Making a $329M maintenance backlog legible enough to act on

A campus-wide deferred maintenance dashboard for a university facilities group. Every building sized by scale and coloured by condition — and flagged where the inspection data behind it could not be trusted. It went on to shape project selection, investment and divestment decisions, and how rehabilitation work was bundled and bought.

Client

University of Alberta

Sector

Post-secondary, facilities & operations

Built

2022–23

Platform

Tableau

Scope

11 linked views · campus to single building

Backlog covered

$328.6M current, $1.77B over 20 years

Data as of

August 2022, source VFA

Every institution with a large building portfolio faces the same question each budget cycle: there is never enough capital to fix everything, so what gets fixed first? The honest answer depends on knowing, building by building, how much deferred maintenance has accumulated and how that compares with what the building is worth replacing.

That data existed. It sat in facility condition assessments, and it was not in a form anyone could make a decision from.

What the dashboard shows

Why a treemap

The landing view is a treemap of every building on campus. Area encodes scale, colour encodes Facility Condition Index — the ratio of accumulated deferred maintenance to asset replacement value. One glance answers the question a capital planner actually has: which buildings are both large and in poor condition?

A table of the same numbers would be complete and useless. Two hundred rows sorted by dollar value hides the building that is small, cheap to fix and about to fail. The form was chosen for the decision, not for the data.

Filters cut the same view by faculty, by campus and by funding type, because the question changes depending on who is asking. A dean wants their own faculty. A capital planner wants the funding envelope. Clicking any building opens its own profile, where the deferred maintenance is broken down by system category.

Being honest about the data

Not every building had a complete facility condition assessment. Rather than quietly averaging the gap away or leaving those buildings out, the dashboard marks them: an asterisk beside a building name means part of its deferred maintenance figure comes from placeholder data rather than an inspection.

Inspection currency is shown the same way. A building whose assessment has aged out reads inspection stale, and one that has never been assessed reads inspection date N/A, rather than silently appearing as if its number were as trustworthy as the rest.

Two further caveats sit on the face of the dashboard rather than in a footnote someone has to go looking for: the data is stated as current to August 2022 with its source named, and a line notes plainly that the figures do not cover every faculty building on campus.

This is the part that matters most, and it is the part most dashboards skip. A capital plan built on a number nobody flagged as an estimate is a plan that fails its first audit. The cost of saying so on the face of the chart is a small asterisk. The cost of not saying so is the credibility of every other figure on the page.

Beyond the snapshot

A current-condition view only answers half the question. The remaining views model what happens next: current versus projected deferred maintenance, the same split by severity, funding levels modelled across a twenty-year horizon, requirement-level detail per building, and a ranking view for comparing buildings on a consistent basis.

That is the difference between a report and a decision system. The first tells you the building is in poor condition. The second tells you what the backlog looks like in a decade under each funding level you might actually choose — which is the question the budget conversation turns on.

What it was used to decide

Three groups used it differently, which is the real test of whether a dashboard has landed. Planners and executives used it to inform investment and divestment decisions — not only what to fund, but what to stop funding. The facilities team used it to assess short-term maintenance priorities. And at project level, current facility condition became an input to project selection rather than a document consulted after the shortlist was already drawn.

The system-level view mattered more than expected. Reading condition across the portfolio one building system at a time — rather than one building at a time — made it possible to bundle rehabilitation work that would otherwise have been scheduled and tendered separately, and to size procurement contracts accordingly. That saving only becomes visible if the data can be pivoted that way, and it is invisible in a building-by-building report.

What changed

Maintenance issues started being addressed before they became replacements, which is the entire economic argument for tracking deferred maintenance rather than simply recording it. Catching a failing system early is the difference between a repair and a capital project.

Individual units secured additional funding on the strength of it, because for the first time they could show concretely what their investment dollars bought — which buildings moved, by how much, and what the backlog would have looked like without the spend. Procurement improved for the same underlying reason: maintenance could be planned far enough ahead to be bought well.

None of that came from the chart being attractive. It came from the numbers being trustworthy enough that someone was willing to put their budget request behind them.