Fair Isaac Corporation: The Three-Digit Fee
The Fee You Cannot Negotiate
At the closing table for a four-hundred-thousand-dollar mortgage, among the appraisals, title searches, and attorney fees, sits a line item for roughly fifty dollars. This is the tri-merge credit report, a consolidated snapshot of a borrower’s history from Equifax, Experian, and TransUnion. Buried inside that report is a three-digit number generated by Fair Isaac Corporation, FICO, that will determine the interest rate, the monthly payment, and ultimately the profitability of the loan. The cost to produce that number is less than one dollar. The price the lender pays for it, passed through two intermediaries and ultimately folded into the consumer’s closing costs, is imperceptible in the context of the transaction. This is not a bug in the pricing architecture. It is the entire business.
FICO’s economic engine runs on a three-layer distribution model that insulates it from every form of price sensitivity. The company licenses its mathematical models to the three national credit bureaus, which account for roughly seventy-five percent of scores revenue. The bureaus embed these scores into their reports, marking them up approximately one hundred percent without resistance because they neither bear the cost of the underlying computation nor compete on the basis of score pricing. Banks purchase the reports, absorb the cost as a trivial operational expense, and pass it to consumers who are psychologically and contractually unprepared to negotiate a fifty-dollar line item in a six-figure transaction. The result is an operating margin above eighty-five percent on a product that costs pennies to deliver, with incremental margins on recent price increases exceeding ninety-five percent. FICO does not gather consumer data, maintain data centers full of personally identifiable information, or manage the regulatory burden of credit files. It merely sells the algorithm that turns others’ data into a standard. The business is, in essence, a royalty on financial anxiety, collected at a remove that renders the toll invisible.
The Empty Balance Sheet
The most durable structural advantage FICO possesses is not its brand recognition or its regulatory entrenchment, though both matter. It is the absence of a consumer data infrastructure. While Equifax, TransUnion, and Experian maintain vast warehouses of credit histories, ssets that require billions in capital expenditure, attract intense regulatory scrutiny from the Consumer Financial Protection Bureau, and carry the catastrophic liability of data breaches, ICO retains nothing. It produces only the model; the bureaus provide the input. When Equifax suffered its 2017 breach, exposing the records of 147 million Americans, the company eventually booked costs exceeding $1.4 billion for remediation, legal settlements, and security upgrades. FICO, whose scores were generated using the very same compromised data, bore none of these costs. Its balance sheet carries no toxic data assets.
This asset-lightness is not merely an efficiency. It is a competitive moat that compounds over time. Any competitor seeking to displace FICO must either build its own data infrastructure, a proposition requiring years of regulatory licensing and capital deployment, or license data from the bureaus, which places them at the mercy of incumbents who can adjust pricing or access terms. FICO’s cash conversion is therefore structurally superior to that of credit bureaus or fintech lenders, who must reinvest heavily in data security and compliance. The eighty-five percent operating margins in scores are not the result of operational genius but of a deliberate architectural choice made decades ago to remain a pure intellectual property company. The cash generated, hundreds of millions annually with minimal reinvestment required, funds the expansion of FICO’s software division while insulating the core business from cyclical pressure. It is a rentier position in the financial plumbing, maintained by never touching the water.
The Twenty-Five-Year Silence
For twenty-five years, from 1991 to 2016, FICO did not raise prices. This was not a failure of nerve but a strategic hibernation, allowing the FICO score to become the unchallenged standard for mortgage securitization, auto lending, and credit card origination. Then, beginning in 2018, leadership initiated a program of annual thirty percent price increases, rotating across mortgage, auto, and credit card categories. Within three years, this added approximately sixty million dollars in annual revenue, flowing through at ninety-five percent incremental margins. The decision represents one of the most aggressive exercises of latent monopoly power in recent financial services history.
The governance implications are stark. Management treated the scores business as a convertible bond, uration risk embedded in regulatory and technological stability, suddenly called for redemption. The pricing power was always there; what changed was the willingness to deploy it. This shift transforms FICO from a passive utility into an active extractor of economic rent. But it also creates a political target. The fact that prices remained flat for a generation and then accelerated sharply suggests to regulators and legislators that the company was previously restraining itself not out of competitive necessity, but out of prudence. The Federal Housing Finance Agency’s 2021 decision to require both FICO 10T and VantageScore 4.0 for Fannie Mae and Freddie Mac mortgages, and simultaneously to reduce the required credit reports from three bureaus to two, cut FICO’s addressable mortgage revenue by one-third overnight. It was a signal that regulatory forbearance has limits, and that pricing aggression invites intervention. The leadership now navigates a narrow ledge: maximize extraction before the political window closes, without triggering price caps or mandatory licensing regimes that would permanently impair the franchise.
The Standard
FICO’s dominance rests not on network effects in the traditional sense, but on a coordination problem embedded in the architecture of securitization. In 1995, Fannie Mae and Freddie Mac institutionalized the FICO score as the required risk metric for all mortgages they purchase, effectively making it the collateral standard for the $11 trillion mortgage-backed securities market. This created a self-reinforcing equilibrium: investors price FICO-based pools with confidence, issuers must use FICO to access the deepest capital markets, and deviating from the standard imposes a liquidity penalty that outweighs any cost savings.
The case of Synchrony Financial illustrates the bind. In 2021, after a three-year transition, Synchrony, America’s largest private-label credit card issuer, migrated from FICO to VantageScore for its consumer credit decisions. The annual savings totaled $2 to $3 million on a revenue base of $18 billion, a negligible gain. However, Synchrony lost the ability to securitize its receivables at the most favorable rates, because investors demanded the FICO-based risk calibration they understood. The switch proved pyrrhic; the company sacrificed funding efficiency for a rounding error in cost. This dynamic protects FICO from direct competition, but it also exposes a fragility. The monopoly depends entirely on the continued preference of securitization markets for FICO’s specific risk stratification. Should regulators or investors ever accept an alternative standard, particularly one derived from artificial intelligence with demonstrably superior predictive power, the coordination problem dissolves, and the lock-in evaporates.
There are early warnings. The Gini coefficient, a measure of predictive accuracy, has plateaued at approximately 0.70 for FICO models for fifteen years, including versions 8, 9, and the newer 10. VantageScore 3.0 and 4.0 cluster at the same level. Despite FICO’s marketing of FICO 10 as a ten-percent improvement, adoption has been glacial; the majority of scores delivered in 2021 were still generated by FICO 8, a model launched in 2009. Lenders see insufficient marginal benefit to justify the operational cost of migration. This stagnation suggests that traditional credit scoring has hit a ceiling of extractable signal from conventional data. If an AI-native competitor, Upstart, ZestAI, or a yet-unamed entrant, demonstrates a Gini coefficient above 0.75 in a specific vertical, the rationale for adhering to FICO weakens. The securitization standard persists only so long as it remains the most predictive option available.
The Cyclical Subsidy
The scores business is cyclical. During the 2007-2009 financial crisis, revenue from scores declined by one-third as mortgage originations collapsed. Yet the company’s strategic future depends on software, fraud detection, customer management, and originations platforms, that requires sustained, heavy investment regardless of the credit cycle. FICO’s Falcon system currently protects approximately two-thirds of global credit card transactions from fraud; its TRIAD platform manages two-thirds of the world’s credit card accounts. These are market-leading positions, but they are under assault from fragmented competition, Pegasystems, SAS, IBM, Moody’s, Verisk, and from the threat of AI-driven disruption that could alter the competitive landscape unpredictably.
Software currently generates half of FICO’s $1.3 billion revenue but only twenty-five percent of operating income, with margins hovering near thirty percent compared to the mid-eighties in scores. The new FICO Platform, a unified architecture for software applications, is growing at forty to sixty percent annually and now represents twenty percent of software recurring revenue. The transition is working, but it is expensive. The cash flow from scores must fund this expansion while protecting against cyclical troughs. Over-invest in software during a recession, and returns crater; under-invest now, and AI competitors capture the next generation of risk management infrastructure. The leadership’s capital allocation decisions, how much of the scores extraction to reinvest versus return to shareholders, will determine whether FICO emerges as a diversified software company or remains a melting ice cube of credit scoring royalties.
The Free Score Trap
In 2013, FICO launched the Score Open Access program, allowing banks and credit card issuers to provide free FICO scores to their customers. The strategy builds ecosystem lock-in: by saturating the market with free FICO scores, the company prevents VantageScore from gaining cognitive foothold among consumers. If borrowers know their FICO score, they demand FICO scores from lenders; if lenders provide VantageScore, consumers perceive a lesser product. The program has been effective at marginalizing the competitor, but it carries a self-inflicted wound.
Approximately one-third of scores revenue comes from business-to-consumer channels, myFICO subscriptions and partnership royalties, which are not cyclical and command high margins. If free distribution through banks becomes the norm, the B2C business faces structural decline. Conversely, if FICO restricts free access to protect myFICO revenue, it cedes ground to VantageScore, which could then achieve the consumer recognition necessary to challenge FICO’s institutional position. The leadership is attempting to thread a needle: provide enough free scores to maintain standard-setting power, while preserving sufficient scarcity to sustain direct consumer monetization. This is a balance that grows more precarious as consumer financial literacy increases and regulatory pressure mounts for transparent, no-cost credit information.
The Predictive Ceiling
The ultimate threat to FICO is not VantageScore, nor regulatory pricing caps, nor even the cyclicality of mortgage markets. It is the possibility that credit scoring, as a mathematical discipline, has been solved as far as traditional data permits, and that artificial intelligence can now extract materially superior prediction from alternative data sources, cash flow, employment history, behavioral biometrics, that FICO does not possess and cannot easily incorporate.
FICO’s models are generalized, designed to function across the entire credit spectrum using bureau data alone. AI competitors can train narrow, vertical-specific algorithms, specialized for subprime auto, or thin-file millennials, or gig economy workers, that achieve significantly higher accuracy within those niches. If a lender can demonstrate to securitization markets that its AI-derived risk assessments reduce default rates by fifteen percent compared to FICO 10, the collective action problem that protects FICO begins to fracture. Investors will eventually accept new standards if the yield improvement is sufficient.
The company’s reliance on the Gini coefficient as its defensive moat is therefore a temporal gamble. The plateau at 0.70 has held for fifteen years, suggesting diminishing returns from iterative model improvements. FICO 10 and 10T have not driven a mass migration away from FICO 8 because the marginal gain is insufficient to justify operational disruption. But an external technological discontinuity, an AI breakthrough in predictive accuracy, would render the switching cost logic obsolete. The securitization standard is only as durable as its predictive supremacy.
The Narrow Path
FICO stands at the intersection of three narrowing corridors. In the first, management must extract maximum pricing from the scores business, thirty percent annual increases, rotated across product lines, before regulatory bodies impose caps or mandate open-access scoring standards that eliminate the B2C revenue stream. In the second, they must deploy that cash aggressively into software, building the FICO Platform into a genuine standard for fraud detection and customer management before AI-native competitors fragment the market or surpass FICO’s predictive capabilities in high-value verticals. In the third, they must hope that the Gini ceiling remains intact, that no competitor achieves a step-change in risk prediction that breaks the securitization coordination equilibrium.
The business is not merely a monopoly; it is a timed monopoly, its value leaking gradually toward the moment when either politics or algorithms render the FICO score a commodity. The eighty-five percent margins and the asset-light structure provide the cash and the flexibility to attempt a transition. But the cyclical vulnerability of the scores business means that a recession, a 2008-style collapse in mortgage originations, could arrive precisely when software investment is most needed, forcing a choice between defending current returns and funding future relevance. The leadership’s decision to begin aggressive price extraction in 2018 suggests they recognize the window is finite. They are racing to convert regulatory inertia and distribution opacity into software scale before the mathematics of credit risk moves beyond their models. Whether they succeed depends on whether the predictive ceiling holds long enough for the software engine to fire.
The Regulatory Fracture
The Federal Housing Finance Agency’s 2021 decision arrived without fanfare but with surgical precision. By mandating that Fannie Mae and Freddie Mac accept both FICO 10T and VantageScore 4.0 for mortgage originations, and simultaneously reducing the required credit reports from three bureaus to two, regulators accomplished in a single administrative action what decades of competitive pressure could not: they reduced FICO’s addressable revenue in its largest single vertical by one-third. The mechanism was subtle: reduction in the tri-merge requirement meant that lenders now needed only two bureau reports rather than three, eliminating one slot in the distribution chain where FICO had previously collected a toll. The credit bureaus, suddenly competing for two positions rather than three, began negotiating on price for the remaining slots, compressing margins throughout the distribution layer.
This was not merely a revenue event. It was a demonstration of structural vulnerability. For thirty years, the three-bureau system had functioned as a cartel of distribution, with FICO comfortably licensing to all three without fear of exclusion. The reduction to two bureaus introduces a zero-sum dynamic that did not previously exist. If a lender chooses Equifax and TransUnion, Experian receives nothing; FICO loses one-third of its volume for that origination. The bureaus, facing existential pressure to secure one of the two available slots, are now incentivized to resist FICO’s thirty-percent annual price increases or to accelerate investment in proprietary scoring models that could replace FICO entirely. Equifax, Experian, and TransUnion have long maintained VantageScore as a joint venture precisely for this contingency: hedge against FICO’s pricing power that now becomes a live weapon in a battle for shelf space.
The FHFA decision reveals that FICO’s regulatory protection is reversible. The 1995 requirement that Fannie and Freddie use FICO scores was an administrative choice that created the standard; the 2021 decision to admit VantageScore alongside FICO demonstrates that administrative choices can dilute it. The regulatory moat flows both ways. If the FHFA can mandate dual-score acceptance to promote competition, it can mandate tri-merge restoration if competition proves chaotic, or it can mandate price caps if FICO’s thirty-percent annual increases attract Congressional scrutiny. The political economy of mortgage lending, where every basis point of cost is scrutinized by consumer advocates and where “junk fees” have become a bipartisan target, creates a ceiling on extraction that FICO is rapidly approaching. The company has raised prices aggressively since 2018 precisely because the window for such extraction is visible from the outside; regulators have already demonstrated the capacity to restructure the distribution model.
The Free Distribution Paradox
In 2013, FICO launched the Score Open Access program, allowing over 170 financial institutions, including nine of the ten largest U.S. credit card issuers, to provide free FICO scores to their customers. The strategic logic was elegant: by saturating the consumer market with free FICO scores, the company would prevent VantageScore from achieving cognitive recognition. If a borrower checks their Capital One app and sees a FICO Score 8 of 720, they internalize that number as their “real” credit score. When that same borrower applies for a mortgage and the lender uses VantageScore 3.0, the resulting discrepancy, often twenty to forty points, creates friction and confusion. The borrower demands to know why their “real” 720 has become a 690, pressuring the lender to standardize on FICO. The program converts consumer mindshare into institutional lock-in.
However, this strategy creates a direct conflict with FICO’s business-to-consumer revenue stream, which accounts for approximately one-third of scores revenue and generates margins comparable to the core licensing business. The myFICO subscription service, where consumers pay $19.95 to $39.95 monthly for continuous score monitoring, credit reports, and identity theft insurance, depends on scarcity. If free scores become ubiquitous through banking apps, the value proposition for myFICO erodes. Why pay forty dollars monthly for a FICO score when your Chase app updates it weekly at no cost?
The tension manifests in partnership negotiations. FICO must calibrate the availability of free scores precisely: sufficient to maintain standard-setting power against VantageScore, but insufficient to cannibalize the high-margin B2C subscription base. Recent years have seen FICO restrict certain score versions from the Open Access program, ICO Score 9 and 10 remain largely unavailable through free channels, reserved for myFICO subscribers and premium lender tiers. This creates a bifurcated market where consumers see free FICO 8 scores (a model launched in 2009) while lenders increasingly adopt FICO 10T for mortgage decisions. The disconnect generates the very confusion the program was designed to prevent, potentially accelerating regulatory pressure for standardized, free disclosure of the scores actually used in lending decisions.
If Congress or the CFPB mandates that lenders provide the specific score used in the credit decision, rather than any educational score, at no cost to the consumer, the B2C business collapses. This is not speculative; Senator Elizabeth Warren has previously introduced legislation requiring free access to credit scores, and the CFPB under Director Rohit Chopra has signaled aggressive scrutiny of “rent-seeking” in credit reporting. FICO’s Open Access program is a preemptive strike against such regulation, but it is also a slow-motion cannibalization of the only non-cyclical revenue stream in the scores business. The leadership walks a tightrope: give away too much, and myFICO revenue declines; give away too little, and VantageScore captures consumer mindshare or regulators mandate universal free access.
The Alternative Data Frontier
The plateau in the Gini coefficient, roughly 0.70 for fifteen years across all major models, suggests that traditional credit bureau data has reached the limits of predictive utility. FICO 10 and 10T incorporate trended data, examining payment patterns over twenty-four months rather than static snapshots, but the improvement remains marginal, approximately ten percent. This stagnation creates an opening for competitors using alternative data sources that FICO cannot easily access.
Consider Upstart Holdings, which originates personal loans using an AI model that incorporates over 1,600 variables, including education, employment history, and banking transaction data, nformation not contained in traditional credit files. In specific verticals, particularly thin-file borrowers with limited credit history, Upstart’s model demonstrates significantly lower loss rates than traditional FICO-based underwriting. While Upstart currently uses FICO scores as one input among many, the direction of travel is clear: as AI models train on cash-flow data and behavioral signals, the marginal contribution of the FICO score diminishes.
The threat is not that Upstart replaces FICO entirely, but that it proves the Gini ceiling is breakable. If a major lender, Capital One, Discover, or a fintech like SoFi, demonstrates that an AI-native model achieves a 0.75 Gini coefficient in credit card originations, the justification for FICO’s securitization standard weakens. Investors in asset-backed securities care about loss rates, not methodological purity. If a pool of loans underwritten without FICO scores performs better than FICO-underwritten pools, the coordination equilibrium that protects FICO begins to unravel. The switching costs that currently prevent migration, operational integration, legacy system compatibility, investor familiarity, become surmountable if the predictive gap widens sufficiently.
FICO’s response, the FICO XD score and UltraFICO, attempts to incorporate alternative data, utility payments, property records, and bank account history, but these products remain niche, capturing only the unscorable population rather than redefining the mainstream. The structural problem is data ownership. FICO does not possess the banking transaction data or the employment records; it must license them or rely on lenders to provide them. The credit bureaus, meanwhile, are expanding their own alternative data capabilities, potentially rendering FICO’s models obsolete by providing raw data directly to lenders who build their own AI tools.
The Governance of Extraction
The decision to initiate thirty-percent annual price increases in 2018, after twenty-five years of price stability, represents a fundamental shift in corporate strategy that can only be understood as a capital allocation choice by leadership. Under CEO Will Lansing, who took the helm in 2012, FICO transitioned from a company that prized market share and standard-setting ubiquity to one that maximizes near-term extraction from a mature monopoly. This was not inevitable. The company could have continued the low-price strategy, preserving regulatory goodwill and fostering ecosystem growth while slowly building software revenue. Instead, leadership chose to monetize the regulatory moat before it erodes.
This decision reveals management’s internal assessment of the business’s durability. They do not believe the thirty-percent annual increases are sustainable indefinitely; otherwise, they would have implemented them gradually over decades. The aggression suggests a view that the window for such pricing power is finite, perhaps ten to fifteen years, before regulatory intervention or technological disruption closes it. The cash generated is not being returned entirely to shareholders through buybacks, though the company maintains an active repurchase program. Instead, it funds the FICO Platform, a $300 million-plus investment in cloud architecture and AI capabilities designed to transform the software division from a collection of legacy mainframe products (Falcon, TRIAD) into a modern, integrated platform.
The economic variable here is time preference. Lansing’s team is effectively betting that the scores business can fund a complete transformation of the company before the scores business itself declines. If the software platform achieves fifty-percent operating margins and becomes the standard for fraud detection and customer management, FICO becomes a diversified software company with a declining but high-margin legacy business: model analogous to Oracle or IBM in their respective transitions. If the platform fails to achieve scale before the scores business faces regulatory price caps or AI displacement, FICO becomes a melting ice cube of extraction, liquidating a monopoly for cash until it expires.
The cyclical risk compounds this tension. The 2007-2009 financial crisis reduced scores revenue by one-third. A similar recession in 2024 or 2025 would arrive precisely when the software division requires sustained investment to compete with AI-native entrants. Management would face the choice of cutting software R&D to protect margins, ceding the future, or accepting depressed returns and potential activist investor pressure. The governance structure, with a classified board and concentrated institutional ownership (Vanguard, BlackRock, and State Street collectively hold over twenty-five percent), provides some protection against short-term pressure, but not immunity. The extraction strategy is a bet that the cycle will hold long enough for the transition to complete.
The Bear Case: Fragmentation
The most credible threat to FICO is not a single competitor but a fragmentation of the market into vertical-specific standards. In auto lending, lenders may adopt AI models that incorporate vehicle telematics and dealership behavior data. In credit cards, cash-flow underwriting may supplement or replace FICO for subprime segments. In mortgages, VantageScore or proprietary bureau models may capture share through regulatory mandate. FICO becomes not the standard, but one of many standards, losing the coordination advantage that sustains its pricing power.
In this scenario, the three-layer distribution model becomes a liability. Credit bureaus, facing margin pressure from the reduction to two-report requirements, develop their own scoring models for specific verticals, xperian already offers proprietary scores for auto lenders that compete directly with FICO Auto Score. If Equifax and TransUnion follow suit, FICO’s seventy-five percent revenue concentration in bureau licensing becomes a vulnerability. The bureaus control the data pipeline; if they choose to prioritize their own models or to offer VantageScore as the default in tri-merge reports, FICO becomes a premium option rather than a requirement.
The fragmentation bear case gains force from the AI transition. Machine learning models require continuous retraining on recent data; they age quickly as consumer behavior shifts. FICO’s traditional model of releasing new scores every five to seven years (FICO 8 in 2009, FICO 9 in 2014, FICO 10 in 2020) may prove too slow for a market where competitors update algorithms monthly. The FICO Platform is designed to address this, offering real-time model deployment, but it requires customers to migrate from legacy systems: process that has historically taken years. FICO 8 remains the dominant model in 2024, fifteen years after its release, because lenders see insufficient benefit to justify the operational cost of migration. If AI competitors offer continuous improvement without requiring costly infrastructure overhauls, FICO’s installed base becomes a millstone rather than a moat.
The Endgame
The business now operates in a state of managed decline disguised as growth. The thirty-percent annual price increases generate the cash flow necessary to fund transformation, but each increase accelerates the regulatory and competitive pressures that will eventually cap or reverse pricing. The software platform must scale rapidly enough to offset the inevitable slowing of scores revenue, yet the cyclicality of the scores business means that precisely when software investment is most needed, during a credit downturn, the cash flow to fund it may contract by thirty percent or more.
The structural advantages remain formidable. The securitization standard, the three-layer distribution insulation, and the asset-light model provide a runway that few competitors can match. FICO will likely remain the dominant risk metric for prime mortgage lending for another decade, regardless of AI advances, because the coordination costs of switching the $11 trillion MBS market are prohibitive. But the margin structure is already under pressure from the FHFA’s dual-score mandate, and the pricing power that seemed absolute in 2018 now faces visible ceilings.
What emerges is a company extracting maximum value from a regulatory standard while racing to build a new business that can survive the standard’s eventual erosion. The leadership’s decision-making, aggressive pricing, heavy software investment, strategic free-score distribution, reflects a clear-eyed view that the current monopoly is temporal. The question is whether the transition can complete before the extraction strategy triggers a regulatory response that permanently impairs the pricing architecture, or before an AI competitor demonstrates that the Gini plateau was not a ceiling but a complacency. The cash flows are real, the margins are exceptional, and the competitive position is dominant. But the window is visible, and it is closing.
The mechanics of the two-bureau squeeze deserve closer scrutiny. Under the previous tri-merge regime, Equifax, Experian, and TransUnion operated as a comfortable oligopoly, each collecting tolls on one-third of mortgage originations with little incentive to undercut one another on FICO pricing. The shift to dual-bureau reports transforms this stable triumvirate into a brutal auction. For any given mortgage, one bureau will be excluded entirely, losing not merely the FICO licensing revenue but the entire report fee, pproximately fifty dollars that flows directly to their top line. The desperation to avoid being the odd bureau out creates defection pressure. If TransUnion offers a proprietary credit score that undercuts FICO’s thirty-percent annual increase, or bundles VantageScore at a discount to capture the slot, Equifax must match or risk exclusion. The coordination that maintained FICO’s pricing power for decades collapses into a race to the bottom.
This dynamic is already visible in auto lending, where lenders frequently use single-bureau reports, forcing bureaus to compete aggressively on price for exclusive placement. The mortgage market, previously insulated from such competition by the tri-merge requirement, now faces similar fragmentation. As bureaus invest in their own scoring algorithms to avoid FICO’s escalating tolls, FICO’s seventy-five percent revenue concentration becomes a concentrated vulnerability. The company has no alternative distribution channel; it cannot sell scores directly to lenders without alienating the bureaus that currently comprise its primary revenue stream. The three-layer model, once the foundation of pricing power, becomes a strategic noose when the middle layer turns hostile. FICO may find itself forced to choose between margin compression in its largest vertical or the gradual erosion of its monopoly as bureaus commoditize the score itself.