A benchmarking number — the EUI on the front page of a disclosure submission, the GHGI on a board deck — is only as defensible as the data underneath it. In British Columbia's commercial and multifamily stock, the gap between defensible and probably right is where most owners get stuck.
This piece is the framework Vandecarb uses to set up whole-building energy data so that the resulting benchmarks hold up under regulator review, lender scrutiny, and your own retrofit decisions.
The five questions every benchmarking dataset should answer
Before anything is loaded into ENERGY STAR Portfolio Manager (ESPM), the underlying data should answer five questions. If it doesn't, the benchmarks built on top will inherit the gaps.
- Which physical building, exactly? Property line, civic address, gross floor area (GFA) basis, and which spaces are in scope versus parking/exterior/leased-out.
- Which meters serve it? Every utility account, every fuel, with a clean line from each meter to the spaces it serves.
- What was the consumption, not just the cost? Invoice totals are not consumption. Estimated reads, retroactive true-ups, and account transfers will distort year-over-year comparisons if not normalized.
- What weather-normalized period is this? A single calendar year is fine for disclosure. For trend analysis and retrofit measurement, weather-normalization on consistent heating/cooling degree-day basis is non-negotiable.
- Who does the meter belong to — and who pays for it? Tenant-paid utilities, common-area accounts, and net-metered renewables each get treated differently in whole-building totals.
The four issues that fail first-pass review
Across portfolios we've audited, the same handful of issues account for the majority of rejected or flagged submissions:
1. Tenant data gaps treated as zeros
In a multi-tenant commercial building, tenant electricity is often invisible to the owner. A submission that treats missing tenant consumption as zero will under-report the building's EUI by 30 to 60 percent. ESPM has tools for this — aggregated whole-building data requests through the utility, or default-value flagging — but they have to be used, not skipped.
2. Account-to-space mismatches
A single ESPM property can hide a portfolio-level mess: a parking-lot account billed against the main building, a separately-titled annex on the same meter as the head office, an old service drop that's been disconnected for years. Until each account is mapped to a confirmed list of meters and spaces, the building "saving energy" might just be losing an invoice.
3. Mixed-fuel reporting in inconsistent units
Natural gas in GJ in one column, ekWh in another, propane delivered by truck still on a paper log. The conversions are deterministic, but the reporting trail has to be explicit — every fuel, native unit, conversion factor, and source. Aggregated kWh-equivalents without that trail will not survive a regulator's first follow-up question.
4. Weather years not stated
"EUI dropped 12 percent year over year" is meaningless without the weather context. A mild winter in Vancouver will move EUI by more than most building-level operational changes. Reporting unweather-normalized year-over-year deltas as performance is the single most common reason benchmark trends don't tell the truth.
A working sequence
When Vandecarb sets up benchmarking for a new asset or portfolio, the order of operations is fixed:
- Asset definition — confirm GFA basis, in-scope spaces, vacancy treatment, and parking/exterior boundaries. This is the document that will be referenced when something is questioned a year from now.
- Account inventory — list every utility account, every meter, every fuel, with provider, account number, and the spaces it serves. Gaps are flagged here, not later.
- Whole-building data request — where the utility supports it, request aggregated whole-building data directly. This eliminates the tenant-data problem at the source.
- Twelve-month load — load a clean 12 months of consumption per fuel, with native units preserved alongside ekWh.
- Anomaly review — flag readings outside ±2 standard deviations from the building's monthly profile and reconcile each one. Most of these are billing artifacts; some are operational findings.
- Weather normalization — apply degree-day normalization on consistent base temperatures. Document which weather station was used.
- Submission package — ESPM-ready data plus the documentation trail that produced it.
Why the documentation trail matters more than the number
A benchmark number on its own is brittle. The same building, run through ESPM by two consultants, can produce two different EUIs depending on assumptions about vacant space, mixed-use floor area, and tenant data treatment. Neither is "wrong" — they're answering slightly different questions.
What makes a benchmark defensible is the documentation trail behind it: the assumptions, the source data, the reconciliations. When a regulator follows up, when a lender questions, when next year's submission shows a change, the trail is what answers.
That's the deliverable Vandecarb hands over at the end of a benchmarking engagement. The number is on the cover page. The trail is the engagement.
Working through a benchmarking deadline or a portfolio audit? Book a discovery call to talk through scope.