A useful score
shows its work.

Who qualifies for comparison?

The first Kansas City research sample contains 52 firms gathered through barndominium, shouse, residential post-frame and related searches. It is a sample, not every result in Google or every company serving the metro. The original Google and registry records are dated September 27, 2026. The current revision corrects eligibility and classifications on September 29.

Before scoring, a company needs a working company site with evidence of a barndominium, shouse, residential post-frame product or named project; at least five recorded Google reviews; a recorded rating of at least 3.8; and a location or stated service area relevant to the KC region. General custom-home construction alone is insufficient.

• A regional base within approximately 80 straight-line miles is a research-screening criterion, not proof that the firm accepts work at every metro address.
• Companies whose stated service area excludes the metro do not qualify.
• A company with a commercial relationship to the publisher—client, tenant, sponsor or partner—is excluded from scoring.
• Missing scope evidence is an eligibility failure. Missing footprint or registry tenure is an unscored input, not a finding of poor quality.

The five-review floor is an editorial choice for this lower-volume category. Five reviews do not establish statistical reliability. All 40 exclusions appear on the Kansas City page with their recorded reasons.

The components and base weights.

Footprint describes an evidenced operating model, not a legal audit of beneficial ownership. Local/family labels are company claims. A supplier’s manufacturer affiliation alone does not establish who owns the business. Version 1.1 replaces the earlier “independence” wording and omits unsupported classifications.

Reputation: moderate small review samples.

B = (v × R + m × C) / (v + m)
R = recorded Google rating
v = recorded Google review count
m = 25 equivalent reviews (fixed prior weight)
C = 4.6 stars (fixed editorial prior)

reputation = clamp((B − 3.5) / 1.5 × 100, 0, 100)

The 4.6 prior is not a measured metro mean. It is an explicitly chosen reference value. This adjustment pulls small samples toward 4.6; it neither validates reviews nor guarantees that a high-volume company outranks every smaller one. Ratings may cover services other than barndominiums.

A worked example.

For 5.0 stars and 14 reviews: B = (14 × 5.0 + 25 × 4.6) / 39 = 4.7435897, giving a reputation score of 82.9059829. With full-home scope (100), local/family footprint (100) and missing tenure, the total is:

(45 × 82.9059829 + 35 × 100 + 10 × 100) / 90
= 91.4529914 → displayed score 91.5

For 4.4 stars and 88 reviews, the adjusted reputation score is 62.9498525, displayed as 62.9. Each component is displayed to one decimal place where needed, but calculations use unrounded values.

Missing inputs and normalization.

The score is the weighted sum of available component scores divided by the sum of their available weights. If tenure is missing, divide by 90. If footprint and tenure are both missing, divide by 80. If only footprint is missing, divide by 90. Missing inputs can move the score up or down; companies are not all scored from an identical evidence set.

total = Σ(available component score × base weight)
        / Σ(available base weights)

Order by the unrounded total, then unrounded reputation, then review count, then company name. Round only the displayed result to one decimal place. Two displayed totals can therefore match without an exact underlying tie.

Scope: read the label and the evidence.

These categories distinguish purchasing arrangements. They are not a quality ladder.

• Full-home construction advertised — 100: the company describes managing a complete home build or work through finish-out. Its exact contract still needs confirmation; we do not certify a complete list of included trades.
• Structure / shell — 60: structural construction is supported, but interior finish responsibility is limited or not established. Some companies explicitly stop at the shell; others do not clearly define the handoff. Each profile states which is the case.
• Custom builder with barndominium evidence — 50: a named project or product supports inclusion, but a dedicated barndominium service scope is not itemized. This is a deliberately conservative category, not a claim of lesser skill.
• Package / design dealer — 25: material supply, design or procurement guidance is separate from the general contractor’s field-construction agreement.

We do not infer complete plumbing, electrical, HVAC and interior finishing from photographs, a plans service, insulation, concrete or interior framing alone. A source URL supports the classification, and the profile explains any uncertainty.

Entity age is not trading history.

Where the September 27 record contains a matched Missouri entity name, charter number, formation date and registry status, the score uses that date. Kansas records and unresolved entity matches remain unscored. The state search link and identifiers are included so the match can be checked; these registry records were not refreshed in the September 29 content correction.

years = (score date − formation date) in days / 365.2425
score date = 2026-09-27

tenure = 0                              if years < 2
       = clamp(100 × ln(years/2) / ln(10), 0, 100)
                                        otherwise

A fictitious-name registration is not an entity formation date. A founder’s experience, a predecessor business and the current LLC’s age may differ. An active entity record is not proof of continuous trading, a contractor license, insurance, solvency or construction quality.

What the ranking does not establish.

• Project quality, completed-project counts, financial stability or customer satisfaction beyond the dated Google inputs.
• Current licenses, insurance, inspection history, availability or acceptance of a particular build address.
• Per-company prices, construction schedules, warranties or whether a bid includes a specific trade.
• An unbiased definition of “best.” The weights, review floor, prior and scope categories are editorial decisions.

Research and classification are assisted by software. This is desk research from public sources; it does not include site visits or a construction-quality audit. Follow the linked company and government sources when making a decision.

Updates, corrections and reuse.

The current release is 2026-Q3-r2. Ratings and registry records remain a September 27 snapshot; scope and content corrections are dated September 29. A newer content date does not imply that every source was refreshed. Only the latest revision is currently published at the main data URLs; no quarterly refresh or public archive schedule is promised.

Material corrections are recorded on the affected page. The JSON and CSV downloads contain the current classifications, dates, score components and exclusion records. The JSON includes the available weights per company. Our original selection, classifications and explanatory text are offered under CC BY 4.0; third-party ratings and source content retain their own rights.