How AMI + SCADA Integration Finally Lets You Measure Non-Revenue Water in Real Time

The American Water Works Association's most recent benchmark puts utility-wide non-revenue water at 16-18% on average, with stressed systems running well above 30%. That's water you produced, treated, pumped, and never billed for. It's the single largest avoidable line item on most utility P&Ls.

Most utilities measure it quarterly. With paper reconciliation. Months after the loss has already happened.

If you have AMI meters in the field and a SCADA system at the plant, you already have everything you need to measure NRW daily — by district, by class, by anomaly. The data is sitting there. Most utilities just don't have a layer above it that connects the two.

This is what changes when they do.

The math hasn't changed. The latency has.

NRW is conceptually simple:

NRW % = (Water Produced − Water Billed) / Water Produced × 100

Production data lives in SCADA — flow meters at the treatment plant, at booster pumps, at zone inlets. Billed consumption lives in your AMI head-end or billing system — every meter, every account.

For decades, the gap between those two systems was operational, not technical. SCADA was the plant operator's tool. AMI lived in customer service. Nobody owned the join. So NRW got calculated once a quarter by hand, with the production volume from operations and the billing volume from finance, and the gap was reported to council as "system loss."

By the time you saw a spike, the leak had been running for 60-90 days.

What "real-time" actually unlocks

When you join SCADA production data and AMI billed consumption on a daily cadence — at a district level, not just system-wide — three things become possible.

1. You see leaks before quarterly meter reading does

A continuous flow of 5 gpm at a single household is a $40/month leak — annoying for the resident, invisible at the system level. But fifty of those in one district is $2,000/month of lost water, and it shows up as a 4-point bump in district NRW within 48 hours.

With real-time NRW by district, the leak signature is unmistakable. Without it, you wait until the quarterly reconciliation, by which point you've already burned the loss for two billing cycles.

2. You separate real loss from meter error

Apparent loss — under-registering meters, billing system mismatches, theft — accounts for roughly 30-40% of total NRW in most utilities. Real loss is the leak in the ground.

The shape of the gap tells you which is which. A district that runs steady at 18% NRW month over month is showing apparent loss — almost certainly meter aging. A district that jumps from 12% to 24% in two weeks is showing real loss — a new leak, almost certainly mainline.

If your only NRW metric is system-wide quarterly, you can't tell the difference. The blended number tells you nothing actionable.

3. You can prove improvement to rate-setters

State utility commissions and council rate hearings increasingly ask the same question: "What are you doing about your water loss?"

A utility that can show district-level NRW trending from 22% to 14% over two quarters has a real story to tell. A utility that reports a single annual NRW percentage has nothing to defend with.

The dashboard becomes the artifact. Real-time means the regulator (and council) can see it whenever they ask.

Why your AMI vendor's portal doesn't do this

Most AMI head-end software is designed to do one thing: ingest meter reads and pass them to billing. It might give you a meter status view, maybe a high-consumption alert. But it doesn't pull production data from SCADA, and it doesn't know what "billed" means — that lives in your CIS.

Same for SCADA. SCADA shows you the production side perfectly. It has no idea what got billed downstream.

The gap is the whole point. NRW lives in the join between two systems that don't talk to each other. Without a layer that pulls both, you can't measure NRW operationally. You can only measure it after the fact, on a spreadsheet.

What district-level NRW looks like in practice

Real-time NRW measurement is most useful when it's broken down by district. System-wide is too blurry — a 6-point spike in one district shows up as a 1-point spike in the citywide number, easy to miss. District-level, the same event lights up like a flare.

A typical utility we work with looks something like this on day one:

• North District — 12% NRW. Newer infrastructure, mostly single-family, low irrigation. Healthy.
• Downtown — 19% NRW. Older mains, commercial mix, more meter aging. Trending the wrong direction over the last 90 days.
• Southside — 31% NRW. Old pipes, high seasonal irrigation, one or two persistent mainline leaks suspected.

The Southside number is what makes the leak-detection program pay for itself. You don't go chase NRW system-wide. You send the crew to the district where the gap is widening.

The top-anomaly list — turning the AMI signal into a daily task

A real-time NRW dashboard is only useful if it tells you what to do next. The actionable output isn't the percentage. It's the top 20 accounts driving outsized consumption right now.

The math is straightforward:

• Pull each account's 30-day average consumption (baseline)
• Compare to last 24 hours
• Rank by absolute and relative change
• Classify the likely cause: continuous flow (leak), irrigation overrun, new tenant, commercial change, meter fault

What lands on your operations team's screen every morning is a list of 15-20 service addresses — ranked by likely dollar impact — with a one-line cause hypothesis. That's not a dashboard. That's a work order.

Most leak-detection programs spend 80% of their time looking for leaks. With the top-anomaly list, you spend 80% of your time fixing them.

What this stack costs vs. what NRW costs

A reasonable mid-size utility — say, 15,000 service connections, 4 districts, average production of 4 MGD — runs roughly the following math:

• Average NRW around 18% = ~720,000 gallons/day unaccounted
• At $4 per 1,000 gallons fully-loaded cost (treatment + pumping + chemicals + amortized infrastructure), that's $2,880/day of cost on water you'll never bill for
• Annualized: just over $1M per year of NRW cost

A real-time NRW analytics layer — including SCADA integration, the district-level dashboard, and the daily anomaly list — typically runs $25-50K/year for a utility that size. Most pilots see NRW drop 3-5 percentage points within two quarters, just from the visibility itself plus the directed leak-detection that the anomaly list enables.

That's a 5-8x return in year one, and the savings compound every year after.

What to ask any vendor selling you "AMI analytics"

The AMI analytics market is full of products that show you pretty charts of consumption data. Most of them are dashboards built on top of your billing system. They don't connect to SCADA. They don't compute NRW. They can't.

Three questions to filter the real platforms from the dashboards-in-a-coat:

1. "Do you ingest production flow from our SCADA system?" If the answer is no, you don't have NRW measurement. You have a consumption dashboard.

2. "Can you show me NRW broken down by district, day by day?" If the granularity is monthly or system-wide, the data isn't actionable.

3. "What does the operations team look at every morning?" If it's a chart, the answer is wrong. The answer should be a ranked list of service addresses driving today's loss.

NRW isn't a reporting problem. It's a daily operations problem. The tooling has to match.

---

Fluvio combines your AMI consumption data with SCADA production flow to give you real-time NRW visibility — district by district, anomaly by anomaly. See the analytics dashboard.