How these cost estimates are built

Roofing pricing (the trade with full coverage today - see “What’s not covered yet” below) is built from real, itemized unit costs, not a single averaged figure:

• Craftsman Book Co., 2026 National Home Improvement Estimator - real material and labor costs for individual line items (a specific shingle product, a specific length of drip edge, a specific pipe flashing), plus a published crew-rate table and a job-condition labor-adjustment table we cite directly rather than re-deriving.
• Craftsman’s 2026 ZIP-Based Area Modification Factors - ZIP-level cost variation, aggregated up to the county level with a population-weighted mean (falling back to state, then a national median, for ZIPs with no county-level match). Applied separately to material, labor, and equipment costs, not blended into one multiplier.
• US Census Bureau - ZCTA population figures (for the population-weighting above) and county reference names, so a result is labeled with the actual county it was priced at, not a ZIP-derived city name that can span the wrong county in a multi-county metro.
• State sales tax rates - public state revenue department data, applied to the material portion of a roofing estimate.

We verified the crew-rate arithmetic directly: checking labor cost against published manhours × crew rate across the full costbook found a 5.8% exact-match miss rate, with a maximum discrepancy of $4.84 on any single line - consistent with ordinary rounding in a book that prints rates to the cent and manhours to a few significant figures, not a data error. Nothing here was invented to fill a gap; where we didn’t have a real number, we left it out and said so (see below).

1. Price real line items, not an average. Each tier (budget, mid-range, premium, plus tile and wood shake as their own priced options) is a specific, real assembly - a specific shingle or panel product, specific underlayment, specific flashing and trim - not a single blended “roofing costs $X” number.
2. Scale to your roof. Roof area is derived from your house size, story count, and roof pitch, not assumed flat. Material-quantity line items (shingles, underlayment) scale with that area; trim line items (drip edge, ridge, rake) scale with a derived edge length, not area; a few line items (pipe flashing count, vent count) stay fixed or step in the way Craftsman’s own data describes.
3. Apply real jobsite-condition adjustments. Story count and roof pitch carry a labor-time adjustment from Craftsman’s own published table (steeper and taller means slower work), applied to the labor cost specifically, before markup.
4. Apply regional cost factors. Your ZIP resolves to a county (or state, or national, if no county-level match exists - see “Regional confidence” below), and that region’s real material/labor/ equipment cost factors are applied.
5. Convert cost to price. Material cost is taxed at your state’s rate and then marked up; labor and equipment are marked up at the same rate without a tax step. This replaced an earlier flat overhead-and-profit assumption once we had a real formula to work from.

Every estimate is resolved at one of three levels, and the range shown widens as the resolution gets coarser:

• County (most ZIPs) - range is the typical estimate ±20%.
• State (used when your specific county has no cost data) - ±30%.
• National (last resort, when even the state can’t be resolved) - ±50%.

The ±20% county-level band reflects Craftsman’s own stated experience that a high-to-low bid spread on the same job is commonly 20% or more - even at our most confident resolution level, real bids vary.

Puerto Rico and the US Virgin Islands currently resolve at the national level rather than a real regional figure - our licensed source data has no ZIP-level factors for either territory, so there’s nothing to build a county or state figure from yet. This is a known gap, not a guess we’re standing behind at national confidence.

A few real, small figures aren’t in Craftsman’s data at all and needed a documented assumption to compute anything. Every one of these is tagged as an assumption in our underlying data, not presented as sourced:

• A 10% waste allowance on felt underlayment (the one material-only line item where waste isn’t already built into an installed price).
• Ice-and-water-shield coverage width along the eaves.
• How much roof area needs spot deck repair at “minor” vs. “major” condition.
• Treating a chimney as exactly one chimney, and two plumbing-stack penetrations as the default count, since our questions capture presence, not count.
• Roof edge lengths (for trim quantities) are derived assuming a square footprint, since house size alone doesn’t tell us the actual shape - real houses are usually more rectangular, so this likely slightly understates trim quantities.

Real costs a roofing job typically incurs that aren’t in our number at all - not an oversight, just not sourceable from our current data without inventing a figure:

• Permits.
• Material delivery / staging fees.
• Debris disposal and dumpster costs.
• Scaffold, staging, or fall-protection equipment rental for steep or multi-story jobs - this is a known, systematic underestimate for those jobs specifically, not a small gap: our labor-time adjustment accounts for a steep or tall roof taking longer to work, but not the cost of the equipment that job often requires.
• Structural rafter or framing repair found under damaged decking (we do price spot deck repair, not framing).
• Ice dam removal or prevention work.

Roofing is the only trade priced this way today. The other six trades on this site (siding, windows, HVAC, walk-in showers, walk-in tubs, bathroom remodeling) don’t have this level of real, itemized pricing built out yet - their cost guides will say so plainly rather than show a number we can’t stand behind. Within roofing, roof complexity (simple gable vs. multiple valleys and dormers) affects labor time but not material quantity yet - valley and hip flashing costs are real line items in our data but aren’t priced in, since we don’t have a sourced figure for how much valley footage a “complex” roof typically has.