Skip to content
Case study

Meter data that finally agreed with the bills

Three metering systems produced three consumption figures. We modelled them into one reconciled series and cut billing disputes by two thirds.

Billing disputes
-64%
Meters reconciled
410k
Estimated reads
-38%
Time to first value
9 wks

The situation

Westvale bills 410,000 connections from three metering systems installed across two decades. Each produced its own consumption figure, and the billing engine read whichever was available first.

Customers noticed. The contact centre handled roughly 1,240 disputes a month, and investigating one meant an analyst opening three systems and reconciling by hand.

Constraints

Billing could not pause. Any change had to run in parallel with the existing engine until the regulator-facing figures matched for three consecutive cycles.

What we did

We modelled all three sources into one reconciled consumption series, with explicit rules for which reading wins and why. Where sources disagree beyond a threshold, the record is flagged rather than silently resolved.

Estimated reads were separated from actual reads in the model, which nobody had done before. That alone explained a large share of the disputes.

What changed

Disputes fell by nearly two thirds within a quarter. The more useful outcome is that an analyst investigating one now opens a single record showing every reading, its source and the rule applied.

Results

MeasureBeforeAfter
Monthly billing disputes1,240450
Bills based on estimates22%14%
Days to investigate a dispute92
Consumption series in use31
Our call centre stopped being a complaints line about numbers. That is the whole project in one sentence.
Dana WhitfieldDana WhitfieldHead of Customer Operations, Westvale Utilities

What we left behind

  • A project in the client's own repository, with tests and CI
  • Documented definitions with named owners
  • Runbooks for every scheduled job, written for a non-engineer
  • Two weeks of training for the team taking it over

Tell us which number you cannot trust

Send one paragraph about the report that always starts an argument. A data lead, not a salesperson, replies within one business day.