We never ask for quote contact details. No lead forms, no contractor ads—just numbers.

Method version 2026-07-27.4

How we build our foundation repair cost ranges.

The model is deliberately simple: named sources, visible units, dated review and no percentage adjustment unless a source supports it.

No. 01Principles

Uncertainty belongs in the result.

Use method and quantity before national averages.

A national project average mixes limited crack repairs, piering, waterproofing and much larger structural scopes. We use it as context, then calculate from the specific method and measured quantity where a verified unit range exists.

Keep shared upstream data visible.

Angi and HomeAdvisor are part of the same cost-data ecosystem. This Old House states that it uses Angi-derived estimates, and other editorial publications may repeat the same numbers. Multiple URLs therefore do not necessarily mean multiple independent datasets.

Widen or qualify—never average away disagreement.

If current sources disagree, we publish a defensible wider range and explain why. An older source can add context, but it is never the only basis for a 2026 range.

Give every number a claim ID.

Every published range resolves to a dated claim, its unit and scope, the supporting source records, and an evidence grade. Supported means compatible evidence from at least two independent families. Qualified means the number is useful but depends on a single family, an older source or a material scope caveat. Draft evidence cannot change a result.

No. 02Source hierarchy

Different sources answer different questions.

Evidence classes used by the foundation model
ClassExamplesUsed forNot used for
A · Official authorityICC, NRCS, NCEES, FTC, municipal authoritiesTechnical scope, licensing, permits and contract checksNational repair-price averages
B · Task-level cost dataEquivalent task-level dataFuture local/national ratios after a method match and auditGeneric metro or cost-of-living multipliers
C · Market and editorial dataAngi/HomeAdvisor-derived pages; HomeGuideNational ranges with compatible scope and unitsTreating derived pages or overlapping scopes as independent evidence
D · Validation dataBLS OEWS and BEA RPPLabor-component and broad-market sanity checksDirect repair-cost factors
E · Local and anecdotal contextDe-identified permit aggregates; forumsSeparate permit-valuation context, backtesting, language and edge casesInstalled-cost claims, automatic fees or multipliers

No. 03Model choices

What the calculator does—and refuses to do.

  • No regional multiplier until an equivalent task-level location ratio or independent installed-cost benchmark, local backtest and independent review are complete.
  • No automatic soil multiplier until a defensible percentage source is verified.
  • No automatic access multiplier; verified sources support only a directional effect.
  • No automatic permit add-on because requirements and fees are local.
  • No additive labor line because method ranges already include labor.

The market selector exposes the pilot geography and fallback clearly. It does not change today's range. For three method-matched research samples, it can now show permit valuation percentiles as a separate, clearly labeled comparison. For unit-compatible pier samples, those percentiles are also scaled by the quantity entered so the local signal is visible as a distinct total. That arithmetic remains outside the national model. Generic cost-of-living indexes are not a substitute for method-specific foundation evidence.

No. 04Regional model

Ten pilots. Zero active factors.

A market-method adjustment stays off until the repair task is matched, backtested locally and reviewed independently. If a method has no defensible match, it keeps the national range.

Regional research status, July 2026
Pilot marketOfficial local coveragePermit researchCalculator status
Dallas–Fort Worth, TXPartial source coverage0 screened; 0 qualitative; 0 installed-cost readyNational fallback
Houston, TXFoundation-specific source coverage464 screened; 20 qualitative; 0 installed-cost readyNational fallback
Los Angeles, CAFoundation-specific source coverage20 screened; 8 qualitative; 0 installed-cost readyNational fallback
Sacramento, CAFoundation-specific source coverage125 screened; 49 explicit-unit valuation candidates; 0 installed-cost readyNational fallback + separate permit context
Miami, FLPartial source coverage38 screened; 5 qualitative; 0 installed-cost readyNational fallback
Tampa, FLFoundation-specific source coverage0 screened; 0 qualitative; 0 installed-cost readyNational fallback
Raleigh, NCFoundation-specific source coverage72 screened; 37 explicit-unit valuation candidates; 0 installed-cost readyNational fallback + separate permit context
Charlotte, NCPartial source coverage0 screened; 0 qualitative; 0 installed-cost readyNational fallback
Columbus, OHPartial source coverage120 screened; 86 explicit-unit valuation candidates; 0 installed-cost readyNational fallback + separate permit context
Cleveland, OHFoundation-specific source coverage20 screened; 6 qualitative; 0 installed-cost readyNational fallback
Activation requires all three gates.

We need an equivalent task-level cost match, a de-identified local backtest and review by an independent U.S. estimator or licensed professional engineer. Paid task databases are not being used in the current zero-budget research cycle. Soil, access, foundation type, permit fees, BLS wages and BEA Regional Price Parities remain qualitative or validation inputs. Draft records carry no usable factor.

No. 05Fair Range Meter

Editorial thresholds, clearly labeled.

The meter flags a quote below 60% of the lower bound as “suspiciously low” and a quote above 130% of the upper bound as “well above.” The spaces between those thresholds use neutral “below range” and “above range” states.

These thresholds are an editorial comparison policy, not a market statistic. They help identify questions; they do not establish misconduct, poor workmanship or an unnecessary repair.

No. 06Changelog

Range review history

Foundation model 2026-07-27.4

Promoted compatible Raleigh and Sacramento permit evidence into a separate quantity-scaled valuation total. The displayed local total uses the selected pier count but still does not alter the national planning range or Quote Checker assessment. Columbus remains per-item only because strips, beams and anchors cannot be converted into the calculator's wall-section unit.

Foundation model 2026-07-27.3

Added separate method-level permit valuation context for Raleigh helical piers, Sacramento push piers and Columbus wall reinforcement. These percentiles never alter the national estimate or Quote Checker assessment and are explicitly labeled as permit valuations rather than invoices or installed costs.

Foundation model 2026-07-27.2

Added HomeGuide as a separate proprietary estimate family and audited every national number source by source. This Old House, the matching NerdWallet anchor and the older RepairPricer page no longer count as independent from Angi/HomeAdvisor for these claims. No dollar range changed; evidence grades became more conservative, except the default underpinning range now has two compatible families.

Foundation model 2026-07-27.1

Replaced the stale Los Angeles permit source with the current LADBS dataset. Screened 20 recent residential repair candidates, retained eight for qualitative scope context and rejected every declared value for cost-factor use. Separated the editorial author and publisher identities in structured data and expanded the two pause-first decision guides.

Foundation model 2026-07-26.1

Added claim-level evidence status, source-class and upstream-family metadata. Added ten metro pilots with draft, null adjustments and a mandatory national fallback. Added FTC contract guidance and official code, licensing, soil, labor, price-parity and local permit context without turning those sources into cost multipliers.

Foundation model 2026-07-24.1

Verified the F0 table. Updated the engineer range to $300–$1,000, defined major underpinning as $10,000–$30,000 typical with a possible $50,000 extended high, and removed unsupported state, soil and access multipliers.

No. 07Source register

Every source currently in the model.

  1. Source date 2026-03-17; verified 2026-07-26. Evidence class: market-data; family: angi-homeadvisor. Primary upstream cost survey used for national, method, issue, engineer-report, permit and labor ranges.

  2. Source date 2026-07-10; verified 2026-07-26. Evidence class: market-data; family: angi-homeadvisor. Primary upstream source for underpinning, pier counts, project labor and permit ranges.

  3. Source date 2026-07-08; verified 2026-07-26. Evidence class: market-data; family: angi-homeadvisor. Inspection pricing varies by complexity; suspected foundation issues can exceed the typical inspection range.

  4. Source date 2026-06-20; verified 2026-07-26. Evidence class: market-data; family: angi-homeadvisor. Shares an upstream cost-data ecosystem with Angi and should not be treated as an independent dataset.

  5. Source date 2026-03-06; verified 2026-07-26. Evidence class: editorial-market-data; family: angi-homeadvisor. Editorial presentation derived from Angi cost data. It is not an independent price-data family.

  6. Source date 2026-02-25; verified 2026-07-26. Evidence class: editorial-market-data; family: angi-homeadvisor. Secondary editorial synthesis. Its national anchor and engineer figure trace to Angi, so it is not counted as independent price evidence.

  7. Source date 2022-01-18; verified 2026-07-26. Evidence class: market-data; family: angi-homeadvisor. Older secondary page that attributes its national benchmark to HomeAdvisor and repeats the same method figures. Retained for provenance, not independence.

  8. Source date 2026-02-06; verified 2026-07-27. Evidence class: market-data; family: homeguide. Independent proprietary cost-database comparison. Several scopes differ from the runtime claims, so overlap alone does not make a range supported.

  9. Source date 2024-11-26; verified 2026-07-27. Evidence class: market-data; family: homeguide. Reports $10,000–$30,000+ for underpinning and $1,000–$3,000 for a single pier using a proprietary estimate database and editorial review.

  10. Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: city-of-austin. Official local example showing why permits cannot be represented by a single national automatic add-on.

  11. Source date 2024-01-01; verified 2026-07-26. Evidence class: official-code; family: icc. Technical context for foundations, soil, drainage and waterproofing. It does not publish repair-cost multipliers; local adoption varies.

  12. Source date 2024-01-01; verified 2026-07-26. Evidence class: official-code; family: icc. Technical context for underpinning design and sequencing. It does not publish repair-cost ranges.

  13. USDA Natural Resources Conservation Service. Web Soil Survey

    Source date 2026-01-01; verified 2026-07-26. Evidence class: official-data; family: usda-nrcs. Site-specific soil context only. The survey does not provide foundation-repair cost coefficients.

  14. Source date 2026-01-01; verified 2026-07-26. Evidence class: official-licensing; family: ncees. Directory for the relevant state engineering licensing authority; it does not provide price data.

  15. Federal Trade Commission. How to Avoid a Home Improvement Scam

    Source date 2026-01-01; verified 2026-07-26. Evidence class: official-consumer-guidance; family: ftc. Supports written-estimate, contract and payment checks. It does not support a fair-price threshold.

  16. Source date 2025-05-01; verified 2026-07-26. Evidence class: official-data; family: bls. Validation context for the labor component only; occupational wages are not installed repair costs.

  17. U.S. Bureau of Economic Analysis. Regional Price Parities by State and Metro Area

    Source date 2024-01-01; verified 2026-07-26. Evidence class: official-data; family: bea. Broad regional price-level sanity check only; never a direct foundation-repair factor.

  18. Source date 2026-01-01; verified 2026-07-26. Evidence class: vendor-information; family: gordian-rsmeans. Candidate method-specific regional source. No task-level match has been accessed, purchased or approved.

  19. Source date 2026-07-14; verified 2026-07-27. Evidence class: vendor-information; family: municipal-permit. Aggregates municipal permit records. API job_value is reported in cents and was normalized to dollars. It is a jurisdiction-reported valuation, not an invoice or verified installed cost. Raw responses and identifying fields are not retained.

  20. Source date 2025-06-27; verified 2026-07-26. Evidence class: official-local; family: city-of-dallas. Recognizes Foundation Repair Permit as a structural repair permit type. No regional price factor.

  21. Source date 2026-07-16; verified 2026-07-26. Evidence class: official-local; family: city-of-houston. Twenty de-identified qualitative records supported method and pier-count context; the report has no valuation field and raw rows are not retained.

  22. Los Angeles Department of Building and Safety. Building Permits Issued from 2020 to Present

    Source date 2026-07-18; verified 2026-07-27. Evidence class: official-local; family: city-of-los-angeles. A current 24-month query found 329 residential alteration/repair candidates. Twenty recent records were screened; eight were qualitative-only and none had comparable cost scope. No regional price factor.

  23. Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: city-of-sacramento. The v3 target reviewed 125 records and retained 49 push-pier observations with explicit counts. The official current-year comparison is directional; permit valuations are not used as a cost factor.

  24. Miami-Dade County. Building Permits Dataset

    Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: miami-dade-county. Five residential finalized records lacked valuation, repair-method and pier-count fields. County scope only; no regional price factor.

  25. Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: city-of-tampa. Foundation-specific permit, engineering documentation and inspection requirements. No regional price factor.

  26. City of Raleigh. Building Permits Dataset

    Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: city-of-raleigh. The official allowlisted audit reviewed 71 records and directionally confirmed the helical-pier permit sample. Official and aggregated records remain one municipal upstream family; no regional price factor.

  27. Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: mecklenburg-county. Load-bearing repairs require permits despite the general monetary exemption. No regional price factor.

  28. City of Columbus. Building Permits Layer

    Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: city-of-columbus. The official layer lacks descriptions, so it cannot method-match the permit-derived carbon-fiber and steel/anchor samples. The carbon-fiber sample is truncated and no regional price factor is active.

  29. Source date 2026-01-01; verified 2026-07-26. Evidence class: official-local; family: city-of-cleveland. Six of 20 screened records identified a pier method; declared values were rejected because scopes were mixed. No regional price factor.

Published ranges summarize available sources. They are not a local bid, appraisal, diagnosis, engineering opinion or guarantee of price.

Use the foundation calculator