Insulation Before Panels: The Audit That Sets the Order

A solar proposal quotes an array and a first-year bill saving. What it will not mention is whether the house burns gas for heat, because if it does, no panel on that roof touches a therm of it.

That is not a scandal. A solar salesperson sells solar. But it means the number at the bottom of a proposal answers a narrower question than the one most people are actually asking, which is: what is the cheapest way to spend less on energy in this house?

EIA's own summary of the 2020 Residential Energy Consumption Survey puts space heating and air conditioning together at 52% of a typical household's annual energy use, with water heating, lighting and refrigeration adding 25% and everything else — televisions, cooking, laundry, electronics — making up the remaining 23% (EIA, Use of energy in homes, read 18 August 2026). An array changes what a kilowatt-hour costs you. Insulation and air sealing change how many you need, and in a gas-heated house they reach a fuel the array cannot see at all.

So the question is ordering. And the order is not a matter of opinion — somebody already wrote the method down.

The ranking method is a federal regulation, and it is three paragraphs long

The Department of Energy's Weatherization Assistance Program cannot spend public money on a measure that does not pay for itself, so the rule that governs it had to define, precisely, what "pays for itself" means and in what order measures get considered. I pulled the current text from the eCFR versioner API on 18 August 2026.

The cost-effectiveness test, at 10 CFR 440.21(d), is this: each measure and each package of measures "must result in energy cost savings over the lifetime of the measure(s), discounted to present value, that equal or exceed the cost of materials, installation, and on-site supervisory personnel."

Note what that is not: simple payback. It is lifetime savings, discounted, against installed cost. A measure lasting 40 years and one lasting 10 are not comparable on payback years alone, and the regulation refuses to compare them that way.

Then the ordering, at 440.21(e). Audit procedures "must assign priorities among individual weatherization materials in descending order of their cost-effectiveness," after two adjustments:

1. Adjusting for interaction between architectural and mechanical measures "by using generally accepted engineering methods to decrease the estimated fuel cost savings for a lower priority weatherization material in light of fuel cost savings for a related higher priority weatherization material."
2. Eliminating any measure that is no longer cost-effective once that derating has been applied.

Paragraph (f) adds the rest of the discipline: use the local climate data, determine existing energy use "from actual energy bills or by generally accepted engineering calculations," treat the dwelling "as a whole system," and document the reason for any manual reordering. State procedures have to be re-approved by DOE every five years.

For scale, the companion spending rule caps the program at an average of $6,500 per dwelling unit weatherized in the State (10 CFR 440.18(a)). Paragraph (c) has DOE raising that figure every year since calendar 2010 by the lesser of the CPI increase or three percent, so the number actually in force is well above $6,500 and is published per program year in DOE's guidance — do not use the regulation's base figure as a current one. Either way it is a ceiling for low-income weatherization, not a market price. What it tells you is the order of magnitude the federal government's own arithmetic treats as recoverable on an existing house.

What a private auditor is supposed to hand you

The consumer-market equivalent of that rule is a standard: ANSI/BPI-1100-T-2023, Home Energy Auditing Standard, from the Building Performance Institute, which publishes it as a free PDF (read on 18 August 2026). It defines what an audit must contain; its companion ANSI/BPI-1200-S-2017 covers how to carry one out.

Seven clauses in it are worth knowing by number, because each one gives you something specific to ask for:

• 2.2.3 — the report includes a baseline energy-use analysis where consumption records are available. If nobody asked for twelve months of your bills, the baseline came from a default, not from your house.
• 4.3 — the report has to state the "Method/s used to determine cost benefit of the recommended upgrades." This is the clause to point at when a savings figure arrives with no working behind it.
• 5.1 and 5.3 — a cost-benefit analysis is mandatory, and savings estimates "shall clearly indicate whether savings are projected for baseload, heating, cooling, or total household energy consumption." A savings figure with no stated denominator is not compliant.
• 5.2 — the benefit may be expressed as reduced fuel use, reduced energy cost, a fractional improvement on existing performance, or a score on a benchmark such as Home Energy Score, Home Energy Rating or Energy Performance Score. Note what this permits: energy simulation software is "an option, but not a requirement," so a hand-calculated audit is not automatically a deficient one.
• 6.2.2 — there must be a prioritization hierarchy in the report so you can see how the ranking was arrived at. Read this one carefully: the hierarchy is the one "established by the AHJ, program or contractor," not a ranking the standard itself fixes. Health and safety ahead of efficiency is the example the standard gives, not a universal rule it imposes.
• 10.2 — the envelope evaluation includes air leakage "as determined by blower door diagnostic testing." Not estimated. Tested.
• 4.5 and the note to clause 4 — the auditor must disclose any other products and services the firm sells, and the report "shall not include recommendations for measures based primarily on a specific product line, services of a contractor, or convenience."

That last one is the reason to read the standard at all. It gives you a published, citable sentence to hold up when the audit report's top recommendation happens to be the thing the auditor's company installs.

The DOE Home Energy Score sits alongside this as a lighter option: a 1-to-10 asset rating built from a home's physical facts, issued through a DOE gateway operated by PNNL. It is a benchmark, not a diagnosis. It will not give you a blower door number.

Putting one measure on the same yardstick as the array

Here is the arithmetic. Every input comes from a document you already have or can get.

annual_saving = (therms_saved x $/therm) + (kWh_saved x $/kWh)

simple_payback_years = installed_cost / annual_saving

Take $/therm and $/kWh off your own bills as all-in rates — total charges for the period divided by total units, so that fixed and delivery charges are inside the number rather than quietly excluded. Take therms_saved and kWh_saved from the audit report, per measure. Take installed_cost from a contractor quote, not from the audit's estimate.

Worked with stand-in figures, purely so the shape is visible. These are not your rates:

Then the same two lines for the array: installed cost after whatever incentives actually survive, divided by the first-year reduction in what you pay the utility. Not the gross value of generation — the reduction in the bill, which depends on your export compensation and is a separate calculation entirely.

Now the honest part. Sometimes that table says the attic wins by a mile, and sometimes, in a house already sealed and insulated near code, it says the array does. Both are real answers. What is not a real answer is a payback figure produced by whoever happened to knock first.

A sanity band while you wait for the audit: EPA estimates homeowners save "an average of 15% on heating and cooling costs (or an average of 11% on total energy costs)" by air sealing and adding insulation in attics, floors over crawl spaces, and accessible basement rim joists (ENERGY STAR, Methodology for Estimated Energy Savings, read 18 August 2026). Read the assumptions before you use the number: a 1,700 sq ft home of 1970–1989 vintage, four bedrooms, 23% total duct leakage, a 25% reduction in total air infiltration, insulation brought up to 2012 IECC levels, modelled in Beacon Residential on DOE-2. The page's own Table 1 shows how much the climate zone moves it — total-house savings run from 5% in Zone 1 to 16% in Zone 8, and the heating-and-cooling-only figure from 7% to 20%. The 11% and 15% are national averages of those columns. Treat whichever row matches your zone as a first-pass ceiling on the whole package, not a promise about any one measure.

The sum that comes out bigger than the bill it is reducing

There is one way to build the table above that guarantees a wrong answer, and it is the obvious way. Put four envelope measures in four rows — air sealing, attic insulation, rim joist, duct sealing — copy the modelled saving printed beside each one, and total the column. On a leaky house that total can land above the heating line of the bill it is supposed to be reducing.

Not a rounding error. A category error. All four measures attack the same heat loss, so whichever goes in first captures most of what was available and every measure after it is working on a load that has already shrunk. This is the precise failure 440.21(e)(1) was written to prevent: decrease the estimated savings of the lower-priority material in light of the higher-priority one, and then, under (e)(2), drop it entirely if the derated figure no longer clears the cost-effectiveness test.

How big the derate should be is a modelling question, and the regulation deliberately does not name a number — it names a standard of practice, "generally accepted engineering methods." To see the shape of it, take an arbitrary 18% off the attic row above: the annual saving falls from $229.96 to $188.57 and payback stretches from 11.3 years to 13.8. Your auditor's number will be different. The direction it moves is not.

That gives a one-line test to run on any audit report. Do the per-measure savings sum exactly to the package total? If they do, interaction was not modelled, and every measure below the first one is overstated. BPI-1100 clause 4.3 already requires the report to state the "Method/s used to determine cost benefit," so asking which method produced those figures is not an imposition — it is asking for a section the report was supposed to contain.

Sizing a measure off your own consumption record rather than a default is the same discipline that decides how many kWh of battery a house actually needs, and the twelve months of billing data BPI-1100 clause 2.2.3 wants are the same twelve months.

R-values tell you the target, not the gap

ENERGY STAR publishes recommended retrofit levels for wood-framed homes by climate zone (Recommended Home Insulation R-Values, read 18 August 2026). The page states its own basis in a footnote — the guidance "is based on the 2021 International Energy Conservation Code (IECC) Residential Provisions Chapter 4 Table R402.1.3" — which is worth reading exactly as written. Recommended retrofit levels derived from a code are not the same thing as the code your jurisdiction enforces, and existing houses are rarely required to meet either. For attics: R30 in Zone 1, R49 in Zones 2–3, R60 in Zones 4A through 8 if currently uninsulated — dropping to R25, R38 and R49 respectively where three to four inches are already in place. Floors run R13 in Zone 1 up to R30–R38 in Zones 5–8.

That table is a destination, and it is routinely misread as a saving. Your payback depends on the distance. Going from nothing to R49 saves a great deal; going from R38 to R49 saves very little, because the heat flow you are stopping falls off sharply once the first layers are down. A quote that cites the recommended level without stating what is up there now has told you nothing.

The money that changed on 31 December 2025

Two federal credits lapsed at the end of 2025, and this one bites the envelope specifically. The Energy Efficient Home Improvement Credit under section 25C — which covered insulation and air-sealing materials, and included up to $150 for a home energy audit — applied to qualifying property placed in service on or after 1 January 2023 and before 31 December 2025, per the IRS credit page read on 18 August 2026. The Residential Clean Energy Credit under section 25D — the 30% one that covered solar panels and battery storage — ran out on the same date: it "is not available for any property placed in service after December 31, 2025," per the IRS page for that credit, also read on 18 August 2026.

Both sides of the comparison got more expensive at once, so the ranking between them may not have moved much even though every payback figure did. If a contractor's paperwork still nets out either credit, that tells you when the template was last updated.

What is left is state and utility money, and it has to be verified as still funded before it enters your arithmetic — the same caution that applies when a battery quote leans on an incentive to make its numbers work.

Three documents to put on the table before the site visit

Gather these before you book an audit. They change what you get out of it:

1. Twelve months of gas and electric bills, so the auditor can build the baseline BPI-1100 clause 2.2.3 calls for instead of using a default house.
2. Your all-in rates, computed yourself: total charges divided by total units, for each fuel, for a full year. These are the numbers that convert modelled savings into dollars, and they are yours, not the auditor's.
3. The solar proposal, if you are holding one, so that its cost and its first-year bill reduction go into the same ranked list as everything else.

Then ask the auditor two questions in writing: whether the report will state savings separately for heating, cooling and baseload as clause 5.3 requires, and how interaction between measures is handled. The answers will tell you whether you are buying a diagnosis or a sales call with a blower door attached.