ARCO Standard 2026.1

Section 1

Scope, Purpose and Definitions

1.1 Scope

This Standard defines the complete method by which the American Roofing Contractors Organization ("ARCO") evaluates roofing contractors operating in the United States and produces accreditation findings. It governs: the population of entities subject to evaluation (§2); the five weighted evaluation categories and the composite scoring formula (§3 and Part III); the automatic disqualifiers that withhold accreditation regardless of composite score (§4); the procedures by which findings are kept current and may change (§5); and the public verification system through which any finding can be independently confirmed (§6).

The Standard applies uniformly to every entity evaluated during Cycle 2026. No alternative criteria, abbreviated procedures, or negotiated terms exist. An evaluation either follows this document in full or does not produce a valid finding.

1.2 Purpose

Roofing is among the most fragmented licensed trades in the United States. A property owner selecting among contractors confronts dozens of operators whose licensing status, operational history, regulatory record, and customer outcomes are scattered across state licensing boards, Secretary of State filings, federal establishment data, and consumer platforms — each with its own format, currency, and coverage gaps. No single public instrument reconciles these records into one comparable, verifiable measure.

The purpose of this Standard is to define that instrument. Specifically, the Standard exists to:

a. establish a single published criterion set against which every U.S. roofing contractor can be measured identically, regardless of size, market, or affiliation; b. require that every input to that measurement be a public record or third-party data point that any member of the public could, in principle, retrieve and check; c. produce a composite score on a fixed ten-point scale, reconciled across record classes by a published formula rather than by judgment applied case-by-case; d. withhold accreditation automatically where specific high-severity conditions exist, regardless of composite performance (§4); and e. make every resulting finding independently verifiable through a public registry record that discloses the score, the standing, and the last review date (§6).

The Standard deliberately measures what the record shows, not what an entity says about itself. Self- description is unverifiable at scale; filings, license rosters, establishment data, and accumulated customer outcomes are not.

1.3 What this Standard is, and what it is not

To prevent misreading of findings produced under it, the Standard fixes its own limits. The following table is normative.

1.4 The records-first principle

Every provision of this Standard descends from one operating rule, stated here as a normative requirement:

NORMATIVE RULE R-1 · RECORDS FIRST

An evaluation under this Standard SHALL be constructed exclusively from records enumerated in the Source Registry (§15), each independently retrieved, verified, and RECORDED under the Data Verification Protocol (§12). No application, payment, interview, self-reported attestation, or contractor-submitted material SHALL be accepted, solicited, or considered at the evaluation stage.

Rule R-1 has three structural consequences that recur throughout this document. First, evaluation is initiated by ARCO, not requested by the contractor: the evaluated population is drawn from public licensing rosters and state filings as defined in §2, which means an entity cannot opt into a favorable evaluation or opt out of an unfavorable one. Second, every sub-score is reproducible: because each input is a retrievable public record, a third party applying the rubrics in Part III to the same records on the same date arrives at the same sub-scores. Third, accreditation cannot be sold: there is no submission to price, because there is no submission.

Members of the accreditation program pay annual dues for registry maintenance, the accreditation license, and member services after a qualifying finding exists — never to obtain, influence, or accelerate the finding itself. The separation between evaluation and membership is restated as a procedural requirement in §14.6.

1.5 Definitions

The following terms carry the meanings below wherever they appear in this Standard, its annexes, and any finding issued under it. Terms are listed alphabetically; definition codes (D-n) are stable across editions.

D-1 Accreditation. The standing conferred when a committee finding under §14 confirms that an entity's composite score meets or exceeds the threshold in §3.4 and that no automatic disqualifier under §4 is active, evidenced solely by a current public verification record.

D-2 ARCO ID. The permanent identifier assigned to an evaluated entity upon first evaluation, used to retrieve the entity's public verification record. The ARCO ID survives changes of trade name and persists whether or not the entity is accredited.

D-3 Composite score. The weighted sum of the five category sub-scores, computed under the formula in §3.3 and the rounding rules in §13, expressed to one decimal place on a 0.0–10.0 scale.

D-4 Cycle. The annual evaluation period to which an edition of the Standard applies. Cycle 2026 runs January 1 through December 31, 2026, governed by this edition (2026.1).

D-5 Disqualifier (automatic). A condition enumerated in §4 whose verified presence withholds or suspends accreditation regardless of composite score, without committee discretion.

D-6 Entity. The legal person evaluated: the licensed or registered business organization (or sole proprietorship) conducting roofing work, identified by its state filing and license records, together with its principals as defined in D-12.

D-7 Evaluation date. The date on which record collection for an evaluation closes under §12.5. All currency requirements (e.g., "as of the evaluation date") refer to this date.

D-8 Evaluation file. The complete RECORDED set of retrieved records, verification logs, sub-score worksheets, and computation outputs for one entity in one evaluation, retained under §12.7.

D-9 Finding. The formal output of an evaluation: either (i) accredited, (ii) not accredited — below threshold, (iii) not accredited — disqualifier active, or (iv) coverage excluded under §2.4. Findings are issued only by the Standards Committee under §14.

D-10 Measured indicator. A specific, record-derived quantity or status defined in a Category Specifier (Part III) that contributes to a category sub-score under that Specifier's rubric.

D-11 Public record. A record created or maintained by a government body (federal, state, or local) and retrievable by a member of the public, with or without fee, including license rosters, Secretary of State filings, OSHA establishment data, and federal exclusion lists.

D-12 Principal. A natural person who is an owner of ten percent or more of the entity, a governing officer or director of record, or the qualifying individual named on the entity's contractor license where the licensing state uses that structure.

D-13 Re-review. A subsequent evaluation of a previously evaluated entity under §5, producing a refreshed finding and resetting the last-review date in the public verification record.

D-14 Source Registry. The enumerated catalog of record classes (§15) from which evaluation inputs may be drawn. A record class not listed in the Source Registry SHALL NOT contribute to any sub-score.

D-15 Standing. The current state of an entity's most recent finding as displayed in the public verification record: current, lapsed, suspended, or withdrawn, as defined in §5.4.

D-16 Sub-score. The 0.0–10.0 value produced for one evaluation category by applying that category's Specifier rubric to the verified record set.

D-17 Third-party data. Structured data maintained by a non-governmental party with no commercial relationship to the evaluated entity's accreditation status, including consumer review platforms, complaint-handling organizations, and manufacturer credential programs, as enumerated in §15.

D-18 Verification record (public). The registry entry described in §6 through which any person may confirm an entity's finding, composite score, standing, and last review date by entering the entity's name or ARCO ID.

1.6 Rules of interpretation

a. Headings and eyebrow labels are for navigation and do not limit or extend the provisions beneath them. b. The singular includes the plural and vice versa; "including" means "including without limitation." c. Where a provision references a record "on file" or "of record," the reference is to the record as retrievable on the evaluation date, not as it may later be amended. d. Time periods stated in months run calendar-month to calendar-month; "twelve months" from March 4 ends the following March 4. e. Where this Standard is silent on a procedural question, the Standards Committee resolves the question by recorded interpretation under §16.3, and the interpretation applies prospectively only.

Section 2

Eligibility Criteria and Coverage

2.1 Covered population

This Standard covers entities conducting roofing contracting work in any of the fifty U.S. states. Coverage is determined by the public record, not by self-identification: an entity falls within the covered population when it appears on a state contractor licensing roster under a roofing or roofing-inclusive classification, or — in states that do not license roofing at the state level — when its state business filing, local licensing record, or established trade presence identifies roofing contracting as a principal activity. Coverage is national by design. Because licensing structures differ materially across states (state-level trade licenses, state registration regimes, county or municipal licensing, and unlicensed-trade states), the Category Specifier for Licensing & Compliance (§7) defines how the rubric normalizes across regimes so that an entity is scored against the requirements that actually apply where it operates, not against requirements its state does not impose.

2.2 How entities enter evaluation

ARCO initiates evaluation; entities do not apply. The evaluation pipeline is constructed from the public record in the following order:

01 02 03 ROSTER CONSTRUCTION ENTITY RESOLUTION SCHEDULING State licensing rosters and Secretary Trade names, legal names, license Resolved entities are scheduled into of State filings are collected for each records, and filings referring to the evaluation tranches across the Cycle. state and reconciled into a single same operating business are Scheduling order is administrative national roster of roofing resolved to a single entity under and carries no evaluative meaning. contractors, keyed by legal name, §12.3, and an ARCO ID is assigned. license number, and filing jurisdiction.

An entity MAY request that its scheduled evaluation be conducted, or that an existing finding be re-reviewed (§5.3); a request affects timing only. No request, and no absence of a request, affects any sub-score, the composite, or the finding.

2.3 Eligibility conditions

An entity is eligible for a scored finding when all of the following hold as of the evaluation date: a. the entity is identifiable in the public record as defined in §2.1 (roster presence or qualifying filings); b. the entity's state business filing status is active or its equivalent in the filing jurisdiction; c. the entity has a resolvable operating identity under §12.3 — a single legal person to which licenses, filings, and outcome data can be attributed without unresolved conflict; and d. sufficient record coverage exists to score every category under the minimum-data rules of each Specifier (Part III).

2.4 Coverage exclusions

The following entities are excluded from scored evaluation. An exclusion is not an adverse finding; it is a determination that the Standard cannot be validly applied. Exclusions are RECORDED with their basis.

EXCLUSION BASIS

EX-1 · Inactive entity The entity's state filing is dissolved, terminated, or administratively forfeited with no reinstatement of record, or every roofing license attributable to the entity is expired, surrendered, or revoked with no active successor record.

EX-2 · Out-of-scope trade The record shows the entity does not conduct roofing contracting as a principal or substantial activity (e.g., a general contractor with no roofing classification, roofing history, or roofing outcome data).

EX-3 · Unresolvable identity Entity resolution under §12.3 cannot attribute records to a single legal person without material conflict (e.g., two unrelated operators sharing a trade name in one market with commingled outcome data).

EX-4 · Insufficient record One or more categories cannot be scored under its Specifier's minimum-data rule, most commonly Customer Outcome Data for newly formed entities (§10.5).

EX-4 entities are re-queued automatically and re-examined in later tranches of the Cycle as their record accumulates. An exclusion under EX-1 through EX-3 stands until the underlying record condition changes.

2.5 Coverage parameters for Cycle 2026

Verification records on file are reported on the public registry and updated continuously as evaluations and re-reviews complete. The count of verification records displayed on the registry (e.g., the figure published at roofingcontractors.org) is a live operational statistic, not a provision of this Standard, and changes as the Cycle progresses.

2.6 Effect of corporate events

A change of legal form, a sale of substantially all assets, a merger, or a transfer of the qualifying license terminates the continuity of the evaluated entity unless the public record shows the same principals and the same license lineage continuing the same operating business. Where continuity terminates, the successor is treated as a new entity: it receives a new evaluation, and it SHALL NOT display or claim the predecessor's finding. Where continuity is ambiguous, the matter is resolved as an identity question under §12.3 and, if unresolved, the entity is excluded under EX-3 pending clarification of the record.

Section 3

Evaluation Framework — Categories, Weights, and the Composite Score

Evaluation Framework — Categories, Weights, and the Composite Score

3.1 Architecture of an evaluation

Every evaluation under this Standard moves through four stages in fixed order. No stage may be skipped, reordered, or collapsed into another.

01 02 03 04 PUBLIC-RECORD DATA VERIFICATION WE I G H T E D S C O R I N G COMMITTEE FINDING COLLECTION Each data point is Sub-scores are computed A composite at or above State filings, license independently verified per category under the 8.5 with no active rosters, OSHA and RECORDED under Part III rubrics, then disqualifier is eligible for a establishment records, §12. Anything that cannot combined under the §3.3 finding. The Standards complaint-organization be verified is not scored. formula and §13 Committee issues the data, and consumer- computation rules. finding under §14. platform aggregates are retrieved per the Source Registry (§15).

3.2 The five weighted categories

Every evaluated contractor is scored across the same five categories. Each category produces a sub-score on a 0.0– 10.0 scale under its Category Specifier (Part III); the weights below are fixed for the entirety of Cycle 2026.

3.3 The composite scoring formula

The composite score is the weight-reconciled aggregate of the five sub-scores:

Composite = (0.30 × SC-01) + (0.20 × SC-02) + (0.15 × SC-03) + (0.20 × SC-04) + (0.15 × SC-05)

where SC-0n is the sub-score for category C-0n on the 0.0–10.0 scale, computed under the corresponding Specifier in Part III. Weights sum to 1.00 by construction; the composite therefore also lies on the 0.0–10.0 scale. The composite is computed at full available precision and rounded to one decimal place only at the final step, under the rounding rules of §13.2. Intermediate values SHALL NOT be rounded.

3.4 Accreditation threshold

COMPOSITE THRESHOLD FOR ACCREDITATION

8.5 / 10.0 A contractor must meet or exceed 8.5 to be eligible for accreditation under Standard 2026.1 — and remains ineligible at any score while an automatic disqualifier (§4) is active.

The threshold is a constant of the edition: it does not flex by state, market, entity size, or evaluation tranche, and the Standards Committee has no authority to waive it for an individual entity (§14.4). A composite of 8.4 is below threshold; the rounding rules of §13.2 are exact and final precisely so that no discretionary "close enough" zone can exist.

3.5 Worked example

The following illustration shows the computation for a hypothetical entity. Values are illustrative only.

SUB- CATEGORY WEIGHT CONTRIBUTION SCORE

C-01 · Licensing & Compliance 9.2 0.30 2.760

C-02 · Operational History & Stability 8.6 0.20 1.720

C-03 · State Filing & Standing 9.0 0.15 1.350

C-04 · Customer Outcome Data 8.1 0.20 1.620

C-05 · Complaint & Public Findings 8.4 0.15 1.260

Composite (sum, then round per §13.2) 8.710 → 8.7

The composite of 8.7 meets the 8.5 threshold. The entity is therefore eligible for a finding provided the disqualifier screen of §4 is clear — eligibility on score alone is never sufficient.

3.6 Design rationale for the weights

The weight structure encodes three deliberate judgments, published here so that the structure can be examined rather than inferred.

• Right-to-transact dominates. Licensing & Compliance (C-01) carries the largest single weight, 30%,

because licensure is the one criterion with legal force: an entity's authority to perform roofing work at all rests on it where the operating state imposes it. A contractor weak on C-01 is weak on the threshold question every other category presupposes.

• Conduct evidence outweighs paperwork evidence. The two conduct-derived categories — Operational

History & Stability (C-02) and Customer Outcome Data (C-04) — together carry 40%, against 15% for State Filing & Standing (C-03). Filings demonstrate existence and orderliness; tenure and accumulated customer outcomes demonstrate behavior over time. The Standard weights behavior over paperwork while still requiring the paperwork.

• Adverse signal is double-counted by design. Complaint & Public Findings (C-05) carries 15% as a scored

category, and the highest-severity adverse conditions reappear in §4 as automatic disqualifiers. A moderate complaint record depresses the composite proportionally; a severe one ends the analysis regardless of the composite. This two-layer treatment is intentional: averaging alone can launder a severe adverse fact under four strong sub-scores, and the disqualifier layer exists to prevent exactly that.

3.7 Category independence and double-counting controls

Each measured indicator is assigned to exactly one category. Where a single underlying record could plausibly inform two categories — for example, a license-board disciplinary order is both a compliance fact (C-01) and a public finding (C-05) — the Specifiers allocate it once, by rule:

a. license status, classification, and board discipline score in C-01 only (§7.2); C-05 expressly excludes license- board actions already scored in C-01 (§11.5(a)); b. filing-derived tenure scores in C-02; filing status scores in C-03 (§8.5(b), §9.2); c. review ratings and volume score in C-04; complaint-organization records and adverse government findings score in C-05 (§10.2, §11.2).

The disqualifier layer (§4) is the sole sanctioned exception to single-counting, for the reason stated in §3.6(c).

3.8 What the composite is not

The composite is a reconciliation of verified records against published criteria as of the evaluation date. It is not a ranking of contractors against one another (two entities may both score 9.1), not a prediction of any future job's outcome, and not a measure of price, speed, or aesthetic quality, none of which appear in any Specifier. Persons relying on a finding should read it as exactly what §6 displays: a score, a standing, and a last review date, each traceable to this document.

3.9 Score bands and reporting

For display in the public verification record and in evaluation reports, composites are described by the following fixed bands. Bands are descriptive labels only; the accreditation boundary remains the single threshold of §3.4.

COMPOSITE BAND REPORTING MEANING

9.5 – 10.0 Exemplary Verified record at or near the ceiling of every weighted category.

9.0 – 9.4 Distinguished Strong verified record with minor deductions in one or more categories.

8.5 – 8.9 Accredited Standard Meets the published threshold; accreditation-eligible subject to §4.

7.0 – 8.4 Below Threshold Not accreditation-eligible in this Cycle; re-examined under §5.

0.0 – 6.9 Substantially Below Material deficiencies across one or more heavily weighted categories.

3.10 Stability of the framework within a Cycle

The categories, codes, weights, formula, threshold, and bands defined in this section are fixed for the duration of Cycle 2026. They may be revised only in a subsequent edition of the Standard, adopted under the maintenance procedures of §16, and any such revision applies prospectively to evaluations conducted under the new edition. No finding issued under edition 2026.1 is re-stated under a later edition's parameters; it is instead refreshed by re-review under §5, scored under whichever edition is then in force.

3.11 Relationship to the Category Specifiers

This section defines the skeleton; Part III supplies the muscle. Each Category Specifier (§7–§11) is the controlling authority for its category and defines, at minimum: (a) the category's scope and exclusions; (b) the measured indicators and their record sources; (c) the ten-point scoring rubric with fixed anchors; (d) the source hierarchy applied when records conflict; and (e) minimum-data and edge-case rules. Where a question of scoring arises that this section and the relevant Specifier answer differently, the Specifier governs (per the precedence rule of the front matter); where neither answers it, §16.3 interpretation procedure applies.

Section 4

Automatic Disqualifiers and Trigger Conditions

Automatic Disqualifiers and Trigger Conditions

4.1 Operation of the disqualifier layer

Accreditation is withheld if any condition in §4.2 applies — regardless of composite score. The disqualifier screen is applied twice in every evaluation: once at the opening of record collection (so that disqualified entities are identified early) and once immediately before the committee finding (so that a condition arising mid-evaluation is caught). A verified disqualifier produces the finding not accredited — disqualifier active, and no composite is published for the entity while the condition stands. Disqualifier determinations are factual, not discretionary: the Committee confirms whether the trigger condition exists in the verified record; it does not weigh severity, context, or mitigation.

4.2 The four automatic disqualifiers

DQ-1 · FEDERAL EXCLUSION SOURCE: SAM.GOV

TRIGGER— An active exclusion (debarment, suspension, or proposed debarment) on the federal System for Award Management applying to the entity or to any of its principals (D-12). VERIFICATION— Direct query of the SAM.gov exclusions dataset by legal name, trade name, and principal names, RECORDED with query date and match basis per §12. CLEARANCE — The exclusion no longer appears as active on SAM.gov, confirmed on two retrievals at least fourteen days apart.

DQ-2 · OSHA WILLFUL OR REPEAT PATTERN SOURCE: OSHA ESTABLISHMENT DATA

T R I G G E R — A willful violation of record against the entity, or three or more repeat violations of record, within the federal OSHA establishment inspection data look-back of five years from the evaluation date. VERIFICATION — OSHA establishment search by entity name, trade names, and known establishment addresses; citation type and current status RECORDED. Citations vacated or reclassified on appeal, as shown in the current federal record, do not count toward the trigger. C L E A R A N C E — The qualifying citations fall outside the five-year look-back, or the federal record is amended such that the trigger condition no longer exists.

DQ-3 · FAILING COMPLAINT-ORGANIZATION GRADE SOURCE: BETTER BUSINESS BUREAU

T R I G G E R — A Better Business Bureau rating of C, D, or F (including minus grades within those letters) attached

to the entity's BBB business profile at the time of evaluation. V E R I F I C A T I O N — Retrieval of the entity's BBB profile resolved to the correct legal entity under §12.3; grade and retrieval date RECORDED. An entity with no BBB profile is not triggered by DQ-3; absence of a profile is handled as a C-05 scoring matter (§11.5(c)), not a disqualifier. C L E A R A N C E — The published grade rises to B− or better, confirmed on re-retrieval.

DQ-4 · FAILING AGGREGATE CUSTOMER RATING SOURCE: CONSUMER PLATFORM AGGREGATES

T R I G G E R — An aggregate customer rating below 4.0 (on the normalized 5.0 scale of §10.3) computed across

twenty or more reviews, aggregating all platforms enumerated in the Source Registry on which the entity maintains a resolvable profile. V E R I F I C A T I O N — Platform-by-platform retrieval of rating and review count, RECORDED; aggregation per the

volume-weighted method of §10.3. Entities with fewer than twenty aggregate reviews are not triggered by DQ- 4; thin records are handled under the minimum-data rules of §10.5. CLEARANCE — The aggregate, recomputed under §10.3 at a subsequent evaluation or re-review, equals or exceeds 4.0.

4.3 Trigger timing and mid-cycle events

A disqualifier is assessed against the record as of the evaluation date. For an entity that is already accredited, a disqualifier condition arising mid-cycle is handled under the standing rules of §5.5: upon verification of the condition, standing moves to suspended, the public verification record is updated to display the suspension, and the suspension stands until clearance is verified. Suspension is not retroactive — it does not invalidate the original finding for the period before the condition arose — but the public record always displays present standing, never historical best standing.

4.4 Disqualifiers are not appealable as judgments

Because each trigger is a factual condition in a third party's published record, there is no ARCO appeal of the underlying judgment: ARCO did not make it and cannot unmake it. An entity that believes a triggering record is erroneous must pursue correction with the record holder (SAM.gov, OSHA, the BBB, or the platform). What an entity MAY raise with ARCO is an identity error — a claim that the triggering record belongs to a different legal person. Identity claims are examined under §12.3 and, where the claim is verified, the record is detached from the entity's file and the evaluation re-run.

4.5 Relationship to scored categories

Three of the four disqualifier sources also appear inside the scored rubrics — OSHA history in §7.2(d), complaint-organization standing in §11, and customer ratings in §10. Federal exclusion (DQ-1 / SR-04) is disqualifier-only and does not enter any category sub-score. For the three overlapping sources, the relationship is layered by severity: below the trigger thresholds, these facts deduct points proportionally inside their categories; at or beyond the trigger thresholds, §4 controls and scoring is moot. The thresholds in §4.2 are therefore the published boundary between "weighs against the composite" and "ends the evaluation."

4.6 Stability of the disqualifier set

The disqualifier set — four conditions, with the trigger parameters stated in §4.2 — is fixed for Cycle 2026. No condition may be added, removed, or re-parameterized within the Cycle. Edition revisions follow §16.

Section 5

Re-Review Procedures and Standing

5.1 Why findings expire

A finding is a snapshot of a moving record. Licenses lapse and renew; filings change standing; complaint records open and close; outcome data accumulates. A finding allowed to stand indefinitely would progressively misdescribe the entity it covers. This Standard therefore treats currency as a property of the finding itself: every public verification record displays a last review date, and the data behind a finding must be no older than the currency limit of §5.2 for the finding's standing to remain current.

5.2 Currency requirement

NORMATIVE RULE R-2 · TWELVE-MONTH CURRENCY

The verified record set underlying a finding SHALL be no older than twelve months, measured from the evaluation date of the most recent evaluation or re-review to the present date. A finding whose record set exceeds twelve months in age moves to lapsed standing automatically and without committee action.

Source updates are ingested continuously where the record class permits (registry-monitored classes are flagged in §15); continuous ingestion refreshes the file but does not substitute for a full re-review, which alone resets the last review date.

5.3 Re-review triggers

A re-review — a complete re-execution of the four-stage evaluation of §3.1 against the then-current record — is initiated by any of the following:

TRIGGER DESCRIPTION

RR-1 · Scheduled Annual re-review scheduled before the twelve-month currency limit of §5.2 is reached; the default path for accredited entities in good standing.

RR-2 · Event-driven A monitored record class reports a material change: license status change, filing status change, new complaint-organization grade, new OSHA citation, new federal exclusion, or an aggregate-rating crossing under §10.3.

RR-3 · Requested The entity, or a member of the public citing a specific record, requests re-review. A request fixes timing only; it carries no evidentiary weight and is RECORDED as administrative.

RR-4 · Corrective An identity or verification error under §12.3 or §12.6 is confirmed in an existing file; the evaluation is re-run on the corrected record.

5.4 Standing states

The public verification record displays exactly one standing at any time:

STANDING MEANING EXIT

Current A qualifying finding exists; the record set is within the Lapses at the currency limit; twelve-month currency limit; no disqualifier is active. suspends on a verified disqualifier; refreshes on re-review.

Lapsed The currency limit has passed without a completed re- Returns to current upon review. Lapse is administrative, not adverse. completed re-review meeting threshold; otherwise moves to the re-review's finding.

Suspended A disqualifier condition has been verified mid-cycle Clearance verified per §4.2, or under §4.3, or a SHALL-violation in the entity's file is corrective re-review under RR-4 under correction (§12.6). completes.

Withdrawn Accreditation removed: a below-threshold re-review, a A subsequent evaluation producing coverage exclusion arising (§2.4), or verified misuse of a qualifying finding. the finding under §17.4.

5.5 Mid-cycle adverse events

Upon RR-2 notice of a potential disqualifier, the condition is verified under §12 before any standing change occurs — monitored feeds raise questions; only verified records answer them. If verified, standing moves to suspended and the verification record updates within the registry's normal publication cycle. If the event is adverse but sub-trigger (for example, a new complaint, or a rating decline that remains at or above 4.0), it is queued as input to the next re-review rather than acted on immediately; the Standard changes public standing only on disqualifier-grade events or completed re-reviews, never on raw signal.

5.6 Effect on the accreditation license

Display rights over the ARCO accreditation mark follow standing mechanically: they exist while standing is current and cease while standing is anything else. The licensing terms themselves are a membership matter outside this Standard; what this Standard fixes is the trigger — standing in the public verification record — so that a reader checking the registry sees the same status the mark asserts. Continued display of the mark during non-current standing is misuse under §17.4.

5.7 No grace periods

There are no grace periods in the standing model. Currency math under §5.2 is exact; suspension under §5.5 takes effect on verification, not after a cure window. What protects entities from flicker is the verification requirement itself: no standing changes on an unverified signal, and identity errors are corrected under RR-4 with the registry record annotated accordingly (§12.6).

Section 6

Public Records and Verification

6.1 The public verification registry

ARCO maintains one public verification record per evaluated contractor, retrievable by entity name or ARCO ID at the public registry on roofingcontractors.org. The registry is the sole authoritative statement of any entity's status under this Standard: no certificate, letter, listing, seal, or representation — including materials issued by ARCO itself — overrides the registry. Where a displayed credential and the registry disagree, the registry governs and the credential is being misused (§17.4).

6.2 Contents of a verification record

Each verification record displays, at minimum:

FIELD CONTENT

Identity Legal name, principal trade name(s), ARCO ID, and operating state(s) as resolved under §12.3.

Finding The current finding under §1.5 D-9: accredited; not accredited — below threshold; not accredited — disqualifier active; or coverage excluded.

Composite Score For accredited entities: the composite to one decimal and its band under §3.9. Composites are not published for disqualifier-active entities (§4.1).

Standing Current, lapsed, suspended, or withdrawn, per §5.4.

Last Review Date The evaluation date of the most recent completed evaluation or re-review.

Standard Edition The edition of the Standard under which the finding was produced (for this Cycle, 2026.1).

Records reflect findings under ARCO Standard 2026.1 and are updated as committee review completes; the registry's publication cycle is operational, but a completed standing change SHALL be published in the registry before it is communicated anywhere else.

6.3 Public verification procedure

Verification is deliberately a three-step act that requires nothing from the contractor:

I II III ENTER NAME OR ARCO ID RETRIEVE THE RECORD REVIEW STANDING & DATE The registry accepts a company The record drawn from the The reader confirms finding, score, name with a state, or an exact ARCO contractor's underlying evaluation standing, and last review date — and ID. No account is required; the file is displayed; it reflects the most that the record is not yet certified contractor learns nothing about who recent committee finding under under a different, lapsed cycle. searched. Standard 2026.1.

6.4 Reference IDs on issued materials

Every certificate, profile page, and accreditation communication issued for an accredited entity carries the entity's ARCO ID (and, where space permits, the full accreditation number). The function of the printed ID is singular: it routes the reader to the registry. A reader presented with any ARCO-branded material should be able to move from the material to the authoritative record in one lookup. Materials lacking a resolvable ID should be treated as unverified.

6.5 What verification does and does not establish

A registry lookup establishes that the named entity holds the displayed finding, at the displayed score and standing, on the basis of records current as of the displayed review date. It does not establish anything outside the Standard's scope (§1.3): it is not a referral, a guarantee of any future job, or a statement about price or workmanship. The registry page states this limitation; this clause makes it normative.

6.6 Accuracy and correction of registry records

Registry accuracy is governed by the same correction machinery as the evaluation files beneath it:

a. any person MAY report a suspected error in a verification record, citing the entity and the field in question; b. reported errors are triaged as identity errors (§12.3), verification errors (§12.6), or computation errors (§13.5), and handled under the cited provision; c. a confirmed error is corrected in the registry within the normal publication cycle, and material corrections are annotated with a correction date; and d. where a confirmed error changed a finding or standing, the corrective re-review path RR-4 (§5.3) applies and the corrected finding supersedes.

6.7 Retention and historical states

The registry displays present state only; superseded findings and prior standings are retained in the evaluation files under §12.7 but are not publicly displayed. This is deliberate: the registry's job is to answer "what is this entity's status now," and displaying historical best states alongside present state invites exactly the confusion the registry exists to prevent. Researchers and authorities with a documented need for historical states may request them under §17.5.

6.8 Availability

The registry is published continuously on the public internet without fee, login, or rate restriction inconsistent with ordinary public use. Planned maintenance windows are posted in advance. If the registry is unavailable, no inference about any entity's standing should be drawn from the unavailability; the registry's content, not its uptime, carries the Standard's meaning.

Section 7

Category Specifier C-01 — Licensing & Compliance

Category Specifier C-01 — Licensing & Compliance

Licensing & Compliance C-01

7.1 Definition and scope

C-01 measures the entity's right to transact: whether the entity holds, maintains, and complies with the contractor licensure and trade-regulatory requirements actually applicable where it operates. The category covers license existence, classification fit, status continuity, bond and insurance conditions attached to licensure where the licensing state imposes them, license-board disciplinary history, and federal workplace- safety compliance posture below the disqualifier thresholds of §4.

C-01 expressly excludes: business filing status (scored in C-03); customer-facing outcomes (C-04); and complaint-organization records (C-05). License-board discipline scores here and only here (§3.7(a)).

7.2 Measured indicators

IND. INDICATOR WHAT THE RECORD MUST SHOW

7.2(a) License existence & An active contractor license, registration, or local equivalent in each operating classification state, under a classification that covers roofing work, attributed to the entity or its qualifying individual.

7.2(b) Status continuity License status history across the look-back: continuous active status versus lapses, delinquencies, or reinstatements, as shown on the licensing roster.

7.2(c) Bond & insurance Where the licensing state conditions licensure on bonding or insurance, the conditions roster shows those conditions satisfied and uninterrupted.

7.2(d) Safety compliance OSHA establishment history below the DQ-2 thresholds: counts, classifications, posture and abatement status of serious and other-than-serious citations across the look-back.

7.2(e) Board disciplinary License-board disciplinary orders, citations, or formal accusations of record, and history their disposition.

7.3 Scoring rubric

The sub-score begins from the anchor matching the verified record and applies the stated adjustments; the result is clamped to 0.0–10.0. Anchors are mutually exclusive and assessed top-down.

Adjustments. Deduct 0.5 for each additional cured lapse beyond those reflected in the anchor (max −1.5); deduct 0.5 for each additional abated serious citation (max −1.5); add 0.5 (to a ceiling of 10.0) where the roster shows voluntary supplemental classifications or certifications maintained without interruption for the full look-back.

7.4 Source hierarchy

When sources conflict, precedence is: (i) the state licensing board roster of the operating state; (ii) the federal OSHA establishment database; (iii) state-plan OSHA records where the state operates its own plan; (iv) local licensing records where the state delegates licensure. A conflict between a roster and any secondary description of the roster (including the entity's own marketing) is resolved by the roster, and the conflict is RECORDED.

7.5 Normalization across licensing regimes

The rubric scores entities against the requirements their operating states actually impose:

a. State-license states. The roster of the state licensing authority is controlling for 7.2(a)–(c). b. Registration states. Where the state registers rather than licenses roofing contractors, registration is treated as the licensure requirement; classification fit under 7.2(a) is assessed against the registration's stated scope. c. Local-license states. Where licensure is county or municipal, 7.2(a) is satisfied by active local licensure in the localities where the record shows the entity operating; statewide coverage is not required where the state does not require it. d. No-license states. Where no roofing licensure exists at any level, 7.2(a)–(c) are scored as satisfied by definition, and the C-01 sub-score is carried by 7.2(d)–(e). The Standard does not penalize an entity for its legislature's choices.

7.6 Edge cases and interpretive rules

a. Qualifier departures. Where a license rests on a qualifying individual who departs, the roster's treatment governs: a statutory replacement window shown as open is not a lapse; an expired window is a lapse from the date the roster shows. b. Reciprocity licenses. A license issued by reciprocity or endorsement is a license; its history in the issuing state is scored, and its underlying out-of-state history is not re-scored. c. Citations under appeal. An OSHA citation shown as contested and unresolved is RECORDED but not scored until final; a citation affirmed on appeal is scored from its original issue date. d. Successor liability. Compliance history follows the entity as resolved under §12.3; a continuity-preserving reorganization (§2.6) carries its history, and a true successor starts clean.

7.7 Minimum data and currency

C-01 is scoreable whenever the operating states are identifiable and their rosters retrievable; there is no volume-based minimum. Roster retrievals must be no older than ninety days at the evaluation date; OSHA retrievals no older than the evaluation tranche's collection window (§12.5). C-01 is a registry-monitored category for re-review trigger RR-2 (§5.3): license status changes propagate to standing review without waiting for the annual cycle.

Section 8

Category Specifier C-02 — Operational History & Stability

Category Specifier C-02 — Operational History & Stability

Operational History & Stability C-02

8.1 Definition and scope

C-02 measures demonstrated durability: how long, how continuously, and under what consistent identity the entity has operated. Roofing carries long warranty tails and seasonal cash-flow stress; an operator's accumulated tenure under a stable identity is the strongest record-derived signal that the business will exist when a workmanship issue surfaces in year six. The category covers operating tenure, identity continuity, principal continuity, geographic and registration footprint stability, and service-line consistency as evidenced in filings and license classifications.

C-02 expressly excludes: present filing status (C-03); the quality of customer outcomes (C-04); and adverse findings (C-05). C-02 asks "how long and how steadily"; the other categories ask "how well" and "in what standing."

8.2 Measured indicators

IND. INDICATOR WHAT THE RECORD MUST SHOW

8.2(a) Operating tenure Years from the earliest of: original state filing date, first license issue date, or first verifiable establishment record attributable to the entity under §12.3.

8.2(b) Identity Stability of legal name and trade names across the look-back: continuous identity continuity versus serial renaming, dissolution-and-refiling patterns, or phoenix sequences.

8.2(c) Principal Stability of principals of record (D-12) across filings and license records over the continuity look-back.

8.2(d) Footprint Registration and licensure footprint over time: steady or organically expanding stability operating states versus repeated entries and exits.

8.2(e) Service-line Roofing as a continuous classification across the look-back, versus roofing appearing consistency only recently atop an unrelated history.

8.3 Scoring rubric

Adjustments. Add 0.5 (ceiling 10.0) where principal continuity spans the entire look-back at the 8.0 anchor or below; deduct 0.5 for each verified market exit-and-reentry beyond the first (max −1.0); deduct 1.0 where roofing classification appears only within the most recent twenty-four months atop a longer non-roofing history.

8.4 Source hierarchy

Precedence: (i) Secretary of State filing history of the formation state; (ii) license issue and renewal history under §7's rosters; (iii) Secretary of State records of foreign-qualification states; (iv) federal establishment records where they evidence operating dates. Marketing claims of tenure ("serving the area since 1987") are not record and SHALL NOT be scored; where such a claim conflicts with the filing record, the conflict is RECORDED and the filing record governs.

8.5 Edge cases and interpretive rules

a. Predecessor tenure. Tenure crosses an identity event only where §2.6 continuity holds (same principals, same license lineage, same operating business in the record). Asset purchases of a brand do not carry the seller's tenure. b. Filing date vs. status. The original filing date and the filing timeline score here under 8.2(a)–(b); the present status of the filing scores in C-03 only (§3.7(b)).

c. Founder succession. A generational or estate succession shown in the record (e.g., transfer to a co-principal of record) is not a principal-continuity break for 8.2(c). d. Seasonal dormancy. Filing reinstatements following seasonal dormancy in markets where the record shows this as a recurring administrative pattern are scored as one administrative event, not serial refiling — but only where licenses remained continuous through the dormancy. e. Multiple entities, one operation. Where one operating business runs parallel entities (e.g., a license-holding entity and a contracting entity), resolution under §12.3 designates the evaluated entity and the affiliated entity's record is read for continuity purposes only.

8.6 Minimum data and currency

C-02 requires a resolvable filing or license origin date; an entity with neither is unresolvable under EX-3 (§2.4) rather than zero-scored here. Entities younger than twelve months are handled by the 0.0–3.0 anchor range or, where the overall record is too thin to score other categories, excluded as EX-4 and re-queued (§2.4). Filing-history retrievals must be no older than the tranche collection window (§12.5).

8.7 Design note

C-02 is deliberately a slow-moving category: an entity cannot meaningfully change its tenure profile inside a cycle, in either direction. That is the point. The categories that can move quickly (C-01 status, C-03 standing, C-04 outcomes, C-05 findings) are monitored and re-scored on the twelve-month rhythm of §5; C-02 supplies the low-frequency ballast that distinguishes a durable operator having a bad quarter from a transient operator having a good one.

Section 9

Category Specifier C-03 — State Filing & Standing

Category Specifier C-03 — State Filing & Standing

State Filing & Standing C-03

9.1 Definition and scope

C-03 measures the entity's present orderliness as a legal person: whether the organization exists in good standing where formed, is properly qualified where it operates, maintains a reachable registered agent, and keeps its periodic filings current. These are the records that determine whether a homeowner, supplier, or court can actually find and proceed against the entity — the legal plumbing beneath every contract the entity signs.

C-03 expressly excludes filing history and tenure (scored in C-02 under §3.7(b)) and licensure (C-01). C-03 is a present-state category: it asks what the filing record shows on the evaluation date.

9.2 Measured indicators

IND. INDICATOR WHAT THE RECORD MUST SHOW

9.2(a) Formation-state Active / good-standing status (or jurisdictional equivalent) with the formation standing state's filing authority.

9.2(b) Foreign Qualification to transact in each operating state outside the formation state, where qualification that state's law requires it for the entity's activity.

9.2(c) Registered agent A registered agent of record at a resolvable address in each filing jurisdiction; no agent-resignation pending without replacement.

9.2(d) Periodic filings Annual/biennial reports and franchise filings current; no delinquency flags on the record.

9.2(e) Identifier integrity Entity identifiers (filing number, EIN-linked registrations where public, license cross-references) consistent across jurisdictions under §12.3.

9.3 Scoring rubric

Adjustments. Deduct 0.5 per additional cured delinquency beyond the anchor's allowance (max −1.0); deduct 0.5 where identifiers conflict across jurisdictions in a manner RECORDED but resolved under §12.3 (unresolved conflicts are EX-3 matters, not deductions).

9.4 Source hierarchy

Precedence: (i) the formation state's Secretary of State (or corporations authority) record; (ii) each operating state's foreign-qualification record; (iii) the licensing roster's entity cross-references, used to corroborate identity only. Third-party aggregations of filing data are collection conveniences and SHALL NOT be scored against; every scored C-03 fact must trace to the issuing authority's own record (§12.2).

9.5 Edge cases and interpretive rules

a. Sole proprietorships. Where the entity is an unincorporated sole proprietorship, 9.2(a)–(c) are scored against the jurisdiction's applicable analogues (trade-name/DBA registration and its currency); where the jurisdiction imposes none, those indicators are satisfied by definition and C-03 is carried by 9.2(d)–(e) as applicable. The Standard does not require incorporation. b. States without "good standing." Where a jurisdiction's terminology differs (e.g., "active," "current," "subsisting"), the jurisdiction's own best-standing designation is the 9.2(a) benchmark. c. Grace-period filings. A periodic filing within its statutory grace period and not yet flagged delinquent is current. d. Qualification ambiguity. Whether an operating state "requires" qualification under 9.2(b) is read from that state's published doing-business standard as applied to roofing contracting (physical operations, local crews, local contracts). Marginal cases are resolved in the entity's favor and RECORDED with the basis. e. Mid-evaluation cures. C-03 is scored as of the evaluation date (§1.5 D-7). A cure after that date is not scored in the current evaluation but is a classic RR-2/RR-3 ground for prompt re-review (§5.3).

9.6 Minimum data and currency

C-03 is scoreable whenever the formation jurisdiction is identified; an entity with no locatable filing or proprietorship record in any jurisdiction is an EX-3 identity matter. Filing-status retrievals must be no older than sixty days at the evaluation date — tighter than C-01's ninety, because filing status is the cheapest record for an entity to let silently decay. C-03 is registry-monitored for RR-2 purposes (§5.3).

9.7 Design note

C-03 carries the smallest weight of the five categories (15%) precisely because it is the most binary and the most curable: most entities are either simply current or simply not, and a delinquent filing is a morning's errand to fix. Its inclusion as a full category nonetheless matters — an operator who will not keep a $50 annual report current is broadcasting something about how it will handle a warranty claim — but the Standard declines to let inexpensive paperwork outvote conduct, per the rationale of §3.6(b).

Section 10

Category Specifier C-04 — Customer Outcome Data

Category Specifier C-04 — Customer Outcome Data

Customer Outcome Data C-04

10.1 Definition and scope

C-04 measures accumulated customer outcomes as evidenced by third-party consumer platforms: the normalized aggregate rating, the volume and recency of the review base, the distribution of outcomes, and the entity's pattern of engagement with adverse outcomes. It is the only category scored from third-party data rather than government record, and the Data Verification Protocol applies its strictest provenance rules here (§12.4).

C-04 expressly excludes formal complaint records and government findings (C-05), and any testimonial, reference, or review furnished by the entity itself — only platform-published data resolvable to the entity under §12.3 is scoreable.

10.2 Measured indicators

IND. INDICATOR WHAT THE DATA MUST SHOW

10.2(a) Normalized The volume-weighted mean rating across all Source Registry platforms on which aggregate rating the entity has a resolvable profile, normalized to a 5.0 scale per §10.3.

10.2(b) Review volume Total review count across platforms, applied as the confidence input of §10.3 — volume scales confidence in the rating; it is not independently rewarded.

10.2(c) Recency mix Share of the review base dated within the look-back; a rating carried by stale reviews is discounted per §10.3(c).

10.2(d) Distribution shape Concentration of one-star outcomes and bimodality of the distribution, as a check on means that hide a substantial dissatisfied cohort.

10.2(e) Adverse-outcome Whether the platform record shows substantive responses to negative reviews engagement across the look-back (responsiveness, not sentiment, is measured).

10.3 Aggregation method

The normalized aggregate rating R used by 10.2(a) — and by disqualifier DQ-4 (§4.2) — is computed as follows:

a. Normalization. Each platform's displayed rating is mapped to a 5.0 scale (ten-point platforms divide by two; letter or tier systems use the platform's own published numeric equivalence; platforms without one are excluded from R and RECORDED). b. Volume weighting. R is the mean of per-platform normalized ratings weighted by each platform's review count, so that a 4.9 across 6 reviews cannot outvote a 4.1 across 300. c. Recency discount. Reviews older than the 36-month look-back contribute at half weight where the platform exposes review dates; where it does not, the platform's aggregate is used as published and the limitation RECORDED. d. Shrinkage at low volume. Below sixty aggregate reviews, R is shrunk toward the neutral prior of 4.2 in proportion to the shortfall, so thin records score conservatively in both directions rather than spiking on a handful of outcomes.

10.4 Scoring rubric

Adjustments. Deduct 0.5 where 10.2(e) shows no engagement with any negative outcome across the look- back (entities with no negative outcomes are unaffected); deduct 0.5 where distribution bimodality under 10.2(d) exceeds the threshold RECORDED in the evaluation file's worksheet; add 0.5 (ceiling 10.0) where every platform with 20+ reviews independently sits at or above 4.7.

10.5 Minimum data rule

C-04 requires at least ten aggregate reviews across at least one resolvable platform profile to produce a sub-score. Below that floor the category is unscored, and the entity is handled as EX-4 (insufficient record, §2.4): excluded from scored finding for the tranche and automatically re-queued as the record accumulates. The Standard deliberately refuses to accredit on outcome silence — an entity with no measurable customer record has not failed the Standard; it has not yet generated the record the Standard measures.

10.6 Integrity rules

a. Profile resolution. A platform profile is scoreable only when resolved to the evaluated entity under §12.3 (matching name, location footprint, and license or filing cross-references where the platform exposes them). Same-name profiles in other markets are excluded and the exclusion RECORDED. b. Manipulation indicators. Where the verification process RECORDS platform-flagged review removal, clustered same-day review bursts inconsistent with job volume, or platform sanctions for solicited-review violations, the affected platform's data is excluded from R for the evaluation and the basis RECORDED. Exclusion of a platform under this rule is a data-integrity act, not a sanction; the remaining platforms score normally. c. No entity submissions. Per Rule R-1, screenshots, testimonials, or review exports furnished by the entity SHALL NOT be scored; only direct platform retrievals enter the file.

10.7 Source hierarchy

All Source Registry consumer platforms (§15, classes SR-08 through SR-12) are peers under the volume- weighting of §10.3 — no platform is privileged by brand. Where one platform's displayed aggregate conflicts with its own enumerable review data, the enumerable data governs and the discrepancy is RECORDED. Manufacturer credential programs (SR-13) are not outcome data and do not enter R; they are corroborative context RECORDED in the file only.

10.8 Design note

C-04's machinery — volume weighting, shrinkage, recency discounting, distribution checks — exists to make a noisy data class behave like evidence. The category neither worships a 5.0 built on six reviews nor lets one aggrieved outlier sink three hundred satisfied outcomes; it scores the shape of the accumulated record, which is the closest public proxy available for the question a buyer actually has.

Section 11

Category Specifier C-05 — Complaint & Public Findings

Category Specifier C-05 — Complaint & Public Findings

Complaint & Public Findings C-05

11.1 Definition and scope

C-05 measures the entity's formal adverse record: complaint-organization files and their resolution, government consumer-protection actions, and adverse public findings over the look-back. Where C-04 reads the broad statistical surface of customer experience, C-05 reads the narrow set of grievances serious enough to enter a formal channel — and, critically, what happened next. A complaint answered and resolved is a materially different fact from a complaint ignored, and the rubric scores that difference explicitly.

C-05 expressly excludes license-board discipline (scored in C-01 per §3.7(a)) and ordinary negative reviews (C-04). The DQ-3 grade threshold (§4.2) sits atop this category per the layering principle of §4.5.

11.2 Measured indicators

IND. INDICATOR WHAT THE RECORD MUST SHOW

11.2(a) Complaint- The entity's published BBB rating (where a profile exists), below the DQ-3 organization grade threshold: A-range versus B-range grades.

11.2(b) Complaint volume, Formal complaints filed over the look-back, read against the entity's operating normalized scale as evidenced by review volume and footprint (raw counts are not compared across differently sized operators).

11.2(c) Resolution pattern Share of complaints answered, and share closed as resolved or answered with good-faith response, per the organization's published dispositions.

11.2(d) Government Attorney-general consumer-protection actions, consent orders, or final consumer actions judgments of record involving the entity's roofing operations.

11.2(e) Pattern advisories Published pattern-of-complaint advisories or alerts attached to the entity's complaint-organization profile.

11.3 Scoring rubric

Adjustments. Add 0.5 (ceiling 10.0) where every look-back complaint shows resolution accepted by the complainant; deduct 0.5 where 11.2(b) normalized volume exceeds twice the rate typical of the entity's scale band as RECORDED in the worksheet; deduct 0.5 for each additional unanswered complaint beyond the anchor's allowance (max −1.5).

11.4 Source hierarchy

Precedence: (i) the complaint organization's own published profile and dispositions (SR-07); (ii) state attorney-general action records and court judgments of record (SR-05, SR-06); (iii) platform-published complaint threads, which corroborate but never independently establish a formal complaint. A complaint described in a review but absent from any formal channel is a C-04 fact, not a C-05 fact.

11.5 Edge cases and interpretive rules

a. Single-counting with C-01. License-board accusations, citations, and discipline score exclusively in §7.2(e) and are excluded here, per §3.7(a) — even where the complaint organization's profile references them. b. Settled actions. A government action resolved by settlement without admission is scored at the 4.0 anchor's "completed consent terms" treatment, by disposition shown of record; the Standard scores dispositions, not allegations. c. No-profile entities. An entity with no complaint-organization profile is scored from 11.2(b)–(e) alone, with 11.2(a) treated as the neutral baseline stated in the 10.0 anchor. Absence of a profile is not an adverse fact (and is not DQ-3, per §4.2); plenty of sound small operators have simply never been profiled. d. Complainant identity. Complaints filed by competitors or vendors rather than customers, where the organization's record so identifies them, are RECORDED but scored at half weight in 11.2(b)–(c). e. Withdrawn complaints. A complaint withdrawn by the complainant before response is RECORDED and excluded from 11.2(b)–(c). f. Storm-event clustering. Where a complaint cluster coincides with a declared catastrophic weather event in the entity's market, the cluster is scored as filed, but the worksheet RECORDS the event context for committee visibility under §14.3 — context travels with the score; it does not alter it.

11.6 Minimum data and currency

C-05 has no volume floor: an empty formal record is itself the finding, and it scores at the top anchors. Complaint-organization retrievals must be no older than sixty days at the evaluation date; government-action searches follow the tranche collection window. C-05's grade component is registry-monitored for RR-2 purposes, since a grade movement across the DQ-3 boundary is a standing event under §5.5.

11.7 Design note

The rubric's center of gravity is responsiveness. Complaint volume scales with operating volume, and a large operator with many jobs will accrue formal grievances that a two-crew shop never will; what the record can fairly compare across scale is whether grievances are met or ignored. Hence normalization in 11.2(b), the answered/unanswered architecture of the anchors, and the half-point reward for complainant-accepted resolutions: the category is built to find the operators who treat a formal complaint as work to finish, and to find the ones who treat it as noise.

Section 12

Data Verification Protocol

12.1 Purpose of the Protocol

Stage two of every evaluation (§3.1) exists to convert retrieved data into verified record. The Protocol governs how each data point is retrieved, attributed, logged, and — where it fails verification — kept out of scoring. Its operating rule is absolute: each data point is independently verified, and anything that cannot be verified is not used. There is no provision anywhere in this Standard for scoring an unverified fact, in the entity's favor or against it.

12.2 Provenance requirement

A scoreable data point SHALL be retrieved from the issuing source itself — the licensing board's roster, the Secretary of State's record system, the federal database, the platform's published profile — not from an aggregator, reseller, cache, or summary. Aggregated datasets MAY be used to locate records and to build the roster of §2.2, but every fact that enters a worksheet must be re-retrieved from its issuing source and RECORDED with: source identifier (per §15 class codes), retrieval date and time, retrieval method, and the matched identity basis of §12.3.

12.3 Entity resolution

Every record is attributed to the evaluated entity by at least two independent matching keys drawn from: legal name; registered trade name; license number; state filing number; registered-agent linkage; principal names of record; and address footprint. Single-key matches (a bare name match, most dangerously) SHALL NOT be scored. Where keys conflict — two operators sharing a trade name in one market, a license number pointing to a different filing — the conflict is escalated as an identity question; if unresolved after examination of the underlying records, the entity is excluded under EX-3 (§2.4) rather than scored on a contaminated file. Identity claims raised by entities under §4.4 are examined under this clause.

12.4 Class-specific verification rules

RECORD CLASS VERIFICATION RULE

Government rosters & filings Direct retrieval from the authority's system of record; status fields captured verbatim; screenshots or response payloads preserved in the file.

Federal databases Query parameters RECORDED with results, including null results — a clean SAM.gov or OSHA search is itself a verified fact and is logged as one.

Complaint organizations Profile retrieved and resolved under §12.3; grade, complaint counts, and dispositions captured as published, with capture date.

Consumer platforms Strictest tier: profile resolution under §12.3, rating and count capture per platform, manipulation screening under §10.6(b), and exclusion logging where data is set aside.

12.5 Collection windows and the evaluation date

Each evaluation tranche operates a collection window not exceeding thirty days. All retrievals for an entity occur within one window, and the evaluation date (D-7) is the date the window closes for that entity. Class- specific maximum ages tighten this further where stated (ninety days for license rosters, §7.7; sixty days for filing status and complaint-organization data, §9.6 and §11.6). A retrieval older than its class maximum at window close is re-retrieved before scoring or the affected indicator is treated as unverified.

12.6 Verification failures and corrections

A verification failure discovered after a finding issues — a mis-attributed record, a misread status field, a stale retrieval scored as current — is handled as follows: (a) the error is RECORDED in the evaluation file with discovery date and basis; (b) where the error is material (capable of changing a sub-score anchor, the composite band, the threshold determination, or a disqualifier determination), standing moves to suspended and corrective re-review RR-4 runs on the corrected record; (c) where immaterial, the file is corrected and annotated without standing change; and (d) the registry record carries a correction annotation for material corrections per §6.6. Errors are corrected in whichever direction they cut; the Protocol is symmetric.

12.7 The evaluation file and retention

The evaluation file (D-8) preserves, for each evaluation: every retrieved record or response payload; the verification log with sources, dates, methods, and matching keys; the per-category worksheets showing indicator values, anchor selection, and adjustments; the computation output of §13; the disqualifier screens of §4.1; and the committee finding record of §14.5. Files are retained for no fewer than seven years from the evaluation date, across supersession, so that any finding ever displayed in the registry can be reconstructed from its file.

12.8 Audit reproducibility

The test of a compliant file is reproducibility: a reviewer with access to the file and to this Standard, but with no access to the original evaluator, must be able to re-derive every sub-score and the composite from the preserved records alone. Files failing the reproducibility test in periodic internal audit are treated as containing a material verification failure under §12.6, whatever the finding's direction.

Section 13

Scoring Computation, Rounding, and Reconciliation Rules

Scoring Computation, Rounding, and Reconciliation Rules

13.1 Computation pipeline

Scoring is mechanical by design. For each category: (a) indicator values are read from the verified worksheet;

• the rubric anchor is selected top-down per the Specifier; (c) stated adjustments are applied; (d) the result

is clamped to 0.0–10.0 and carried at one-decimal precision as the sub-score. The composite is then computed under §3.3 at full available precision from the five sub-scores. No step of the pipeline accepts a discretionary input; every number in it traces to a worksheet cell, and every worksheet cell traces to a RECORDED record.

13.2 Rounding rules

a. Sub-scores are expressed to one decimal place. Rubric anchors and adjustments are already one-decimal quantities; no rounding arises within a category. b. The composite is computed unrounded, then rounded once, at the final step, to one decimal place by the round-half-up convention: 8.4500 rounds to 8.5; 8.4499 rounds to 8.4.

c. Intermediate values SHALL NOT be rounded. The §3.5 example illustrates: contributions are carried at three decimals and summed before the single final rounding. d. The threshold test of §3.4 is applied to the rounded composite. The published score and the tested score are therefore always the same number — a reader of the registry never confronts an entity "really" at 8.46 displayed as 8.5; 8.46 is 8.5 under this Standard, by published rule.

13.3 Reconciliation of conflicting records

Where two verified records of the same class conflict (e.g., two state rosters describing one license differently), reconciliation follows the affected Specifier's source hierarchy; where the hierarchy does not resolve it, the conflict is reconciled by recency of the records' own update stamps; where still unresolved, the indicator is scored on the less favorable reading and the conflict RECORDED for committee visibility — the Standard does not award the benefit of a contradiction the entity is best positioned to cure in the public record.

13.4 Determinism and dual computation

The composite for every finding is computed twice, by independent implementations of §3.3 and §13.2 operating on the same worksheet (in current operations, independent computation scripts; the requirement is implementation-independence, not any particular tooling). A mismatch at any precision blocks the finding until traced and resolved. Both computation outputs are preserved in the evaluation file.

13.5 Computation errors

A computation error discovered after issuance — a weight misapplied, an anchor misread from the worksheet, a rounding misapplied — is a material error per se and follows §12.6(b): suspension and corrective re-review RR-4 on the corrected computation. Because the pipeline is deterministic, computation errors are the rarest and most fully correctable error class; the file preserves everything needed to re-run them exactly.

13.6 Boundary cases

a. Exact threshold. A rounded composite of exactly 8.5 meets the threshold. There is no super-threshold review tier; 8.5 is 8.5. b. Clamping. Where adjustments would carry a sub-score beyond 10.0 or below 0.0, the clamp of §13.1(d) applies before the composite; clamping events are RECORDED. c. Unscored categories. There is no provision for re-weighting around a missing category. If any category cannot be scored under its minimum-data rule, the entity is excluded under EX-4 (§2.4) — a four-category composite does not exist under this Standard. d. Ties and comparisons. Composites are not rank-compared between entities for any purpose under this Standard (§3.8); no tie-breaking rule exists because no ranking exists.

13.7 Publication precision

The registry publishes the rounded composite only. Unrounded values, worksheet contents, and per- category sub-scores reside in the evaluation file and the entity's evaluation report; publishing a second, higher-precision figure alongside the official one would create exactly the dual-number ambiguity the single- rounding rule of §13.2 exists to prevent.

Section 14

Committee Finding Procedures

14.1 Role of the Committee in stage four

The Standards Committee closes every evaluation by issuing the finding (§3.1, stage four). The Committee's function at this stage is deliberately narrow: it confirms procedure, it does not re-score. The Committee verifies that the file before it shows a compliant evaluation — and then issues the finding that the computed outputs dictate. The design intent is stated plainly: a body that could adjust scores would become the natural target of every pressure this Standard exists to exclude.

14.2 The finding checklist

For each entity, the Committee confirms on the record:

i. coverage and eligibility under §2, including any exclusion basis; ii. collection-window and currency compliance under §12.5; iii. entity resolution adequacy under §12.3 (two-key minimum, no unresolved conflicts); iv. both disqualifier screens of §4.1 completed, with results; v. worksheet completeness for all five categories, including minimum-data rules; vi. dual-computation agreement under §13.4; and vii. the threshold determination of §3.4 as applied to the rounded composite.

14.3 What the Committee sees

The Committee receives computed outputs and the verification log — sub-scores, the composite, screen results, and RECORDED annotations (including context notes such as §11.5(f)) — together with access to the underlying file. It does not receive, and the file does not contain, any channel for entity-submitted material, per Rule R-1. Annotations inform the Committee's procedural confirmation; they do not authorize departure from the computed result.

14.4 Limits of Committee authority

In issuing findings the Committee SHALL NOT: waive or adjust the 8.5 threshold for any entity (§3.4); alter any weight, anchor, or adjustment within a Cycle (§3.10, §4.6); substitute its judgment for a computed sub- score; or issue a finding on a file failing any checklist item of §14.2. Where a file fails the checklist, the Committee's sole instruments are remand (the file returns to the failed stage for correction and re- presentation) and, for identity failures, exclusion under EX-3. The Committee's discretionary authority lives elsewhere: in interpretations under §16.3 and in edition revisions under §16.2 — both prospective, both published, neither available to move an individual pending finding.

14.5 The finding record

Each finding is RECORDED with: the entity and ARCO ID; the finding category (D-9); the composite and band where published; the checklist confirmations of §14.2; the date of the finding; and the identities of the Committee members confirming it. The finding record joins the evaluation file under §12.7, and the public verification record updates per §6.2. Findings take effect on registry publication, not on internal recording (§6.2's publication-first rule).

14.6 Separation of evaluation and membership

Restating §1.4's third consequence as committee procedure: no person with responsibilities in membership sales, dues, renewals, or member services participates in any stage-four confirmation, and no membership or revenue information appears in any evaluation file or finding record. The finding pipeline runs to completion identically whether or not an accredited entity ever activates membership; activation is a post-finding commercial act with no evaluative content.

14.7 Conflicts of interest

A Committee member with a financial interest in, family relationship to, or prior business relationship with an evaluated entity is recused from that entity's confirmation, and the recusal is RECORDED in the finding record. A member's general interest in ARCO's success is not a recusable conflict — it is the ordinary condition of every standards body — which is precisely why §14.4 strips the Committee of the discretion that such an interest could otherwise reach.

14.8 Cadence

The Committee confirms findings on a rolling cadence matched to evaluation tranches, such that no compliant file awaits confirmation beyond the registry's normal publication cycle. Backlogs are an operational matter; they SHALL NOT be relieved by abbreviating the checklist.

Section 15

Source Registry

15.1 Purpose and structure

This section enumerates the record classes from which evaluation data is drawn. Each class carries a registry code (SR-01 onward), the issuing-source rule of §12.2 as applied to that class, and the categories it feeds. The registry defines classes, not vendors: a class names the kind of record and its authoritative custodian, and remains stable even where a custodian renames a product or changes an access channel. The operational source list — the specific endpoints, portals, and retrieval methods current in a Cycle — is maintained under §16 and conforms to this section.

15.2 Government record classes

15.3 Third-party and aggregate record classes

15.4 Manufacturer credential records

Manufacturer credentials are commercial program memberships, not regulatory facts. They appear on the public record because homeowners encounter them and benefit from seeing them verified against the manufacturer's own directory — but they carry no evaluative weight under this Standard.

15.5 Class rules

a. Provenance. Every record in every class is retrieved from the custodian named for that class, per §12.2. A licensing fact reported by an aggregator is not an SR-02 record; only the board's record is. b. Coverage variation. Not every class yields a record for every entity. Absence handling is governed by the category specifiers (§§7–11), not by this section; this section only fixes where a record, if it exists, must come from. c. Class additions. New record classes may be added only by edition revision under §16.2. Within a Cycle, the registry is closed: no evaluation may draw on a class not enumerated here. d. Custodian succession. Where a custodian is renamed, merged, or succeeded (e.g., platform consolidation), the class persists and the successor custodian is RECORDED in the operational source list without an edition change, provided the record kind is unchanged. e. Single-counting. Where one underlying event surfaces in multiple classes (an AG action reported in SR-05 and docketed in SR-06), the single-counting rule of §3.7 governs: the event is scored once, in the category whose specifier most directly addresses it.

15.6 Retrieval integrity

Each retrieval from each class is logged under §12.6 with class code, custodian, retrieval date, and method. Where a custodian's public channel is unavailable during a collection window, the retrieval is re-attempted within the window; a class that cannot be retrieved for an entity within the window is RECORDED as unretrieved and handled under the relevant specifier's absence rules — never substituted from a secondary source.

Section 16

Standard Maintenance, Governance, and Bylaws References

Standard Maintenance and Governance

16.1 Authoring body

This Standard is authored, maintained, and interpreted by the ARCO Standards Committee, constituted under Section 4.2 of the ARCO Bylaws. The Committee's revision authority and interpretation authority derive from that section; nothing in this Standard enlarges them.

16.2 Edition revision

a. Cadence. The Standard is reviewed annually and revised, where revised, on a Cycle boundary. Editions are designated by year and sequence (this edition: 2026.1). A mid-year edition (e.g., 2026.2) issues only where a defect or external change makes continued application of the current edition unsound. b. Prospectivity. A new edition applies to evaluations whose collection windows open on or after its effective date. No edition is applied retroactively to a closed finding; re-reviews apply the edition current at the re-review (§5.3). c. What revision can touch. Weights, anchors, adjustments, disqualifier definitions, thresholds, record classes, and procedures — all revisable by edition, none revisable within a Cycle (§3.10, §4.6, §15.5(c)). d. Publication. Each edition is published in full at roofingcontractors.org before its effective date, with its revision history per §18.

16.3 Interpretations

Where applying this Standard to a fact pattern requires construction the text does not settle, the Committee issues a written interpretation. Interpretations are published, numbered, prospective, and binding on subsequent evaluations within the Cycle; they are folded into the next edition or extinguished by it. An interpretation SHALL NOT change a weight, anchor, threshold, or disqualifier — those are revision matters under §16.2 — and SHALL NOT be issued to direct the outcome of an identified pending file (§14.4).

16.4 Independence of maintenance

Per §1.4 and §14.6, no revenue, membership, or sales consideration enters revision or interpretation deliberations, and no evaluated entity, member or otherwise, holds any procedural channel into them. Public comment on a published edition may be received and considered; receipt creates no obligation and no entity-specific process.

16.5 Custody of the operational source list

The Committee maintains the operational source list implementing §15 — current endpoints, portals, and retrieval methods per record class — as a controlled internal document, updated as custodians change channels and RECORDED with each change. The list conforms to the registry of §15; it cannot add a class.

Section 17

Document Control, Citation, and Use of Findings

Document Control, Citation, and Use of Findings

17.1 Controlled document

This document is controlled under reference ARCO-STD-2026.1. The controlled copy is the edition published at roofingcontractors.org; any printed or downloaded copy is uncontrolled and is superseded by the published edition. Document control fields — version, effective date, reference, authoring body — appear on the cover and in the document profile (p. 4).

17.2 Citation

Cite this Standard as: ARCO Standard 2026.1, American Roofing Contractors Organization, effective January 1, 2026 — and cite provisions by section and clause (e.g., "ARCO Standard 2026.1 §10.4"). A citation to a clause carries the whole Standard with it; clauses are not severable into freestanding rules.

17.3 Use of findings by accredited entities

An accredited entity in current standing may state its accreditation, display the ARCO mark per its activation terms, and link to its public verification record. Every permitted statement reduces to one verifiable claim: evaluated against ARCO Standard 2026.1, composite at or above 8.5, no automatic disqualifier, verifiable at roofingcontractors.org.

17.4 Misuse

a. Stating or implying accreditation while in lapsed, suspended, or withdrawn standing is misuse; the public verification record controls over any conflicting claim (§6.4). b. Stating or implying that accreditation constitutes a warranty, a guarantee of workmanship, an endorsement of future performance, or a government license is misuse (§1.3). c. Reproducing the rounded composite alongside invented precision, invented rank, or comparative claims against named competitors is misuse (§3.8, §13.7). d. Verified misuse is grounds for suspension of standing pending re-review, and for withdrawal under §5.4 where the misuse is confirmed.

17.5 Historical-state requests

The registry publishes current standing and the most recent finding (§6.2). Requests for the historical state of a record as of a past date — from insurers, attorneys, or courts — are answered from the evaluation file under §12.7 as a written extract identifying the edition then in force, the finding then current, and the standing on the requested date. Extracts are RECORDED and carry the same controlled-document notice as §17.1.

17.6 No reliance beyond stated scope

Findings are issued for the purpose stated in §1.2 and within the limits of §1.3. A party extending a finding beyond that scope — as underwriting advice, as a substitute for due diligence, as a prediction — does so on its own judgment. The Standard's claim is exact and deliberately narrow; its value depends on never exceeding it.

Section 18

Revision History

18.1 Edition record

This table is the permanent edition record of the Standard. Each subsequent edition adds a row identifying its designation, effective date, status, and the material actions taken — additions, retirements, and any change to a weight, anchor, adjustment, threshold, disqualifier definition, or record class — so that any finding can be read against the full text of the edition under which it was issued (§17.5). Rows are never amended or removed once published. When a successor edition takes effect, the prior edition's status is marked Archived; archived editions remain in this record for historical reference and are not applied to evaluations whose collection windows open under a later edition (§16.2(b)).

18.2 Composition of this edition

As the current edition, 2026.1 fixes the complete normative content of the Standard for Cycle 2026:

a. The five category specifiers of §§7–11, with published rubric anchors at every scoring band and a defined look-back period, minimum-data rule, and absence-handling rule for each category. b. The scoring computation rules of §13, including the single-rounding rule and dual independent computation. c. The data verification protocol of §12, including two-key entity resolution and the thirty-day collection window. d. The four automatic disqualifiers of §4 in their complete trigger, verification, and clearance definitions. e. The license-regime normalization map of §7.5, covering all fifty states under the four-regime model. f. The document-control, citation, and use-of-findings provisions of §17, including the misuse provisions of §17.4.

Every weight, anchor, adjustment, threshold, and disqualifier definition stated in this edition is fixed for the entirety of Cycle 2026 per §3.10 and §4.6; the first occasion on which any of them can change is a successor edition under §16.2.

18.3 Interpretation log

No interpretations under §16.3 are outstanding against this edition as of its effective date. Interpretations issued during Cycle 2026, if any, will be published at roofingcontractors.org and consolidated into the subsequent edition's revision history.

18.4 Next scheduled review

The annual review of this Standard for Cycle 2027 concludes no later than December 1, 2026, with any successor edition published before its effective date per §16.2(d). Absent a published successor, this edition remains in force.

End of Document

ARCO Standard 2026.1 — the published standard for roofing contractor accreditation in the United States. Eighteen sections, five weighted categories, four automatic disqualifiers, composite threshold 8.5 / 10.0. Authoring body: ARCO Standards Committee · Effective January 1, 2026 · DOC REF ARCO-STD-2026.1 Verify any contractor's accreditation at roofingcontractors.org.

AMERICAN ROOFING CONTRACTORS ORGANIZATION