# Methodology

> How Window Replacement Lab builds its installed-cost ranges and the calculator model, including sources, arithmetic and limits.

Canonical: https://windowreplacementlab.com/methodology/

Last updated: 2026-11-08

Every number on this site is either sourced or labelled as a planning
    estimate. This page explains the difference, how the calculator is built,
    and what the model cannot see.

## 1. The calculator is component arithmetic, not a lookup table

    A window price is four things stacked: the unit, the labour to set it, the
    removal of the old one, and the weatherproofing and trim around the opening.
    The calculator prices each of those separately and adds them. That is why the
    result is a range rather than a single figure — each component has its own
    low and high band.

    The published component bands are on the
    [calculator page](/calculator/) in the transparency table. If a
    reader disagrees with a result, they can point at the component that caused
    it instead of arguing with a black box.

## 2. Two markets, stated not averaged

    The same opening costs roughly twice as much through a full-service
    replacement company as it does buying the unit and hiring a local installer.
    Both are real installed prices; they are different products. Most published
    cost guides pick one and present it as "the" average, which is why their
    ranges disagree so widely.

    This site exposes the choice as a switch instead of hiding it. A single
    blended average would be easier to publish and less useful.

## 3. What is measured and what is planned

     |
       | Input | Status | How it is handled

       | Unit price by style and material | Published band | Ranges cross-checked against several independent cost guides for the same model year

       | Labour per opening | Published band | Insert and full-frame are priced separately; bay and bow get their own band

       | Size multiplier | Published band | Small 0.8×, standard 1.0×, large 1.3×, oversized 1.7×

       | Regional labour multiplier | **Planning band** | A regional adjustment, not a measured wage. Labelled as such on the tool.

       | Permit allowance | Planning band | $0–$350 flat, scaled by opening count; the local jurisdiction decides

       | Energy savings figures | Published research | Cited to the programme or study that produced the number

    **The honest limit:** we do not have a live wage feed wired into
    the tool. The regional multiplier is a planning adjustment with a stated
    basis, and it is treated as one everywhere it appears. A reader who needs a
    measured local labour rate should start from the Bureau of Labor Statistics
    OEWS series for the relevant trade and metro, which is public.

## 4. Why published window costs disagree

    Three separate reasons, and they compound:

- **Market.** Installer-plus-unit versus full-service contract, roughly 2× apart.

- **Unit specification.** A clear double-pane vinyl unit and a Low-E triple-pane fiberglass unit are both "a window".

- **Scope.** A number that excludes trim, disposal and permit is not comparable with one that includes them.

    When we publish a range we state the market and the specification it
    describes. When a competitor's range differs from ours, the useful question
    is which of the three is behind the gap — not which number is "right".

## 5. What we have not tested

    We do not manufacture windows, install them, or run a test lab. When a claim
    comes from a manufacturer's published rating, a code document, a government
    dataset or an installer report, the page says which. We do not imply
    first-hand testing we did not perform, and we do not publish review scores
    for products we have not handled.

## 6. Corrections

    If a price band moves outside its published range, or a reader finds an error,
    we update the article and its "last updated" date. Email
    [contact@windowreplacementlab.com](mailto:contact@windowreplacementlab.com)
    with the page and the correction.