Methodology

How we source, score, and refresh the data.

Every signal traces back to a public federal source. No proprietary “black box.” No data brokers. No scraped social media. This page documents the full pipeline so you can validate or challenge any score we produce.

Data Sources

RedMapAlert ingests exclusively from official U.S. federal government sources. Every alert and risk signal includes a citation linking back to the agency’s original publication.

DOJU.S. Department of Justice
Hourly

Press releases — indictments, pleas, convictions, settlements

https://www.justice.gov/news
SECU.S. Securities & Exchange Commission
Hourly

Litigation releases, administrative proceedings, enforcement orders

https://www.sec.gov/litigation
IRSInternal Revenue Service
Daily (with IRS publication lag)

Form 990 / 990-PF / 990-EZ filings, EIN registry, EOMF

https://www.irs.gov/charities-non-profits/tax-exempt-organization-search
FTCFederal Trade Commission
Hourly

Cases & proceedings — consumer protection, antitrust

https://www.ftc.gov/legal-library/browse/cases-proceedings
CFPBConsumer Financial Protection Bureau
Hourly

Enforcement actions against financial institutions

https://www.consumerfinance.gov/enforcement/actions/
CourtsFederal Court Dockets (via CourtListener / RECAP)
Daily

Civil and criminal docket metadata across all U.S. federal district, appellate, and bankruptcy courts. Metadata only — full PDFs available at the source.

https://www.courtlistener.com/
State AGsState Attorneys General — NY, CA, FL, TX, IL (top 5 by enforcement volume)
Hourly

Press releases from the top 5 state AG offices. Captures consumer-protection, securities, and AG-led civil suits filed before federal escalation.

https://ag.ny.gov/press-releases

Risk Score (0–100)

Each entity receives a composite score on a 0–100 scale, with higher values indicating greater observed risk. The composite is the sum of four sub-scores (each 0–25), every one derived from machine-readable federal sources:

0–25

Financial Opacity

Missing Form 990 schedules, vague disclosures, omitted self-dealing footnotes, restated financials.

  • Schedule L (Transactions with Interested Persons) absent or empty
  • Schedule O free-text answers shorter than 200 chars on material questions
  • Restated prior-year financials without explanation
  • Officer compensation reported only in aggregate (no per-officer breakout)
0–25

Power Concentration

Related-party transactions, officer loans (Schedule L), family officers, low board independence.

  • Officer loans outstanding (Schedule L Part II)
  • Independent voting members < 50% of total board (Form 990 Part VI)
  • Officer compensation > 10% of program expenses
  • Family/business relationships among officers/directors disclosed
0–25

Mission Efficiency (inverted)

Program-vs-overhead ratio — <40% program → 25 pts, 40–60% → 18 pts, 60–75% → 8 pts, >75% → 3 pts.

  • Program service expenses ÷ total functional expenses
  • Computed from IRS Form 990 Part IX, Column B vs Column A
  • Industry floor: 65% (per BBB Wise Giving Alliance standards)
0–25

Citizen-Report Corroboration

How strongly user-submitted observations align with anomalies independently surfaced from the 990 data.

  • Number of independent reports filed against the same EIN
  • AI-scored alignment between report text and 990 anomaly type
  • Time-decay weighted (recent reports > old reports)

Entity Snapshot Score (terminal & API)

The score shown on /entity/<ein> and returned by the public API uses a separate, deterministic weighting based on canonical federal flags rather than the 990-only rubric above:

Active sanctions hit (OFAC SDN, OpenSanctions, etc.)+50
SAM.gov federal exclusion (debarment)+35
Enforcement action on file (DOJ, SEC, FTC, CFPB, HHS-OIG)+25
IRS 501(c) status revoked+20
Per cross-signal event (capped at 20)+3

Score is capped at 100. Bands: 80+ CRITICAL, 60–79 HIGH, 40–59 MEDIUM, 1–39 LOW, 0 with no flags = CLEAN.

Calibration & limits

Weights are calibrated against historical enforcement outcomes from federal datasets. The score is an intelligence signal for due diligence, not a legal determination. A high score does not imply guilt or wrongdoing. A low score does not constitute a clean record. Always read the underlying citations and consult legal counsel for compliance decisions.

Cross-Source Detection Rules

These rules fire when independent federal feeds align on the same canonical entity. This pattern-recognition layer — not the raw data ingest — is the part of RedMapAlert that competitors cannot replicate by simply downloading IRS records. Rules run nightly and emit immutable cross-signal events on every match.

R1CRITICALSanctioned Active Nonprofit

An active 501(c) nonprofit appears on a global sanctions list. Possible diversion vehicle.

R2HIGHSAM-Excluded with Recent Revenue

Federally debarred entity continues to report revenue. Investigate funding sources.

R3HIGHEnforcement → Revenue Cliff

Significant revenue decline within 18 months of an enforcement action — operational impact confirmed.

R4CRITICALRevoked + Enforcement

501(c) status revoked AND a separate enforcement action on file — pattern of compliance failure.

R5CRITICALMulti-Source Sanctions

Listed by ≥2 independent sanctions sources (e.g., OFAC + EU + UK). High-confidence match.

Source: lib/detection/rules.ts. Rule definitions are auditable on request for enterprise customers.

Refresh Cadence

  • DOJ, SEC, FTC, CFPB — polled hourly. New press releases typically appear in the platform within 60–90 minutes.
  • IRS Form 990 — refreshed daily, but inherits the IRS’s own 12–18 month publication lag (filing window: up to 11 months post-fiscal-year + IRS digitization). RedMapAlert does not claim real-time 990 data. The “real-time” promise applies to enforcement alerts (DOJ, SEC, FTC, CFPB, OFAC, SAM) — your alert fires the same hour the agency publishes.
  • Risk scores — recomputed when any underlying signal updates. Reports always reflect the latest available federal data at generation time.
  • Source citations — every signal embeds a permanent URL to the original government record.

What This Is — and Isn’t

Is

  • • Federal enforcement intelligence
  • • A pre-engagement screening signal
  • • Citable, source-linked, auditable
  • • Updated continuously from public sources

Is not

  • • Legal advice or a legal determination
  • • A consumer report under the FCRA
  • • A sanctions/OFAC screening replacement
  • • A substitute for licensed background checks

Have a methodology question?

Compliance and risk teams can request a detailed methodology brief, including signal definitions, calibration data, and known limitations.

Request methodology brief →