Skip to content

A regional utility leaves Cognos

Fourteen years of outage, load and asset reporting locked inside a Cognos estate nobody had upgraded since 2019. Read, mapped and rebuilt at parity.

Built on synthetic data for a fictional company — Cascade Ridge Public Utility District. The method, the analysis and the outputs are real; the client is not.

Report definitions read
212
Source feeds mapped
9
Places docs disagreed with the system
17
To first working slice
1 week

What the estate looked like

Cascade Ridge ran outage reporting, load forecasting and asset condition tracking out of a Cognos environment that had accumulated fourteen years of report definitions. The last platform upgrade was in 2019. Two people understood the calculation logic and one of them had already given notice.

The reports themselves were not the difficult part. The difficulty was that the definitions of outage duration, customer minutes interrupted and asset criticality had each drifted into three slightly different versions, depending on which report you opened and which decade it was written in.

What the analysis found

We read the report definitions, the underlying Oracle schema, the nightly extract jobs and a slice of real data together, using a harness built for this estate. The harness is deterministic and re-runnable, so the same inputs produce the same findings every time.

Seventeen places where the documented behaviour and the running system disagreed. Nine feeds, four of which were being loaded by hand on a monthly cadence by a single person. Three separate definitions of a metric that appears on a board report.

What the rebuild looked like

Parity first: the same reports, the same numbers, in Power BI over a Fabric lakehouse. Then a single conformed definition per metric in a governed semantic model, with the three historical variants preserved and labelled so the board could see exactly what changed and why.

A fragment of the claim ledger

Every behaviour the harness extracted, tested against the system's own outputs. This is a small part of one demonstration's ledger — enough to show the shape of the work, not enough to hand over the method.

Synthetic. Truncated deliberately.

See what this method reveals in your environment.

Share one representative source, report set or business question. We will work out what can be demonstrated and what a production path would need, before you commit to anything.