Methodology
Two different kinds of number appear on this site, and we keep them apart on purpose.
1. Sourced facts
These come from published data and every one carries a source link and a sample date.
| Input | Source |
|---|---|
| Local construction labour rates | BLS Occupational Employment and Wage Statistics, May 2025, occupation 47-0000, all US metro areas |
| Material price movement | BLS Producer Price Index, monthly commodity series |
| Baseline national cost ranges | EIA, Updated Buildings Sector Appliance and Equipment Costs and Efficiencies (March 2023) for equipment-based jobs. Other services show no figure until a citable source is in place. |
Bringing older figures forward
Published cost studies quote a dollar year. Where a baseline is older than the current month, we carry it forward using that job's own BLS Producer Price Index series, and show the adjustment as its own line under the figure — the original year, the index used and the exact percentage. We do not fold it in silently.
This adjusts for equipment price movement, not for changes in labour rates or installer margins, so it is an approximation. The wage figures are handled separately and are current.
2. Stated model assumptions
These are not measurements. They describe how we think a job's cost is structured, and we publish them so you can disagree with them.
Labour share
Local wages only move the labour portion of a job — materials are bought at close to national prices. A metro paying 30% above the national construction wage, on a job that is 50% labour, lands at a 15% premium rather than 30%.
| Job | Labour share | Baseline | Why |
|---|---|---|---|
| Solar Panel Installation | 30% | pending | Hardware dominates: panels, inverter and racking are bought at national prices, so local wages move the total less than on labour-heavy jobs. |
| HVAC Replacement | 35% | sourced | Equipment is the larger share; installation labour is one to two days for most homes. |
| Roof Replacement | 55% | pending | Tear-off and installation are labour-intensive, so local crew rates move this total noticeably. |
| Basement Waterproofing | 65% | pending | Excavation and interior drainage work are almost entirely labour and equipment time. |
| Foundation Repair | 65% | pending | Pier installation and slab work are labour and heavy-equipment driven. |
| Window Replacement | 40% | pending | Units are manufactured goods; labour is roughly an hour or two per opening. |
| Siding Replacement | 50% | pending | A balanced split between manufactured cladding and on-site installation time. |
| Kitchen Remodel | 45% | pending | Cabinets and appliances are bought goods; trades time still runs several weeks. |
| Bathroom Remodel | 50% | pending | Small footprint, high trade density - plumbing, tiling and electrics all in one room. |
| Water Heater Replacement | 45% | sourced | A half-day plumbing job on a manufactured appliance. |
| Tree Removal | 85% | pending | Almost pure crew time and equipment - there is no material to buy, so local wages dominate. |
| Concrete Driveway | 55% | pending | Forming, pouring and finishing are crew-hour heavy; the concrete itself is a commodity. |
Where a baseline covers something specific
Baseline is a gas furnace plus a 3-ton split central air conditioner at the 2023 standard. The range spans EIA's typical and high-efficiency equipment tiers; it does not capture contractor-to-contractor variation, so real quotes may fall outside it.
Baseline is a 40-gallon gas-fired storage water heater, total installed cost. The typical figure is the midpoint of the published low-high range.
Factor multipliers
Each calculator option carries a low and a high multiplier. The low bound of the band is multiplied by the low multipliers, the high bound by the high multipliers, and the typical figure sits at the geometric mean of the two. Ranges widen as your job gets less standard, which is the honest result: unusual jobs really are harder to price.
3. What we will not publish
- A dollar figure with no source behind it.
- A city page with no local data — those redirect to the state page instead.
- A national figure dressed up as a local one by scaling it with an uncited index.
Corrections
If a figure here disagrees with a quote you have actually been given, that is useful information. The quote checker on each page feeds real numbers back into the dataset over time.