Analyze existing meter and billing data across the full meter population.
- Score each meter’s likelihood of accuracy loss.
- Estimate the Volume at Risk and Revenue at Risk for each meter.
- Rank meters so investigation and testing start where the estimated loss is highest.
The result is a ranked list, not a verdict. The model ranks the risk. Your team decides what to test, repair, or replace.
This is the gap our poll surfaced. Only 10% of responses named a risk-based model, which leaves a large opportunity untapped in most meter programs.
What do utilities gain from risk-based meter testing?
The impact goes beyond finding bad meters. When limited capital goes to the meters most likely to be underperforming, utilities can:
- Recover lost revenue sooner by testing and correcting the highest-risk meters first
- Reduce unnecessary spending by keeping well-performing meters in service
- Focus field work on the meters most likely to create financial impact
- Reduce customer disputes by addressing accuracy issues before they lead to large billing corrections
- Support investment decisions with clear estimates of Volume at Risk and Revenue at Risk
This is not about replacing a current meter program. It is about making that program smarter, more defensible, and more financially focused.
The bottom line on water meter accuracy
AI won’t replace the expertise of utility managers or the hands-on work of field crews. It gives them a view across the entire meter population, so they can see where the greatest risks lie and act with confidence, instead of relying on the “rough sense” that 60% of poll responses described.
In an industry where every gallon counts, water meter accuracy is not just an operational issue. It is a revenue protection issue. Using AI to prioritize the right meters for investigation, testing, and replacement helps utilities protect revenue, reduce apparent losses, and decide with confidence when resources are limited.
Learn how VODA.ai Meters ranks meters by Volume at Risk.
This article is part of our AI for Utilities series, where we break down how AI and machine learning can transform water asset management. From risk prediction to proactive prevention, we cut through the hype to share what really works.
🔔 Subscribe to our blog so you don’t miss the next article in the series.
Frequently asked questions about water meter accuracy
What causes water meter accuracy to decline?
Mechanical water meters lose accuracy over time because of wear, water quality, flow velocity, total throughput, and how the meter was installed and handled. Most degraded meters under-register, which means they record less water than the customer actually uses. The loss is gradual and silent, so it rarely shows up in billing data on its own.
Meter accuracy loss happens quietly. A meter may appear normal while silently under-registering. Without precise visibility into which meters are failing, utilities can’t prioritize where to focus limited resources. Traditional approaches, age-based replacement, customer complaints, regional schedules, are blunt instruments. They often result in replacing meters still performing well while missing ones losing revenue silently.
Is meter under-registration part of non-revenue water?
Yes. Water that reaches a customer but is not recorded by the meter is an apparent loss, and apparent losses are one component of non-revenue water. Unlike leaks, the water is not physically lost. It is delivered but never billed, which makes customer metering inaccuracy a direct revenue issue for the utility.
Why is age-based meter replacement not enough?
Meter age is a weak proxy for meter accuracy. An old meter can still measure well, while a newer meter can under-register because of its size, model, usage pattern, or water quality. A replacement schedule based on age alone can spend budget on healthy meters and miss the ones putting the most revenue at risk.
What is Volume at Risk for a water meter?
Volume at Risk is an estimate of how much water a meter is likely delivering but not recording. Combined with billing rates, it becomes Revenue at Risk. Ranking meters by these two values shows a utility which meters to investigate and test first, instead of treating every meter of the same age the same way.
Does AI replace water meter testing?
No. AI does not test meters or confirm their accuracy. It analyzes existing meter and billing data to rank which meters are most likely under-registering. Field crews still test, repair, or replace the meters. The benefit is focus: testing time and budget go to the meters with the highest estimated Volume at Risk.
Last updated: October 2, 2026
Sources
¹ Chen, H.-L., Lo, S.-L., Kuo, J., et al. (2023). Estimate measurement errors of household water meters using a large amount of on-site data feedback. Sustainable Environment Research, 33, 19. https://doi.org/10.1186/s42834-023-00180-z
² Stoker, D.M., Barfuss, S.L., & Johnson, M.C. (2012). Flow measurement accuracies of in-service residential water meters. Journal AWWA, 104(12), E637-E642. https://doi.org/10.5942/jawwa.2012.104.0145
³ VODA.ai live poll, 2026 Smart Water Summit (SWS26), Westminster, Colorado, Aug 31 to Sept 2, 2026. 75 participants, 276 responses, multi-select (Q1 n = 87, Q2 n = 89, Q3 n = 100).
