Methodologyv1.0.0

Anatomy of the Homestead Score.

A 0 to 100 number that describes the physical condition of a home. Built to be defensible to home pros, insurance underwriters, and lenders. Every input is traceable, tiered by evidence, and tied to remaining useful life. Every score is reported with a confidence number so a thin profile is never mistaken for a thorough one.

Transparent

Weights, sources, and formula are public. Versioned.

Auditable

Every score returns its version, inputs, and confidence.

Evidence-tiered

Self-reported, documented, and verified are different products.

The formula

Weighted condition, adjusted for what's actually left.

Each system contributes a score from remaining useful life and reported condition. Systems are averaged within their category, categories are weighted by consequence of failure, and a small maintenance adjustment is added.

score          = round( weighted_avg(category_scores) + maintenance_adjustment )
                 clamped to [0, 100]

system_score   = rul_score(70) + condition_score(30)
adjusted_life  = base_life × climate_factor(zone, category)
remaining      = max(0, adjusted_life − age) / adjusted_life

confidence     = 0.30 × category_coverage
               + 0.20 × system_density
               + 0.35 × avg_evidence_weight
               + 0.15 × profile_freshness

Seven categories

Weighted by consequence of failure, not replacement cost.

A failing roof ends a closing. A failing dishwasher is an inconvenience. The weights reflect that reality. Anchors drawn from NAHB lifespan studies, IBHS loss research, NFPA safety standards, and ISO insurance frequency data.

Structure

Roof and foundation drive the largest insurance losses and most closing-blocking deficiencies.

22

weight

Envelope

Water intrusion is the leading homeowner claim by frequency.

18

weight

Climate Control

Cost of failure and habitability consequence are high.

16

weight

Safety

Life-safety consequence beyond replacement cost. Sources: NFPA, CPSC.

14

weight

Plumbing

Loss frequency, plus known-defect materials like polybutylene.

12

weight

Utility

Water heater and softening have moderate consequence.

10

weight

Appliances

Rarely affects insurability or collateral value.

8

weight

Remaining useful life

Real lifespans, cited.

Each system has a published expected useful life. We use the manufacturer or trade-association number, then adjust for the climate the home actually sits in. Below is the public-facing subset. The full table lives in the source code.

SystemYearsSource
roof22NAHB Study of Life Expectancy of Home Components (2007), updated by IBHS shingle research
metal roof50NAHB / Metal Roofing Alliance
siding30NAHB
window25NAHB / NFRC
hvac15ASHRAE / ACCA
mini split18ASHRAE
pex50PPI (Plastics Pipe Institute)
copper50Copper Development Association
polybutylene0ASTM withdrawn; known-defect material
water heater12DOE / NAHB; tank-type
tankless20DOE; tankless gas
panel40NFPA 70 / NEMA; known-defect brands excluded
smoke10NFPA 72; sensor degradation
dishwasher10NAHB; major appliances
refrigerator13NAHB
washer11NAHB

Evidence tiers

A 78 / Verified is a different product than a 78 / Self-reported.

Every data point carries a tier. The tier does not change the score directly. It changes the confidence number and the verification level the score can claim. An underwriter, a lender, or a buyer can tell at a glance how much weight to put on the number.

Tier 1

Self-reported

Homeowner attestation with no attached evidence.

Confidence weight 0.50

Tier 2

Photo or receipt

Homeowner-supplied photo, receipt, or product label.

Confidence weight 0.70

Tier 3

Dated document

Permit, invoice, manufacturer registration, or inspection report on file.

Confidence weight 0.90

Tier 4

Third-party verified

Licensed inspector, utility, or municipal record signed within 24 months.

Confidence weight 1.00

Verification level: Verified when minimum tier is 4, or average is at least 3 with minimum at least 3. Documented when average is at least 2.5. Self-reported otherwise.

Inputs and refusals

The score describes the house. Not the people who live in it.

A score that quietly absorbed credit data or demographic signals would not be a score about the home. It would be redlining with a new coat of paint. Homestead draws a hard line so it can be useful to every homeowner.

The score uses

  • Homeowner attestations and uploaded receipts, photos, manuals, inspection reports
  • Public parcel and assessor records when available for the property
  • Manufacturer and trade-association lifespan tables (NAHB, ASHRAE, IBHS, NFPA, DOE)
  • Local climate exposure as a category-specific multiplier
  • Completed maintenance and inspection history logged inside Homestead

The score refuses

  • Credit score, income, debt, or any data from a credit bureau
  • Race, ethnicity, religion, or any demographic of the household
  • Neighborhood crime data, school ratings, or social signals
  • Anything that would penalize a home for who lives in it

Governance and versioning

Versioned. Changelog public.

Every computed score returns its methodology version. Any change to weights, lifespan tables, climate factors, evidence-tier weights, or the formula bumps the version and is recorded in the changelog. A score generated under v1.0 can always be reproduced under v1.0, even after a v1.1 ships.

v1.1 will add persistence of every score event with its full inputs, a homeowner-and-pro dispute workflow, and external anchors for permit history and climate hazard exposure. v2.0 will require sign-off from a methodology board including a licensed home inspector, an underwriter, and a homeowner advocate.

We do not publish a backtest yet because no one publishes a credible home-condition backtest in their first year. The honest answer is that v1.0 is grounded in published lifespan and loss data, and v1.x will be validated against real claim and repair data from a partner pilot in the first market.

Want to see your home's score?

Start free. Two minutes to a draft profile, your first score, and a plan for the year.