Measurement & verification

The short version

Measurement and verification (M&V) is the practice of quantifying energy savings from efficiency projects — proving that the new equipment, controls, or operations actually saved the energy claimed. Because post-installation energy use is influenced by weather, occupancy, production volume, and operating changes, you can't simply compare last year's bills to this year's. M&V uses statistical methods to isolate savings specifically attributable to the installed measures versus changes in everything else.

The baseline problem and adjusted comparisons

The core challenge of M&V is that energy savings are never directly measured — they are calculated as the difference between what was used and what would have been used without the project. The "would have been" baseline is a counterfactual that must be modeled.

A simple before-and-after comparison fails because conditions change. If you retrofit lighting in January and compare it to the previous January's bill, weather may have been mild, occupancy may have shifted, production may have varied. Sophisticated M&V regresses pre-retrofit energy use against drivers (heating degree days, cooling degree days, production volume, occupancy hours), then applies the regression model to post-retrofit drivers to compute an "adjusted baseline" — what energy would have been at current operating conditions if the retrofit hadn't been installed.

The four IPMVP Options

The International Performance Measurement and Verification Protocol (IPMVP) defines four M&V approaches sized to project complexity. Option A (partial retrofit isolation with key parameter measurement) is typical for lighting upgrades — measure operating hours of the new fixtures, stipulate wattage based on equipment specs. Low cost, moderate accuracy. Option B (full retrofit isolation) measures all variables at the equipment level — used for motor or HVAC upgrades where both load and operating profile matter. Higher cost, higher accuracy. Option C (whole facility) uses utility billing data with regression analysis on the entire building — used for comprehensive multi-measure projects where isolating individual measures isn't practical. Option D (calibrated simulation) uses calibrated energy models — for new construction or complex projects where pre-retrofit baseline isn't available or doesn't reflect intended post-retrofit operation.

ASHRAE Guideline 14 and statistical rigor

ASHRAE Guideline 14 provides detailed statistical methods that complement IPMVP. It specifies minimum requirements for regression modeling quality (R² thresholds, residual analysis), uncertainty quantification (confidence intervals on calculated savings), and reporting (what variables, methods, and data must be documented). For large projects and performance contracts, ASHRAE 14 compliance is often required by funders and counterparties as evidence that the M&V methodology is technically defensible.

Uncertainty calculations matter. A project that calculated 1,000,000 kWh savings with ±20% uncertainty at 95% confidence is delivering different value than one with ±5% uncertainty. Some performance contracts include a "minimum guaranteed savings" floor that triggers payment adjustments when measured savings fall below the lower confidence bound — making rigorous baseline and uncertainty analysis economically critical.