Savings verification

Understand what changed, and how much energy it saved

Compare measured energy use after an improvement with a baseline for the reporting conditions. The calculation, data coverage and adjustments sit beside the result.

Baseline against actualSample data
MWh / week0100200measure installedbaseline period — model fitted herereporting period16.2% below the baselineweek 1week 10week 18measuredbaseline (counterfactual)
Measured use tracks the baseline until a measure is installed, then falls below it.Sample data, not a customer result. The band is the baseline’s own uncertainty.

Why not compare with last year?

There is no meter for energy that was not used, so a saving is a comparison with what the building would have used. Last year’s total, with no adjustment, answers a question nobody asked: the two years differ in weather, occupancy and often tenants. A baseline run over the conditions that actually occurred does not. Read the argument

How it works

How is a saving worked out?

  1. 1

    Baseline

    Fit a model to a period before the change.

  2. 2

    Reporting period

    Measure energy after the change.

  3. 3

    Comparison

    Run the model over the reporting conditions and compare.

The result

What does a result include?

  • Energy saved, with the baseline and reporting periods.
  • The share of the period that was actually measured.
  • Any adjustments, with their reasons.
  • The method, fit and limits.

A result analyses the data supplied. It is not a certification or third-party assurance.

A generated reportProduct · demo data
A generated report in the report list of a sample workspace.
A generated report in the demo workspace.

Method

Which IPMVP option fits?

OptionBoundaryMeasured
A · key parameterThe affected systemThe driving parameter; others stipulated
B · all parametersThe affected systemEverything, by submeter
C · whole facilityThe incoming supplyUse, regressed on weather and other drivers

Option D, calibrated simulation, gives modelled rather than measured savings, and is labelled that way. Drivers, fit criteria and adjustments

Changes the model cannot explain, such as a floor taken out of service, are adjusted for explicitly with the evidence attached. Otherwise a new baseline is set.

What data does a project need?

A baseline period, a reporting period and the energy records for both. Send us what you have and we will say which option it can support.

Talk to us about your project