Cost Index Methodology

Each row of the index is the headline cost record of one of our 59 published cost guides. A cost record is a reviewed national planning band: a low, a typical, and a high figure for a defined standard scope, in U.S. dollars, with no ZIP or city multiplier applied.

Records are built by an editor, not scraped. The editor reads published contractor pricing pages, manufacturer data, and industry cost references, normalizes them to a common unit and scope, and records the original reported range alongside the normalized band. Every record is marked VERIFIED with a reviewer and a review date before it can appear on the site. The Reviewed column on the index shows each figure's review date.

What the bands mean. Low is the realistic floor for a straightforward job with standard materials; typical is the planning point most homeowners land near; high covers premium materials, complex conditions, or major-market labor. Units differ by project (per sq ft, per project, per window, and others), so rows are not interchangeable. Each linked guide spells out the exact scope, the labor and material split, and the line items included and excluded.

What we do not claim. This is not a survey of homeowners or contractors, and no figure is invented to fill a gap. Where published sources disagreed or the scope was unclear, the record says so in its notes, which are quoted on the guide. If a source updates its pricing, the record is re-reviewed and the index refreshes with it.

One scope, one unit. Published sources rarely price the same job the same way: one quotes per square foot, another per project, a third for a defined room size. Before any band is set, the editor restates every source figure against the record's single standard scope and unit, and the original wording of each source is preserved in the record's notes so the conversion can be audited on the guide.

When sources disagree. The band brackets the normalized sources rather than copying one of them: the low sits near the lowest credible figure for the standard scope, the high near the highest, and the typical is the planning point the corroborated figures cluster around. When one source is a clear outlier for an identifiable reason (a different scope, a premium-only sample, a stale year), it is excluded from the band and the exclusion is noted. No figure is averaged blindly, and no band is widened or narrowed to hit a target.

Units are never mixed. A per-square-foot rate and a whole-project cost measure different things, so the index keeps them in separate groups everywhere: the dataset, the rankings, and the derived findings on the index page. Only the 33 records priced as a complete installed project are ranked or averaged together. Two records whose scope duplicates another record (a cabinet refacing comparison and a furnace-versus-heat-pump comparison) are excluded from the ranked set for the same reason: the underlying projects are already counted once.

Confidence. Each record carries a confidence mark. HIGH means multiple independent published sources corroborate the band. MEDIUM means the band leans mainly on one dataset, or the corroborating sources only partially agree, so the figure should be treated as a planning range, not a benchmark.

Labor and material splits. A record stores a combined labor and material split only when a published source reports one whose components nest inside the installed band. Many sources publish component ranges that do not nest (a material band drawn from premium samples, a labor band from a different scope), so for those records the split is stored as zero and the published components are quoted verbatim in the record's notes instead. We would rather show no split than adjust a published figure to make one fit, which is why no labor-versus-material ranking is published anywhere in the index.

Every record names its sources in three places: the dataset itself, the list below, and the linked cost guide, where the exact figure taken from each source is quoted. Sources are published contractor and marketplace pricing pages, manufacturer data, trade publications, and the Remodeling Magazine Cost vs. Value report. No source is anonymous, and no figure enters the index from a source we cannot link or name.

Update cadence. Records are re-reviewed on a rolling basis, not in one annual pass: when a cited source republishes its pricing, when a guide is revised, or when a reader flags a figure under our corrections policy. Each re-review updates the record's review date, and the index and its derived findings are regenerated from the dataset at the same time, so the pages never drift from the CSV. The review date on each row is the honest last-checked date for that figure.