GBC Carrier Score — Methodology & Formula
The GBC Carrier Score is a 0–100 composite rating per carrier (or per carrier × insurance line) that answers the whole buyer question: "is this a good carrier to actually buy from?" — not just "what does it cost." It blends three independent data pillars, every one of them publicly verifiable. This page publishes the full formula so the score is transparent + citable.
The formula (v1)
+ 0.30 × PriceScore
+ 0.30 × ServiceScore
Weights reflect what buyers actually care about: financial strength is the largest single factor (will they pay your claim?), with price and service tied for second. Weights are tunable and any future revision will be reflected here + recomputed across all carrier records.
The three pillars
1. StrengthScore (40% weight)
Source: AM Best financial strength rating (primary) or S&P rating (fallback/cross-check). Ratings are publicly accessible at web.ambest.com. AM Best letter is mapped to a 0–100 score:
| AM Best Rating | StrengthScore |
|---|---|
| A++ (Superior) | 100 |
| A+ (Superior) | 95 |
| A (Excellent) | 88 |
| A- (Excellent) | 82 |
| B++ (Good) | 72 |
| B+ (Good) | 65 |
| B (Fair) | 55 |
| B- (Fair) | 45 |
| Below B- or Not Rated (NR) | 30 |
2. PriceScore (30% weight)
Source: GBC filed-rate data + observed quote data. Each carrier's rate for a given insurance line is rank-percentiled against the market for that line. Cheapest decile = 100; median = 50; most expensive decile = 0. Per-line scoring means a carrier can score well on one line and poorly on another — which is honest, because that's how the market actually works.
Filed-rate data is sourced from state Department of Insurance filings via SERFF Filing Access. See our Insurance Rate Changes Tracker for the live feed.
3. ServiceScore (30% weight)
Source: NAIC Complaint Index, publicly available at content.naic.org. The index normalizes complaint volume against market share: an index of 1.0 means the carrier received complaints in proportion to its market share (median); >1.0 means more complaints than peers; <1.0 means fewer.
The mapping:
Examples: index 1.0 → score ~100; index 2.0 → score 50; index 0.5 → score 100 (capped).
How we handle missing data (the integrity rule)
Per the GBC integrity rule (same standard as our primary methodology page): we never invent a value to fill a missing pillar. Three possible states for any carrier:
- All 3 pillars available — full score computed, displayed with no caveat.
- Partial data (1 or 2 of 3 pillars available) — score is computed from available pillars only, with available weights renormalized to preserve the 0-100 scale. Display includes a visible "based on N of 3 pillars" note.
- No data / pending verification — no score displayed; renders a "Verification in progress" placeholder until primary-source data is captured.
Structural limitations of the PriceScore pillar in specialty markets
The PriceScore pillar is calculated as a percentile of each carrier's filed or observed rate vs. market. In typical commercial lines (workers comp, commercial auto, general liability), every carrier's rates are public via state DOI rate-filing portals — making percentile comparison straightforward.
In specialty markets, this isn't always the case. The equine pilot vertical surfaced three structural exemptions that make per-carrier price comparison structurally impossible from public records:
- Excess & Surplus carriers are exempt from rate filing. Westchester Surplus Lines (Chubb Limited subsidiary) writes the majority of high-value equine in the U.S. As a non-admitted E&S carrier, Westchester is statutorily exempt from rate-filing requirements in nearly every state (confirmed via NY ELANY 2022 manual, CA, TX, FL, KY). There is no public SERFF rate filing to compare.
- Inland Marine equine specialty filings are often "informational-only." Equine mortality is classified as inland marine (not workers comp or commercial auto). In approximately 35 states, inland marine specialty filings are filed as informational disclosures (rates filed but not publicly disclosed) under use-and-file rules. Markel — the largest U.S. equine specialty writer — files this way.
- Specialty MGA programs use manuscript / per-risk rates, not class-rated. Hartford writes equine through a dedicated "Hartford Livestock Department" with agent-only distribution; rates are negotiated per-risk against an underwriting schedule, not class-rated like workers comp. Same pattern at most equine specialty programs.
What this means for the GBC Score: for the 5-carrier equine pilot, only 1 carrier (American Equine Insurance Group, underwritten by Argonaut Insurance Company) publishes a verifiable per-policy rate ($350 annual base for tier-1 Care/Custody/Control, AEIG form ELP-APP111-1018). The other 4 carriers' rates cannot be primary-source compared, so the PriceScore pillar renders as "structurally unavailable" rather than a fabricated percentile. We disclose this honestly rather than guess.
Implication for other GBC Score pilot verticals: lines with admitted-market bureau filings (NCCI WC, ISO Commercial Auto loss costs) will have full PriceScore coverage; lines dominated by E&S or specialty MGA programs (cyber, EPLI, professional liability for niche verticals) may face similar transparency limits and will be disclosed similarly.
Editorial independence (non-negotiable)
Carriers can never pay to raise their GBC Score. The moment any score is buyable, the whole metric is worthless. GBC monetizes via lead delivery, not score manipulation. The AM Best rating + NAIC complaint index are publicly verifiable for any reader to cross-check; the GBC price percentile is derived from the same SERFF filings published on the Insurance Rate Changes Tracker.
When scores update
Scores are recomputed as the underlying inputs change:
- StrengthScore — recomputed when AM Best publishes a rating change (typically quarterly per-carrier).
- PriceScore — recomputed when new SERFF filings are captured into our data layer (continuous).
- ServiceScore — recomputed when NAIC publishes annual complaint-index updates (annually).
Each score change is preserved in our history table so trend content ("Carrier X's GBC Score fell 8 points after 2026 rate hikes") can be sourced back to the underlying input change.
Formula versioning
The formula above is v1. Future revisions (different weights, additional pillars, per-line nuances) will be versioned here. Old scores remain queryable under their original formula version for fair historical comparison.
Spec source: SEOexpert SPEC-gbc-carrier-score-2026-06-07.md.
Implementation: FastBusinessQuote.Infrastructure.Services.GbcScoreService.
Carrier inputs: FastBusinessQuote.Domain.Data.GbcCarrierScoreData.