Author: Ben Tewey
Date: 5/20/22
Share Price: $395

Overview

Fair Isaac Corporation, or FICO, was founded in 1956 with the sole mission of improving decision making through data. The company is now known for its FICO Score, a three digit score that rates a consumer’s propensity to repay debt. Banks, insurers, auto dealerships, credit card issuers and investors rely on FICO’s solutions to make informed lending decisions.

For framing, FICO has a $10bn market cap and did ~$1.3bn in revenue in 2021. The company has two reportable segments: Scores and Software which account for 50% of revenue each, however the scores segment has been growing at a much faster clip relative to software. In 2019, scores were only 36% of revenue. As scores revenue has become a larger portion of the pie, FICO’s margins have steadily improved due scores ridiculously high 86% operating margins. The company has compounded FCF at 14.8% since 2012.  

Thesis

I believe that the market is significantly over reacting to the FHFA decision and that the stock is currently in limbo until there is further clarity presenting a fantastic opportunity for long term investors. Scores is a phenomenal, monopoly-like business with spectacular margins and a long runway for continued price taking surrounded by strong trust and high switching costs. The market is also ignoring a fast-growing SaaS segment with a sticky customer base and growing recurring revenue. This segment has been underappreciated bby the market as a result of a lack of disclosures which has recently changed. As investors become comfortable analyzing this side of the business the price implied expectation should revise higher. My base case assumes 16% IRR from today’s price as a result of 8.4% growth in FCF, 3% of shares retired per year through buybacks and a 4% FCF yield, in line with historic average. Management is conservative, shareholder aligned (the CEO owns 1.38% of shares out, a significant portion of his net worth) and has demonstrated shareholder friendly capital allocation. 

Business Segments

Scores
“What would you like our margins to be? Because we can do that.”-Will Lansing CEO, Fair Isaac Corporation

Without a doubt every reader will have heard of the FICO Score. The 3-digit score, ranging from 300-850, that conveys the credit risk of an individual buyer and their propensity to repay debt. Indeed, this score has become the standard measure of consumer risk in the U.S. and is used in most U.S. credit decisions, by nearly all major banks, credit card issuers, mortgage lenders, and auto loan originators. 

The Scores segment has grown revenue at an 18.1% CAGR since 2016 (when FICO began raising prices for the first time in 25 years) and did over $650mn in revenue in 2021. Scores have almost no capital needs and segment operating margins have steadily expanded from 77% in 2016 to an absurd 86% in 2021. For the graph below I annualize Scores results from the H1. FICO CEO Will Lansing has taken advantage of the little incremental cost involved in running the algorithm to produce the scores and kept segment SG&A costs low by reducing headcount in the segment. Scores contribute just over 87% of operating income. 

FICO scores are generated by the company’s algorithm using the data from the three large consumer ratings agencies (CRAs) TransUnion,  Equifax and Experian (which account for >30% of total revenue) and then resold back to CRAs. FICO has also recently begun to grow their consumer facing business selling scores directly from their website, myFico.com, where consumers can get a full report, including an explanation of factors affecting their score.  FICO focusing more on this segment recently is an obvious strategic play: they are trying to reduce their dependence on the rating agencies by building their brand with consumers directly. 

Scores are used across the credit lifecycle, from origination (determining whether or not to execute a loan and if so at what rate),  account and risk management and consumer marketing. The Scores are necessary because they offer a common language for various organizations to communicate about the credit risks of individuals and loan portfolios. In a way, they are similar to the SAT or any other standardized test: they provide a common language for quick communication.

FICO earns royalties (hence the incredibly high margins) on each score the rating agencies generate using the FICO algorithm. These scores (distilled versions of the data they encompass) are then sold to the large money-center banks such as Bank of America. In fact, 90 of the top 100 lenders use FICO scores.

The purpose of the FICO score is incredibly simple: provide a simplistic view of a borrower’s propensity to repay in a cost effective manner. Scores can cost anywhere from only a few cents all the way up to a dollar  to get, but provide a generally sufficient overview of an individual’s creditworthiness for their cost. I estimate that the average score cost ~4.95 cents based on Will Lansing, FICO’s CEO, comments about issuing 14bn scores and TTM scores revenue of $694mn. 

In the grand scheme of things the cost of a FICO score is a rounding error. Scores are usually sought by banks deciding whether or not to extend credit to an individual buying a house, car or any other large item. These loans are often tens of thousands or hundreds of thousands of dollars and the banks care very little about tacking on an additional dollar (which is usually passed through to the consumer anyways in closing costs) for additional information on credit worthiness. Quite literally if the scores provide any information whatsoever they are worth buying because they cost so little. 

Investors also use scores to appraise the creditworthiness of various asset backed securities. FICO is the the benchmark these consumer-based securitizations with 98%+ share

in cards, mortgage and auto. This creates high brand loyalty and trust as well as substantial switching costs as investors will demand banks to provide the score to them. 

Management didn’t raise the price of the scores the 25 years following 1991. They only recently started taking price (initially to keep pace with inflation in 2016 and 2017) but have begun issuing special price increases since then. These 25 years of lag create a chasm between the price of the score and the value it provides. Lansing recently said at a Barclays conference last September “We don’t think about ceiling [in terms of price] we think about not shocking the system.” This decision has seen Scores segment revenue CAGR 26% since 2018. 

As I mentioned earlier, the FICO score is primarily used as a starting point. Not a be all end all,  and it isn’t designed to be. Banks often pull on their own data to supplement the scores to make more informed lending decisions. 

Currently, 232 million Americans are  FICO visible meaning they can get FICO scores (most of the distributed scores are FICO 8 and 9). The scores are calculated exactly how you think they would be: FICO bases the score on payment history, the amounts owed, length of credit , and the mix of your credit (credit card, mortgage, auto loan etc).

FICO has also recently been investing in R&D to expand scorable Americans through FICO XD (extra data). This is an onboarding score so consumers who have limited credit  history have an easier time moving to a more traditional score. The XD score draws on information such as utility, cable or cell phone bills. 

The end state of the consumer rating industry is likely to resemble what we have seen in the credit rating agencies. Moody’s and S&P can both effectively have 100% market share because their bond ratings are used as a supplement, not a replacement, of each other. Fintech solutions and in-house systems developed by banks are likely to be used in addition to FICO scores, not as a replacement of them because of the scores’ minimal cost in relation to the underlying loan (which even with continued price increases remains a rounding error). 

Growth is likely to be driven by moderate increases in volume of scores, however the primary driver will be untapped pricing runway. 

Software

The scores are no doubt what FICO is known for. However, the company has also been nurturing a fast-growing software business inside of it as well. With recent new disclosures around their FICO platform starting at the end of last year (FICO’s fiscal year ends in September) I believe the market is currently underappreciating this side of the business. The increased disclosure should close the gap between FICO’s current price and the business’ true value.  

Software does $660mn is a mid teens to low twenties operating margin business which should expand as upfront investment eases. 

Because of recent disclosure changes (FICO rearranged their reportable segments) I have attempted to reconstruct as best I can what I believe to be the Software segment pre-2019. My method was to subtract out “professional Services revenue” from both the Decision Management Software and Application segments to back into comparable numbers to today’s “On Premises and SaaS Software” segment. 

The large drop from 2020 to 2021 was a result of a change in revenue recognition where a substantial portion of revenues were moved from point in time recognition to being recognized over the life of the contract. FICO also sold the C&R business to Constellation in 2021 which accounted for 6% of segment revenues for $92.8mn.

FICO has been using profits from their Scores business to build the Software side for years now. The software has high upfront fixed costs but low incremental capital needs, highly recurring revenue and high switching costs leading to strong net retention. Once running, margins will expand as FICO eases investment through the income statement creating operating leverage. The beauty of the business model is more than reflected in the multiples applied to software businesses today, but FICO’s has gone largely unappreciated.  

The software business provides account origination, customer management, customer engagement, fraud detection, financial crimes compliance, and marketing software. The software is bought in multi-year subscriptions, with payments based on usage metrics such as the number of accounts, transactions or decisioning use cases deployed. 

To give a sense of the end users:  96 of the 100 largest financial institutions in the U.S., and ⅔ of the largest 100 banks in the world are customers. 600 insurers, 300 retailers and merchandisers and more than 200 government or public agencies. This is a serious business.

The software segment is really a story of two businesses. FICO is in the midst of a multi-year shift from their on-premise solution to their SaaS platform solution which allows for greater revenue visibility. Management has guided toward a 50%+ growth rate for the platform side for “the next few years” and the legacy business to be flat with the goal to move everything to the FICO platform eventually. This change can even be seen over the last few quarters as Platform ARR has moved from 9% to 18% of total ARR in just the last 8 quarters

Because FICO recently rearranged their reportable segments and added additional disclosures around ARR and DBNRR, some of these time frames are shorter than I’d prefer, however they still provide a solid image of the two businesses. 

Platform has been adding ARR at 40% year over year for the last two years and 10% QoQ since Q4’19. The cross selling opportunities on the platform have evidently been enormous as dollar based net retention ratio, a measure of how much existing customers are spending indexed to what they spent in the same period the previous year, has remained at about 115% since Q4’20. This indicates that customers have been increasing the amount of platform services they are using (which strengthens switching costs and causes them to spend more) and FICO has been able to pass through price increases without significant churn or customer attrition. 

On the other hand, the non-platform business has remained resilient with flat revenue and only slightly decreased DBNR over the last 8 quarters. This is surprising because of the strong results of the platform business which I thought would cannibalize non-platform much faster than has been the case. Over time this business should be flat to slightly decreasing as FICO transitions customers to the cloud platform. I don’t think business will die as fast as I originally thought which speaks to the switching costs this software has. In all likelihood, there will be some customers who do not want to transition to the cloud. Bureaucracy and uncertainty around the cloud’s security/peace of mind of an on-prem solution (no matter how flawed that may be) will keep at least some customers on-prem. It makes sense that FICO wouldn’t leave these customers behind. 

Growth within the company’s platform offerings accelerate as the demand for financial data  and analytics grows.This is similar to S&P’s Market Intelligence or Bloomberg in that it is the industry standard. Insurers writing a policy or banks looking to lend will use FICO’s Customer Acquisition Marketing, Originations, Customer Management or Credit Card Fraud Detection software to help in their decision making. Fair Isaac’s strongest franchises, customer management and fraud detection,  are the number 1 player in the category.

The last business segment, which CEO Will Lansing has stated FICO is in the process of deemphasizing is Professional Services. This business is low margin and does ~$100mn in revenue per year. It includes services such as software implementation, consulting, model development and training services for customers. This service won’t fully diminish (all SaaS companies have an on-boarding segment like it), but certainly will not be a revenue driver going forward. I have lumped its revenue and operating profit in with the software segment above. 

Competitive Threats and Risks

I like to segment the bear case for FICO into three different buckets: 1) FHFA decision (customer concentration at large) and general disruption 2) Economic slowdown 3) Tough Comps and the durability of the special pricing runway.

Bear Argument  #1- FHFA and General Disruption 

FICO’s share price has been punished recently (I know I know it’s not that easy to point to pure cause and effect relationships in markets because they are complex adaptive system with nonlinear relationship and so on but I digress) on concerns about the Federal Finance Housing Agency (FHFA) reviewing their rule that every Fannie Mae or Freddie Mac mortgage have a FICO score. 

The options the FHFA is proposing are:

• Option 1: maintaining the single score requirement for each borrower on every loan;
• Option 2: requiring multiple scores;
• Option 3: allowing lenders to deliver loans with any approved score; and
• Option 4; the “waterfall” approach allowing a primary and secondary score.

Option 1 is obviously the path of least resistance with the lowest cost and what is currently mandated. Option 2, requiring multiple scores, would likely not hurt FICO because their score would most likely be included. They would lose the monopoly tag which could impact the cadence of pricing, but this is not a terrible outcome for them. Option 3 would be the most impactful as it provides choice to the lenders and is currently what the market has priced in and would have management think twice about pricing because the market would have become a hell of a lot more competitive overnight. Option 4 is in the same bucket as Option 2.

The FHFA is not only evaluating VantageScore (FICO’s main competitor and the brain child of the CRAs), but also looking at multiple FICO scores including FICO 10 that claims the same ability to score more borrowers.

According to former FHFA Director Mel Watt, Option 3 would take several hundred million dollars and two years to implement. He said, “After all, we all believe that competition is good, don’t we? However, the more we looked into this issue, the more complicated it became and it is turning out to be among the most complicated decisions I have faced during my tenure at FHFA.” 

The primary added benefit of the VantageScore is that it scores more Americans, namely non-prime borrowers. In reality, this is a small portion of the mortgage market and only provides marginal benefit. Certainly not worth overhauling the entire system when the FICO 10, which uses the same data and scoring scale as the one already in place, is available. 

Management also recently disclosed that mortgage related scores are currently only 7% of total revenues, I think it further diminishes the FHFA risks around pricing. 

Turning away from the mortgage risk. Wall Street darlings will from time to time position and market themselves that they will replace the FICO score in some way. Upstart, Affirm, VantageScore, take your pick there’s plenty. However, there are numerous misconceptions around it. Let’s start with “What is the FICO Score?” The FICO Score is a very broad-based, low-cost effective way to evaluate credit if you had to rely on a single data set. Companies like Upstart or the bank themselves will add information on top of that. There’s nothing wrong with that and having more data at your fingertips is always nice to double and triple check. For a few pennies, what’s the harm?

 The key is, they are still using FICO scores as a starting point and as a way to communicate with other organizations as opposed to using their own, unheard of proprietary score that carries no weight for the other party. 

Fair Isaac’s CEO Will Lansing had a lot to say on this point at the recent Barclays TMT Conference at the end of last year. “The other misconception is that these Fintechs are competing with the FICO score, they’re really not. These Fintechs, including the ones that say they are going to replace, use millions and millions of FICO Scores every month. Fintechs use FICO scores in their own algorithms, in their own secret sauce, FICO Scores can be used as a filter for the lender as to which of the leads are candidates for loans.”

Brian Cassin, CEO of Experian also recently touched on this point in the Experian H1 Call

 

FICO’s 90%+ market share in the non-mortgage categories provides some read across that it can continue to successfully compete in the mortgage market with or without a mandate. The mortgage market is the only market that requires FICO scores and the abundance of other markets that FICO scores are used in suggests that the score is preferred in unregulated markets. 

I think the more useful debate for investors is comparisons between the FICO origination software and Upstart’s origination software, not to the FICO Score. They serve different purposes. In the case of the score, using one does not preclude the use of the other. 

Last note on general disruption is to several points made in a recent WSJ article about Citizens and Synchrony shifting away from FICO Scores. Here are the two blurbs I found most interesting “

FICO is becoming a smaller factor in underwriting decisions at Citizens Financial Group Inc. When shoppers apply for its buy-now-pay-later loans, the bank considers factors including the products they are buying, according to a person familiar with the matter. Fitness equipment, for example, is viewed as a sign of creditworthiness because it suggests a positive change in the applicant’s behavior. By next year, the bank plans to use cellphone, cable and utility-payment information to help vet loan applicants, the person said.”

“Synchrony Financial earlier this year said it had switched to VantageScore—a rival offering from the major credit-reporting firms. It produces scores for more U.S. consumers, the bank said, expanding its pool of potential borrowers.”

As far as the Citizens comment goes, FICO XD uses cellphone, cable and utility payment history to score previously invisible consumers. If I were a bank and my competitors were underwriting a $500,000 mortgage based on someone’s Peloton usage, I’d feel a little likeBuffett walking into Las Vegas and seeing people at the slots machine. I’d think that maybe the people I’m competing against aren’t as brilliant as I thought

The Synchrony Financial case is more interesting. It took five years for them to move away from FICO scores. The primary benefit of VantageScore is that it claims it can score 50-60 million more individuals, however with the recent rollout of FICO 10 and FICO XD this advantage seems to have fallen to the wayside. 

Bear argument #2 – Economic Slowdown

In some capacity, FICO is exposed to the broader economy. If big ticket consumer spending slows down, say for example the housing or auto market slow this would hurt their business. Less individuals looking for credit means less scores are used. 

On the Q2 ’22 earnings call Lansing stated that, “Consistent with what we’ve shared in the past, the three segments of the originations [mortgage, auto, personal and credit card], part of our B2B business, are roughly the same size. That varies from quarter to quarter, one grows or shrinks, but in approximate terms, that’s the way to think about and we don’t get into specifics beyond that.” 

Some back of the envelope math gets us to Mortgage revenues at about 14% of Scores revenues and therefore 7% of total company revenues. These slowdowns are generally temporary in nature and will be made up on the back end once the economy rejuvenates.

Bear Argument #3 – Tough Comps and Durability of Special Pricing

Bears have also cited that FICO is lapping spectacular results in the platform business and special pricing on the scores side which makes year over year comps difficult and could lead to disappointing growth rates.

It is abundantly clear through listening to calls and various conferences that FICO believes they still have a large gap to take in pricing and over the last few quarters Will Lansing has beat to the same drum: 

• “The short answer is there is a lot of runway left [in terms of the price vs value gap for scores]”-Will Lansing, Barclays TMT Conference December 2021.
• “For many many years to come I imagine we will touch the portfolio in terms of pricing”–September 2021
• “We don’t think about ceiling [in terms of price] we think about not shocking the system”-April  2022
• “It is really difficult to find businesses that are as attractive as our own”- April 2022

These are not comments you hear from a CEO who sees comps being an issue. He’s also talked about 50% growth rates for the platform over the next two years. Lapping over 40%+ growth rates and DBNRR of 140%+ is not a small task but noting how many non-platform customers still have to convert and how converted customers have stepped up spending on the new platform can help illustrate why Lansing is so confident. 

 It’s concerning to me, however, how openly FICO talks about their monopoly and margins. Generally, strong businesses/true monopolies try to hide their strength. Look at Visa and Mastercard or Google, these companies are stuffing every last expense they can find into their income statement to keep margins down and avoid regulatory scrutiny. I wonder if the regulatory threats are because FICO hasn’t done a good job hiding it. 

Competitive Positioning Against the CRAs

Lastly, I wanted to riff on FICO’s competitive position for a minute. FICO has a strong moat locking in banks and consumers which I cover in the next section, however, their position against the credit agencies is more precarious.  First of all, the agencies make up 30% of revenues and distribute nearly all of the B2B scores. The agencies also provide the data that FICO uses in its algorithm to generate scores. It speaks to how strong the moat is that FICO hasn’t been ripped out and replaced by the agencies. It seems like such a logical step for the agencies to take to cut out FICO or stop providing them with data. They certainly tried to do that in 2006 with the founding of VantageScore, but it’s taken a decade and a half for it to pick up any steam and at it’s less than 10% of the market. 

Moat 

FICO’s Scores business exhibits extremely high switching costs as well as being the low cost producer of scores with high accuracy and brand loyalty. By way of example, it took Synchrony Financial over three years to move away from using FICO. Not many other FIs are going to be enthusiastic about undergoing such a change unless the benefits are enormous. 

Investors, banks, credit cards companies and mortgage lenders have all had FICO ingrained into their systems and vernacular which makes switching incredibly expensive and time consuming. Eagle Point Capital said it well in his write up, “Switching would be like getting Americans to start using the metric system or quoting temperature in Celsius.”

These high switching costs play a vital role in the upcoming FHFA decision. The FHFA is considering changing their rule in which Fannie and Freddie require a FICO score for a mortgage.  Lansing described the moat and its role in shaping how the FHA could decide, saying, “On the FHFA decision a system in which you mandate multiple scores is more costly, a decision where you change from FICO is more costly because of switching costs for little if any benefit.”

It is hard to critique the score for what it is. It is a cheap, efficient way to get a grasp on the propensity for an individual to repay their debt. It is the best building block. Start with a FICO score and then go and build on top of it, go get your extra data. If it is available it doesn’t make sense to not get it.

Fico has also become the standard for selling loans between banks. Banks will develop their own in-house methods for underwriting and communicating about a loan’s credit quality, problems, arise however when they go to talk to other banks or insurers about it. A little like how the GPA scale varies at different colleges, credit scores do too. FICO’s role in the ecosystem is to provide that common language. This is evidenced by the fact that 98.8% of dollars securitized in the US solely cite Fico scores as a risk measurement.

FICO scores also make it easy to communicate with borrowers as well. Explaining to a borrower that their FICO score wasn’t up to the standards is a mch easier than having to explain an in-house or unheard of algorithm. 

Lastly, FICO scores aren’t expensive enough to encourage a lender to switch. No credit analyst is going to be able to look their boss in the eyes after losing thousands of dollars on a loan and say “Well,  no I didn’t think it was worth it to buy the FICO score. It’s 20% more expensive this year so I took the low bid and saved a few cents.” In a world of loss aversion, we are likely to pay up for certainty. 

The burgeoning software business also has susbanital switching cost and it becomes more and more embedded in customer workflows and employees become entrenched in the software. The 90%+ retention rate leads to highly visible recurring revenue. 

Management and Capital Allocation

I have mentioned him a few times so far but FICO is headed by Will Lansing who has been the CEO of Fair Isaac since 2012. Lansing has proved to be an incredibly focused and blunt leader with a straight-shooter attitude toward shareholders. He tells you when the stock is cheap and exactly how the company thinks about capital allocation. It helps, too, that he owns about 1.38% of the company. 

FICO has been eating up their share count, retiring 2.5% of shares out since 2012 from ~35 million down to just over 26mn as of Q2’22. 

Lansing has also stated that the company is not afraid to lever up for opportunistic acceleration of buybacks. He stated recently that, “We think of ourselves as kind of a “go private publicly” organization. We’re an LBO for the benefit of our public shareholders. So we’ve levered up somewhat. We usually run somewhere north of 2 and south of 3x EBITDA leverage. It could be higher than it is today, but we want that to be healthy.”

Below you can see a layout of management’s capital allocation over the last decade. Over 70% of capital has been returned to shareholders through buybacks since Will Lansing took over in 2012, 25% has gone to R&D to build out the software segment and turn out new scores models and the remaining 5% was allocated to cap-ex. 

Management has done a spectacular job taking advantage of the reinvestment opportunities in the business by growing the software side generating ROIIC in at 22% over the last 5 years. The durability with which the company has been able to generate ROIC above it’s cost of capital speaks to the strength of the moat. 

Valuation

“Anyone who analyzes our stock price today comes to the conclusion that scores accounts for 100% or 120% of our stock price meaning that the software business really is disrespected”- Will Lansing, CEO, FICO

I believe that there is a somewhat wide distribution of outcomes for FICO, however the expected is positively skewed. My base case assumptions are:

• Scores Revenue to grow at a 12% CAGR for the next 5 years 
•  Software grows at 6.8% CAGR 
• Professional Services drops 30% this and continues to decline at 10% for through 2027
• Operating margins are steady at 40% (a 200bps increase from 2021)as Professional services runs off and software operating leverage kicks in
• Net income margins remains flat through 2027 
• FCF Grows at 8.5% through 2027
• Shares out decreases by 3% per year 

This produces a ~16% IRR from today’s current price of $395 at a 4% FCF yield exit. With slightly more bullish assumptions (15% scores revenue growth and 10% software growth) IRR would be 20%. Below are my segment and overall models. 

DCF with a 9% discount rate and 3% terminal growth 

Bibliography

2020 10K 
2021 10K 
Q4 Financial Highlights 
March 2022 Investor’s Presentation 
Consumer Credit Reporting Industry Overview 
FHFA Seeking Alpha Article 
Eagle Point Capital on FICO 
YoungMoney Capital On FICO 
Barclays TMT 2021 Conference
Barclays Sept 2020   

Disclaimer: I am not a financial advisor. These articles are for educational purposes only. Investing of any kind involves risk. Your investments are solely your responsibility and we do not provide personalized investment advice. It is crucial that you conduct your own research. I am merely sharing my opinion with no guarantee of gains or losses on investments. Please consult your financial or tax professional prior to making an investment.