Reporting a Savings Claim That Survives Review
You have completed an assessment of a 30-unit building and modeled 24% source-energy savings from a package of measures: attic insulation, air sealing, LED lighting, and a boiler reset control. Post-installation billing for the first year shows 11% savings against the pre-installation year. The installation year had 8% more heating degree days than the baseline year, and the property added six new tenants mid-year. The program requires reporting against a baseline and asks you to explain the discrepancy. The client is unhappy and believes the contractor underperformed.
1. Explain the gap between modeled 24% and observed 11%. Which factors are measurement artifacts and which might be real underperformance?
Read the worked answer
The gap has at least three components and they must be separated before blaming the contractor.
Measurement artifacts: (1) weather. The installation year had 8% more HDD, which inflates consumption and depresses apparent savings. Weather normalization would raise the observed figure, though not all the way to 24%. (2) Occupancy. Six new tenants in a 30-unit building is a 20% increase in occupied units, which raises DHW, plug loads, and setpoint-driven heating. That is a real increase in consumption but not a measure failure. (3) Baseline choice. If the baseline year was unusually mild or low-occupancy, the modeled percentage is measured against a soft denominator.
Possible real underperformance: attic insulation installed with gaps or compressed to less than specified R-value; air sealing that did not achieve the target CFM50 reduction; boiler reset not commissioned or overridden by the maintenance staff; LEDs partially installed. Each of these is verifiable.
The correct response is to re-normalize for weather and occupancy, then compare the residual to the measure-level expectations, and only then assess contractor performance. Presenting the 11% raw figure as the outcome is both unfair and technically wrong.
2. Draft the reporting language you would use, including how to present uncertainty without undermining the client's confidence.
Read the worked answer
Reporting language should be transparent, quantified, and forward-looking. A defensible structure:
"First-year observed savings were 11% of baseline consumption. After normalizing for 8% higher heating degree days and a 20% increase in occupied units, weather- and occupancy-adjusted savings are estimated at approximately 17-19%. The remaining difference from the modeled 24% is attributable to measure-level performance, which is under review. Verified attic insulation depth and blower-door results will be compared to the modeled assumptions, and the boiler reset sequence will be re-verified with the maintenance staff."
This language does three things: it reports the raw number honestly, it shows the adjustment method, and it commits to a specific verification path rather than a vague promise. It avoids claiming savings that were not measured and avoids asserting contractor fault before the measure-level check is complete.
The report should also state the normalization method and data sources, note that a single year is a small sample, and recommend a second year of monitoring with a documented measurement-and-verification plan. Presenting uncertainty as a method statement, not an apology, preserves credibility with both the client and the program reviewer.