Fair Isaac Corporation

FICO owns the credit-score toll every US lender must pass through to underwrite. The FICO Score is embedded in ~90% of US lending decisions and is the de-facto mandated score for conforming mortgages (Fannie Mae / Freddie Mac). The bottleneck is not code — it is the regulatory + switching-cost lock: a lender replacing FICO faces Federal Reserve model-risk re-validation (SR 11-7), GSE approval, securitization-investor acceptance, and Fair Credit Reporting Act compliance. The proof the toll is real: FICO raised its per-score wholesale royalty from $0.60 to $10.00 over five years and doubled it from $4.95 to $10.00 in 2026 — a ~17× cumulative price increase with no loss of share. (Senator Hawley letter, March 2026.)

Not a patent moat — a regulatory-embedded network moat. The FICO Score methodology is not protected by a single blocking patent; the durability comes from:

• GSE mandate — FHFA requires a credit score for conforming mortgages; FICO 10T is approved (with VantageScore 4.0, the one rival, as the counterweight).
• SR 11-7 model governance — any replacement score must survive full model-risk validation before a regulated bank can deploy it.
• Securitization lock-in — mortgage-backed-security investors underwrite on FICO score distributions; a switch re-prices the entire secondary market.
• Bureau distribution — FICO scores are resold through the three consumer reporting agencies (Equifax, Experian, TransUnion), which collectively are 63% of FICO revenue (each >10%) — simultaneously the moat's distribution rail and its single biggest channel risk.

• bypass_test_score = 3 — lenders cannot route around (GSE mandate, ~90% lending share, 17× price power)
• concentration_test_score = 2 — dominant duopoly, not single-source: VantageScore 4.0 is FHFA-approved; the 3 bureaus (63% of revenue) co-own the rival
• hyperscaler_exposure_score = 3 — no hyperscaler path into a regulated, GSE-approved, SR 11-7-governed credit score
• Composite 3 + 2 + 3 = 8 → DIRECT_BOTTLENECK

• product_service: FICO Score credit-risk score plus FICO decision-management/analytics software platform.
• technical function: Scores convert bureau credit data into a 3-digit statistical/ML risk score; software automates originations, fraud, customer management via rules, predictive models, optimization.
• why it is hard: Embedded in GSE mortgage underwriting, securitization, regulatory capital; bureau data/validation history; high model/software switching costs; IP/trade secrets. [10-K FY2024, Item 1, Item 1A].
• substitutes: VantageScore, bureau proprietary scores, lender internal models, cloud-native decision engines. [10-K FY2024, Item 1A].
• pricing power: HIGH — FICO Score is required in GSE underwriting and widely used in lending; price increases have historically stuck. [10-K FY2024, Item 1A].

FICO sells two broad types of products:

1. Credit scores and related scoring products. The best-known product is the FICO Score, a three-digit number that rank-orders consumer credit risk. FICO has multiple score versions, including FICO Score 8, FICO Score 9, FICO Score 10, FICO Score 10 T, and industry-specific scores for auto lending and bankcard underwriting.

1. Decision-management software. The modern product is FICO Platform, which provides decisioning, analytics, and workflow capabilities. FICO also sells established applications:

• Falcon Fraud Manager — payment-card and transaction fraud detection
• TRIAD Customer Manager — account management and adaptive control
• Debt Manager — collections and recovery
• Origination Manager — originations and onboarding
• Blaze Advisor — business-rules and decisioning engine
• FICO Decision Modeler — model development and deployment

These product names are from FICO’s public disclosures and investor materials. For exact current product descriptions and segment categorization, use the FY2024 10-K and the annual earnings release:

Source: FICO Investor Relations

URL: fico.com

Access date: 2026-05-07

The FICO Score is built on credit-report data from the consumer reporting agencies. It converts historical credit behavior — payment history, amounts owed, length of credit history, new credit inquiries, credit mix — into a rank-ordered risk prediction. Later versions introduce trended data, which uses historical balance and payment trajectories rather than only a point-in-time snapshot.

The software products operationalize decisions. For example, a bank can use FICO Platform to approve a loan, set a credit limit, flag potential fraud, or route a delinquent account to the correct collections treatment. The decision engine applies predictive models and business rules inside a governed workflow.

Substitutes and competitive alternatives include:

• VantageScore, a competing credit score created by Equifax, Experian, and TransUnion
• In-house bank scorecards built by internal data-science teams
• Credit-bureau proprietary scores sold without the FICO brand
• Custom machine-learning/underwriting models built on alternative data
• Decisioning platforms from other analytics vendors
• Open-banking cash-flow underwriting models that bypass traditional credit-report scoring

The strongest substitution risk is in mortgage scoring because of regulatory and government-sponsored enterprise acceptance. FHFA, the GSE regulator, has announced that FICO Score 10 T and VantageScore 4.0 will be used for loans sold to Fannie Mae and Freddie Mac, subject to implementation milestones. This is a named-institution government source:

Source: Federal Housing Finance Agency

URL: fhfa.gov

Access date: 2026-05-07

This is structurally significant: FICO historically had deeply entrenched mortgage-score use through the GSEs, but FHFA’s inclusion of VantageScore is a real competitive opening.

FICO is hard to replicate for several reasons:

1. Regulatory and institutional embeddedness. Lenders, investors, and regulators have decades of infrastructure, internal policies, model-validation documentation, and regulatory exams built around FICO Scores.

1. GSE and mortgage-market use. Mortgage underwriting systems used by Fannie Mae, Freddie Mac, mortgage insurers, and originators have long incorporated specific FICO score versions.

1. Data ecosystem without owning the data. FICO does not need to replicate the credit bureaus’ data assets; it sits on top of the bureau data as the accepted algorithm layer.

1. Model governance and switching costs. Replacing a validated credit-risk score requires model risk management, regulatory justification, and operational change across many downstream systems.

1. Network effects at the lender-approval level. The more lenders use a score, the more consumers and originators orient around that score.

These are qualitative moat arguments. The quantitative strength must be tested against FICO’s pricing power and score volume in SEC filings.

FICO likely has high pricing power in Scores, but that is an analytical conclusion, not a verified primary-source fact. The working mechanism is simple: the score is a small input cost relative to the size of the loan, and it is hard to remove without changes to credit policy and model governance. Therefore, FICO should be able to raise per-score fees modestly without losing material share.

However, FICO must manage:

• Reseller dynamics with Equifax, Experian, and TransUnion
• Regulatory pressure if activists or policymakers push for lower-cost or government-provided scores
• VantageScore adoption, especially if mortgage-market share shifts
• Enterprise buyer concentration in the software business

Every quantitative pricing-power claim requires verification against FICO’s pricing disclosures, which are often not fully public. Exact price movement: needs primary source validation.

• What is the actual VantageScore 4.0 adoption rate at Fannie/Freddie as of the latest FHFA quarterly score report? (Determines whether concentration should drop to 1 next cycle.)
• Has the FTC formally opened an investigation into FICO's B2B score pricing (beyond Hawley's referral)?
• Does FICO's Q4 FY2026 guide show any volume response (not just price) to the $10.00 score — i.e., is there a demand-elasticity signal hiding in mortgage-origination volume?

STRENGTHENED. The toll-road thesis is confirmed by exceptional quantitative evidence: 90.8% Scores margin, +49% B2B / +97% mortgage-origination revenue on flat cost of revenue, and ~17× per-score pricing power with zero share loss. The DB scoring gap (ts/ec/ttm NULL, toll-road NULL) is now filled at 8/9/8 with toll-road 8 (DIRECT_BOTTLENECK). The disconfirming regulatory counter-pressure (Hawley/FTC, FHFA VantageScore 4.0) is material but has not yet dented the toll's monetization — it is the thesis's chief forward risk, not a current breach.

WEAKENED (from STRENGTHENED). The 08-29 note was right that the current-quarter P&L is a toll-road: 68% Scores mix, 91% Scores margin, 63% bureau concentration, 98% incremental flow-through, all 10-Q-cited. It was wrong that “regulatory counter-pressure … has not yet dented the toll’s monetization.” September 4 is the dent — not of Q3 numbers, of the forward GSE exclusive that those numbers were capitalizing. Moat quality WEAKENED (DIRECT BOTTLENECK 8 → STRONG PROXY 6); current-quarter fundamentals still print. Research-only; no live capital.

WEAKENED (re-affirmed). The 09-09 codification confirms the moat breach: the GSE exclusive is officially over, the currently-monetized model (Classic FICO) is slated for retirement, and the successor (10T) has no delivery date. FICO's Direct License Program is a competent defensive re-pricing (converts monopoly rent into a $4.95–$10/score performance toll) but concedes the "you have to buy FICO" frame. Moat quality WEAKENED (STRONG PROXY 6); current-quarter fundamentals still print. Research-only; no live capital.