Component

Normalized consumption panels

A utility's billing extract turned into one row per account and period, on a consistent class and meter-size vocabulary, with every period that could not be relied on flagged and the reason for the flag kept beside it.

What it is

A billing extract, as it comes out of a utility’s system, put onto one shape: one row per account per billing period, with the class code, the meter size, the days in the period and the consumption on it. Account classes are mapped onto a single vocabulary so that a commercial account in one export and a commercial account in another are counted as the same thing.

The part that costs a study its first two weeks is not the arithmetic. It is finding out which periods are complete, which class codes mean what, which meters were changed mid-period, and which readings cannot be true. That work happens once here and is written down.

What you hand us

The billing extract. It does not need to be clean, and it does not need to be in any particular format — one file per year, one file per cycle, a report somebody exports by hand, or a database dump. If the utility can only produce it as a PDF, say so early, because that changes what the work is.

We do not connect to a billing system. This is a file you send and a file you get back.

What you get back

The panel, and a reason column.

Every period that could not be relied on stays in the panel and is flagged, with the reason recorded on the row: the period was partial, the class code did not map to anything, the consumption is not physically possible for that meter, or the meter was changed part-way through. Nothing is silently dropped. A panel with a hole in it and no note beside the hole cannot be checked by the person who has to defend the study built on it.

What gets excluded from the study is your decision, not ours. The panel gives you the flag and the reason; which flags you exclude, and what you say about it in the study, is yours.

What it is built from

The utility’s own records, and nothing else. This is not built from the public corpus, no figure in it comes from a regulator’s file, and none of it is published anywhere by us.

What it does not do

It does not correct a reading. Where a consumption figure cannot be true, it is flagged as implausible and left as filed. Replacing it with an estimate would put a number in the panel that no meter produced.

It does not tell you what a class ought to be. Where a class code is not in the utility’s own documentation, it is flagged as unmapped and left for somebody who can ask.

It does not reconstruct a period that was never billed. A gap in the billing record is reported as a gap.

It is not a water balance and it is not an audit. It is the consumption side, put in order, so a cost-of-service allocation can start.

Start a conversation

A short call is usually enough to say whether we are the right people for it.

Email [email protected]

Related

Cite as

Municorn.us, “Normalized consumption panels”, Municorn water loss index, https://municorn.us/build/normalized-consumption-panels/ (source filings retrieved 2026-09-08).

CC BY 4.0 — reuse and redistribution are permitted with attribution to Municorn.us and a link to this page. Terms, and the one caveat that must travel with these figures.

"Water loss" is at least three different quantities depending on the jurisdiction, and nothing here supports a comparison between states.