The Energy Model That Promised 40% Savings
A contractor uses an energy modeling tool to propose a retrofit package for a 2,400 sq ft 1970s home: air sealing to 2,500 CFM50, attic insulation to R-49, duct sealing, and a new 95% AFUE furnace. The model predicts 40% annual energy savings and a 6-year simple payback. The homeowner signs. After the work, the contractor measures 3,100 CFM50 (not 2,500) and the homeowner reports bills only 18% lower than before. The homeowner is angry and wants the contractor to make up the difference. The contractor reviews the model inputs and realizes the baseline infiltration was entered as a default value rather than the measured pre-retrofit blower door result, and the model assumed a duct leakage reduction that was not fully achieved.
1. Diagnose the modeling errors and explain how each contributed to the overprediction of savings.
Read the worked answer
Two modeling errors stand out.
First, the baseline infiltration was a default value rather than the measured pre-retrofit blower door result. If the default was lower (tighter) than the actual measured leakage, the model would show a larger improvement from air sealing than the home actually experienced. The measured post-retrofit 3,100 CFM50 versus a target of 2,500 also means the air sealing fell short of the modeled assumption - so the model overpredicted the airtightness gain on both the baseline and the achieved side.
Second, the model assumed a duct leakage reduction that was not fully achieved. If the model credited a large duct leakage improvement that did not materialize, the predicted savings from duct sealing are phantom savings.
Both errors inflate the predicted savings because the model's baseline was not calibrated to the actual house and the post-retrofit assumptions were optimistic. The result - 18% actual versus 40% predicted - is consistent with a model that overestimated both the starting leakage and the achievable reduction.
The lesson: always calibrate the model to measured blower door and duct blaster data, and model realistic (not best-case) post-retrofit values.
2. How should the contractor handle the client relationship and the discrepancy, and what should be done differently on future projects?
Read the worked answer
Client relationship: Acknowledge the discrepancy directly and transparently. Show the homeowner the pre- and post-test data and explain the modeling errors in plain language. Do not blame the software. Offer a remedy proportionate to the error - for example, re-test and complete any air sealing or duct sealing that fell short of the modeled target at no or reduced cost, and provide a revised, calibrated savings estimate. If the shortfall is substantial and the contract promised specific performance, consider a goodwill adjustment. Document the conversation.
Process improvements: (1) Always use measured blower door and duct blaster values as model inputs, never defaults. (2) Model post-retrofit values conservatively, and state assumptions and ranges in the proposal. (3) Present savings as estimates with a range, not a single number, and explain what could cause actual savings to differ. (4) Verify post-retrofit performance with testing and compare to the model. (5) Include a commissioning step so shortfalls are caught before the client does. (6) Build in a feedback loop so future models are calibrated against actual results.
The contractor should also review whether the proposal language overpromised - a 40% savings claim with a 6-year payback is a strong promise that should have been presented with caveats.
3. What should the revised, calibrated savings expectation be, and how should the contractor communicate it to the homeowner?
Read the worked answer
The contractor should not simply re-run the model to produce a new single number. Instead, recalibrate using measured data: baseline infiltration from the actual pre-retrofit blower door, post-retrofit infiltration from the measured 3,100 CFM50, and actual duct leakage results. Then present a range.
The observed 18% savings is the best empirical anchor. The contractor can explain that the calibrated model now predicts savings in roughly the high-teens to mid-20s percent range depending on weather and occupant behavior, rather than 40%. If additional air sealing can bring the home from 3,100 to 2,500 CFM50, the model can show the incremental savings from that work, giving the homeowner a choice.
Communication: Use plain language, show the before/after test numbers, explain that the original estimate was based on assumptions that did not match the house, and present the revised range with the caveat that weather-normalized bills are the true measure. Offer to weather-normalize the first year of bills against the model to validate the revised estimate. Be honest that the original 40% figure was not achievable and explain what was learned - homeowners generally respond better to transparency plus a concrete remedy than to defensiveness.