The revenue you already earned: finding leakage in meter-to-cash

Every utility is quietly leaking revenue it has already earned. Not through bad pricing or weak sales — through the gap between energy delivered and money collected. A soft shutoff that never got enforced. A tampered or defective meter. Usage that was consumed but never billed. Trailing debt that walked out the door under a new name. Individually small; in aggregate, a margin line hiding in plain sight.

The reason it persists isn’t that the anomalies are invisible. It’s that they’re scattered across the meter-to-cash process — metering, billing, service orders, inspections, collections — and no single system owns the whole chain. Detection is an integration problem first and a modeling problem second.

Detection is the easy half; attribution is the point

It’s tempting to frame this as “build a fraud model.” But a model that says leakage is happening isn’t worth much on its own. What a revenue-assurance team can act on is attribution — leakage broken down by the dimension that tells you what to do:

• by cause — soft shutoffs vs. defective equipment vs. unbilled usage vs. inaccurate billing
• by customer, account, and service agreement — so recovery is targetable
• by rate, tariff, and usage pattern — so the systemic cases surface, not just the one-offs
• by stage of meter-to-cash — so you fix the process, not just chase the symptom

The anomaly detection tells you that revenue is escaping. The attribution tells you where the pipe is leaking — and which crew, work order, or billing rule closes it.

Why “near real time” changes the economics

Catch unbilled usage in the next cycle and you recover it. Catch it six months later and you’re writing it off. The value of leakage analytics is heavily front-loaded in time, which is why low-latency monitoring of the meter-to-cash process — flagging abnormal usage as it appears rather than at audit — is where the recoverable dollars actually live.

None of this needs a moonshot model. It needs the meter-to-cash chain wired together, each anomaly attributed to a cause and an owner, and a latency short enough to recover the dollars while they’re still recoverable. Do that and revenue assurance stops being an annual audit and becomes a standing line of defense — the cheapest margin a utility can find, because it was already earned.