FICO and the Consumer Credit Bureaus: Part 1

This circuitous 2-parter is going to open with the consumer credit bureaus (Equifax, Experian, and TransUnion – EET), switch to FICO, and loop back to the bureaus…apologies in advance for the disorienting structure.

Let’s start with Experian, which like many companies formed through motley acquisitions, wasn’t so much created as it was allowed to emerge.

During the ‘60s, the mail order and home furnishing divisions of UK retail conglomerate Great Universal Stores (GUS) developed a computerized system for keeping track of the millions of consumers to whom it extended credit. The in-house database, which absorbed county court judgments and other public data over the years, was commercialized as Commercial Credit Nottingham (CCN) in 1980 and merged with Manchester Guardian Society (a credit association formed in 1820s[1]) 4 years later to become the largest bureau in the country[2].

In the US, a defense focused

electronics company founded in the 1950s called TRW applied its engineering

talent to automating credit reporting, acquiring Credit Data in 1968 to become

one of the 5 largest agencies in an industry whose ranks had swelled from 50 in

1900 to 1,000 in 1920, to 2,000 by 1960 as more automobiles, appliances, houses

and other comforts of the American Dream were financed with credit, but were

now dwindling as a handful of tech forward players began rolling up smaller,

sleepier bureaus. Eight years following

the somewhat botched 1988 merger between TRW Information Systems &

Services (the credit reporting division of

TRW) and Chilton Corp. (one of the other 5 national agencies, founded in 1897),

TRW sold IS&S to Bain Capital and Thomas H. Lee in 1996, who rebranded the

company Experian and less than 2 months later, merged Experian into GUS for

$1.7bn. 

The combined company’s footprint in the US and UK was augmented by acquisitions in less developed markets. Most notably, in 2006 Experian took a 65% interest in Serasa, the leading credit agency in Brazil[3], spending $2.8bn in total (~8.5x EBITDA) to claw its way to full ownership by 2012 in its largest “acquisition” to date. Experian complemented Serasa with bureau acquisitions in Colombia[4], Peru, and Venezuela, and then launched a credit agency JV in Australia with the country’s leading banks. 

Experian expanded not just across

geographies but verticals. In the early

2000s, with the advent of the consumer internet, the agencies started offering

a direct-to-consumer product, where consumers could pay a subscription fee to monitor

their credit. But from fy04 to fy06 (fy

ending March) Experian stretched its interest in the consumer market to the max,

blowing ~$2bn on misguided

diversions into online comparison shopping search engines and lead gen businesses

(think shopping.com and LendingTree)[5] that were moved into “strategic review”

during the downturn and later divested[6]. 

Transunion has a similarly rich

history with nonobvious origins. The

company was founded in 1968 as a subsidiary of railway equipment leasing

company Union Tank Car. It got into

credit reporting services by rolling up or partnering regional credit reporting

services, many with origins dating back to 1800s[7], until it achieved

nationwide coverage in 1988. In the 90s,

TRU acquired its way into analytics and marketing services, and began its foray

into emerging markets, acquiring the largest credit database in South Africa

(an acquisition that would be followed by scattered minority/majority ownership

stakes of leading credit agencies in Brazil, Mexico, Chile, Colombia, and India

over the next 25 years). 

Transunion has traded hands multiple times: it was sold to Marmon Holdings in 1981, spun-off to the Pritzker family in 2005, who then sold 51% of the company to Madison Dearborn partners in 2010. In 2012, the company was sold for $3bn to Advent International and Goldman Sachs, who took the company public in 2015.

And then finally, there’s Equifax, founded in 1899 to evaluate applications for insurers before then acquiring a bunch of regional credit agencies over decades until by 1978, the company had become so dominant that it was ordered by the FTC to divest several bureaus. Like TransUnion and Experian, Equifax started expanding into analytics and overseas markets[8], planting flags in Latin America through a series of acquisitions in Brazil and Argentina in 1998 and more recently establishing a presence Australia/APAC through its 2016 acquisition of Veda (at $1.9bn, the company’s largest acquisition to date). Equifax also forayed into credit card processing and healthcare claim reviews, businesses that it divested in the late 90s/early 2000s[9]. Other domains like loan origination software, talent management, mortgage settlement services, and direct marketing were sold between 2010 and 2015[10].

From the late 90s through the early 2000s, organic product development took a back seat to Equifax’s strategy of rolling up US credit bureau affiliates (affiliates = franchisees who used Equifax’s credit databases and sold Equifax products to banks in their territories). Then in 2005, Rick Smith took over as CEO and did a fine job invigorating the company’s product development cadence until one day, whoopsie, data breach. 

Here are charts to give you a rough sense of how the 3 largest credit bureaus stack up against one another:

[Experian: LTM through September 2019]

The largest source of revenue for all

3 bureaus comes from North American businesses, who purchase data and software to

help them make underwriting decisions.

All three companies are the product of a tech-forward commercial entity stitching regional agencies together into a nationwide database. Both elements – the databases and the technology that makes them accessible and predictive – are essential, but they play somewhat different roles. Technology was what allowed these bureaus to effectively scale the disparate collection of regional datasets during the ‘60s through the ‘80s and today, it is what underlies the new product development that fuels 40%-50% of organic growth (more on this in Part 2). 

If technology drives growth, data

underpins durability. The CRAs’ entrenched

positions stem from historical dependencies and reflexive scale advantages that

have accrued over many decades of owning the most comprehensive datasets. What I mean by the latter is that the CRAs

effectively function managers of data coops. 

Just as 100 friends, each with a separate piece to the same puzzle, will

have a more comprehensive understanding of the puzzle by joining their pieces

together, lenders mutually benefit from combining the payment histories of

their customers. Each “puzzle piece” is

separately of little worth but collectively valuable. Lender B in Cleveland wants to know if

Customer 1 is has already taken out a big loan from Lender A in Portland. And just as you would rather see a puzzle

that is 3/4 of the way complete than one that has just started, a lender will

find it more valuable to access and contribute to a national bureau database

that captures payment data from every lender in the country than one that only

gathers repayment data from, say, 10 banks in Chicago. Centralizing data also acts as a stick, since

if a consumer knows that his payment behavior is going to be captured no matter

what – that a bank in New Jersey will be privy to his delinquent loan in

Montana – he will be more likely to stay current on his obligations. 

And so you eventually get to a place where data aggregators with the most mass reflexively attract even more data. In my Verisk post, I wrote:

The agreements through which a customer licenses VRSK’s data also allows the company to make use of that customer’s data….so essentially the customer pays Verisk for a solution that costs almost nothing for the company to deliver and Verisk gets to use that customer’s data to enhance its own solutions, which improved solutions reduce churn and attract even more customers (and their data) in a subsidized feedback loop. 

But not only is data collection

self-reinforcing, but different kinds of data can be monetized across different

kinds of customers. For instance, in the

mid-90s, Equifax built out a closed exchange where telcos and other utility

providers contributed their “negative data” (that is, data related to fraud and

delinquency) and use the pooled data to garner better insight into fraudulent

behavior. In 2011, Equifax launched a

separate exchange where telco/utility participants would contribute “positive

data” (on-time payments), with the aim of identifying the significant minority

of consumers who regularly pay their utility bills on time but who were not

captured in traditional credit files. This

data was then used by financial institutions to underwrite borrowers as well. Likewise, the income and employment data

acquired in Equifax’s 2007 acquisition of TALX has been carried beyond the

original use case of verifying new employees, and is now used to underwrite

mortgages, auto loans, and credit cards. 

Finally, not only can the same dataset be leveraged across different

industries, but different datasets can be combined to make smarter decisions. Data on someone’s payment history is far more

informative when augmented with income and wealth data. 

Given the mutual benefit of pooling, most of the data the bureaus collect are contributed by lenders at no cost. This data can be combined in various ways to make smarter decisions and scales across many customers, giving rise to healthy 35%-40%+ EBITDA margins in the mature US market (capex usually runs ~6%-7% of revenue, most of that going towards new product development). 

The industry trends toward natural

oligopoly since only the most widely referenced and complete databases are fed

data from lenders, which further reinforces their dominance and compels the

incremental lender to contribute as well. 

Hence, as I wrote in my Equifax post from 2017:

Today, you can’t take out a mortgage,

credit card, auto loan, or in many cases rent an apartment without first having

a credit report furnished from one or more of the three main consumer credit

reporting agencies – Equifax, Experian, and Transunion – who warehouse years of

credit activity for the 80% of American adults who carry debt.

….From a system-level point of view,

the role that these agencies play in the credit ecosystem is vital. Without them, credit decisions would take an

inordinate amount of time and borrowing costs would be higher in aggregate to

account for information asymmetries and the redundant administrative burden

borne by each lender, who would have to separately research credit risk. 

The data and software sold by the credit bureaus are deeply intertwined in the core processing infrastructure and workflow of financial institutions. Moody’s and S&P are similarly entrenched in the debt issuing ecosystem. Their ratings’ essential role in resolving information asymmetries between debt buyers and issuers and their ubiquitous risk management function basically immunized the rating agencies from serious repercussions post 2008/2009, when it became evident that they were rubber-stamping AAA ratings on precariously leveraged asset backed securities. You may recall that several smaller ratings agencies tried capitalizing on the scandal to steal share to no avail. Moody’s and S&P retained their oligopoly and continue to rate 85% of the world’s outstanding credit. 

As testament to how entrenched the

CRAs are consider that in September 2017, Equifax had its own “Moody’s” moment

when it failed to prevent a data breach that exposed the personal data of 147mn

consumers. In my 2017 post on the matter,

I speculated that the damage to Equifax’s credibility could be so severe that it

risked arresting the feedback loop between the data contribution and

consumption as financial institutions – for security and PR reasons – balked at

entrusting the company with their data.

But my concerns proved totally unwarranted. 

Equifax purged its senior leadership; settled class action lawsuits with consumers and shareholders; resolved civil enforcement actions with state AGs; and modernized its creaky IT and data security infrastructure. Equifax’s customers seem cool with this. The crisis precipitated no major share shifts. Equifax spent a few quarters in the penalty box before the projects that were put on hold immediately following the breach continued as plan. Revenue in its US Information Solutions (USIS) segment – which sells consumer data and applications to US financial institutions and was at the center of the data breach – declined by just 1% (1%!) in 2018[11] before growing 2%+ in 2019 (including 8% growth in 4q19). Equifax’s International segment grew 7% in 2018 (ex. fx), followed by 3% growth in 2019. Excluding breach-related one-timers USIS margins declined by 5 points from the 2016 peak as the company invested in technology and data security…but who’s complaining about 45% margins? International margins, meanwhile, barely budged. 

I drew the parallel to Moody’s earlier only to point out that if your negligence gives rise to the largest data breach in American history and your revenue declines by just 1% the following year, this might be a sign that you have a good business. But the way in which Moody’s and the consumer credit bureaus draw their advantages differ. The credit bureaus draw power from their datasets and technology while Moody’s Investor Service relies on its ratings. The latter’s moat is not so much about ratings accuracy – I bet that many others could design a relative ratings system that was just as good at predicting probability and severity of default – as it is the standardization function that ratings play. As I wrote in my Moody’s post:

Over the last 100+ years since its

founding, Moody’s ratings – derived from a consistent framework applied across

11k and 6k corporate and public finance issuers, respectively, in addition to

64k structured finance obligations – have become the veritable benchmark by

which market participants, from investors to regulators, peg the credit

worthiness of one debt security against another. NRSRO ratings underpin the risk weightings that

banks attach to assets to determine capital requirements, dictate which

securities a money market fund can own, and, in ostensibly surfacing the credit

risk embedded in fixed income securities, make it easier for two parties to

confidently price and trade, enhancing market liquidity.

Because of such industry-wide adoption, a debt issuer has little choice but to pay Moody’s for a rating if it hopes to get a fair deal in the market: an issuer of $500mn in 10-year bonds might pay the company 60bps ($3mn) upfront, but will save 30bps in interest expense every year ($15mn over the life of the bond)….and each incremental issuer who pays the toll only further reinforces the Moody’s ratings as the standard upon which to coalesce, fostering still further participation. This feedback loop naturally evolves into a deeply entrenched oligopoly. In terms of total ratings issued, S&P and Moody’s are at the top of the heap. There are actually 10 NRSROs, but unless you work in credit, you’ve probably never heard of most of them (Egan Jones anyone?)

In that sense, FICO is perhaps a closer

analog to Moody’s. 

FICO was founded in 1956 by engineer

Bill Fair and mathematician Earl Isaac, who while employed at the

Stanford Research Institute collaborated on how to solve problems by applying

math to datasets. The company they

created – Fair, Isaac, and Company (FICO) – performed custom analytics work for

lenders until the late ‘70s, when it standardized its IP into software. In 1989, the company launched the FICO Score,

which has become the de facto metric by which US lenders assess the

creditworthiness of borrowers. 

FICO owns the algorithm that computes the FICO Score, not the data that feeds the algorithm. The latter belongs to the big 3 national credit bureaus, who license FICO’s algorithm, run the algo against their consumer data files (because that data varies somewhat across bureaus, so too will the FICO Scores), and sell their data plus the resulting FICO Scores as a bundle to lenders, who use that information to underwrite risk. 

Like a Moody’s rating, the FICO credit

score’s ubiquity reinforces adoption: it wouldn’t make sense for a bank to

suddenly adopt an alternative, less widely used credit score (even if that

score were equivalently predictive) given how hard it would be to reconcile the

new rubric with the FICO Score that is already woven into its risk management

process and being used to tag, price, and otherwise evaluate millions of

borrowers already in its system (according to management, at least one bank has

FICO Scores embedded in 67 mission critical applications). The FICO Score is the industry standard for

seeing how one borrower’s creditworthiness has trended over time and for

comparing one borrower’s creditworthiness to another’s. 

To a large degree, FICO’s dominance owes a great deal to the three credit bureaus, who control consumer data distribution to lenders and account for 80% of FICO’s Scores segment revenue. Bureau data adoption was fueled by the FICO scoring standard and the FICO scoring standard feeds on bureau data. The symbiosis extends to product development. Over the last 2 years, FICO launched XD and UltraFICO Scores in partnership with Equifax and Experian, respectively, to incorporate the alternative data like income and payments to landlords and utility providers. 

But FICO and the bureaus also compete in numerous domains. The bureaus distribute FICO Scores even as they market their own proprietary scores and even teamed up in 2006 to launch VantageScore, an alternative to FICO that, despite the bureaus’ stranglehold on distribution, has failed to dent FICO’s 90% share of credit scoring across credit cards, auto loans, and mortgages. 

After decades of stable pricing, FICO pushed through significant hikes for scores used in mortgage and auto originations, which fueled y/y revenue growth of 29% in 2018 and 25% in 2019, testament to FICO’s market power but also to the trivial cost of pulling a FICO Score relative to the value that score provides. FICO was charging 6 cents every time a FICO Score was referenced for a mortgage origination, so maybe it’s more like the 9 cents today after the price hikes. Of course, the credit bureaus apply some markup to the score and pass it on to the lender, but even at 3x-4x the wholesale price (just guessing here), the dimes a lender pays for a FICO barely register as a rounding error against the ~$8,000 that it costs to take the average mortgage from application to close. I estimate that the FICO Scores revenue distributed through the 3 credit bureaus account for less than 5% of the latter’s collective operating costs. So while FICO Scores are a mature product with volumes that track GDP growth, they command enormous pricing power that has long been untapped until recently. 

Scores is a royalty business. For each of the 10bn+ times a FICO Score is pulled every year, FICO realizes ~3c-4c of revenue, almost all of which drops down to profits:

The other part of FICO’s business is Software, which consists of: 1/ analytics used to find patterns in data and make automated decisions based on those patterns (“Decision Management”) and 2/ the proprietary applications that draw on that IP (“Applications”). Management disaggregates Software into these 2 distinct segments because in theory a lender could choose to consume FICO’s analytics separately from its apps. But I think it best to just combine them into a single unit because: 1/ Decision management, the “platform” layer, is small anyways, accounting for just 12% of FICO’s revenue and less than 0% of its profits, and 2/ because many of the investments, particularly in cloud infrastructure, relegated to Decision Management also benefit the Applications segment, looking at the latter’s financials in a vacuum gives a distorted view on how things are going.

For instance, here’s Applications…

…and here’s Software (Applications + Decision Management):

The software business is a collection

of applications sold directly to banks and other lenders who use it to automate

loan originations, monitor the credit performance of existing accounts,

identify fraudulent behavior, market products, and facilitate collections on

delinquent loans. FICO has leading

market share in a few of these areas. Its

TRIAD application – which banks use to change limits or rates based on whether your

credit performance is improving or deteriorating – monitors 2/3 of all credit

card accounts in the US. Its fraud

detection franchise, Falcon, used by card issuers and merchant acquirers (First

Data, Fiserv, TSYS) to detect whether a point-of-sale swipe is legitimate, has

been the industry standard since the 1990s and commands 90% share.

In 2011 management referred to Scores as a mature business whose GDP growth profile would be offset by the high-single digit growth in Software. Well, they were half right, with Software growing by ~6%/year (nearly all organic) as predicted, but Scores revenue compounding at 13%, thanks in part to the price-fueled revenue growth over the last 2 years. With virtually no investment required to sustain its growth, Scores EBIT has grown at 16%/year. Software profits, meanwhile, have been flat over the last 8.5 years, thanks to significant investment in cloud infrastructure over the last 2 years[12]. But even if I “back out” those investment and consider the Applications segment on its own, EBIT has grown by less than 5%/year. 

And so I see FICO as the combination of an astoundingly profitable royalty business that enjoys no competition and a so-so software business that overlaps with solutions offered by all three credit bureaus. At 60x trailing/70x fy19 earnings (after significant stock comp expense), the stock trades like the excellent business that it is. I don’t get how you make money at the current multiple. The core US market is tapped out and without a symbiotic relationship with credit bureaus, there is no reason to expect FICO’s algorithm to become a standard in international markets. Judging by the litigious relationship between FICO and EET in the mid-2000s (EET launched VantageScore and FICO sued), I reckon that if the bureaus could go back in time, they’d try harder to claim FICO’s Scores royalty for themselves. It is telling that the US bureaus have not exported the FICO Score to the international markets that they’ve come to dominate over the last 20 years.

As I mentioned earlier, FICO in theory has lots of room to lift pricing, but as a monopoly they must be careful about the degree to which they exercise this power, lest they invite anti-trust action. It is probably no coincidence that the US Department of Justice’s recent anti-trust investigation into FICO follows closely on heels of the company pushing 25%+ price hikes in each of the last 2 consecutive fiscal years.

I expect Scores volumes, which are

tied to loan originations, to experience significant COVID-induced declines

(15%-20%?) over the next year or so, perhaps worse than the 19%-20% contraction

experienced from fy08 to fy10 (ending September), though I suppose some of the

volume declines this time around may be offset by pricing actions. Software revenues declined by ~19%-20% during

the last recession and would expect a similar outcome this time around. 

But let’s be generous and assume a cumulative 10% revenue decline for both segments over the next 2 years, followed by 13% revenue growth at Scores and 9% growth in Software. Let’s say Software revenues rebound at 50% contribution profits, driving margins from 15% today to 25% in year 5, around where they were 2015, before the company began aggressively upgrading its tech. This translates to around $485mn in free cash flow in year 5 (vs. $260mn LTM). Now, unlike the acquisitive credit bureaus, FICO has plowed all its free cash flow into share buybacks, without which I estimate FICO’s share count would have increased by close to 7%/year over the last 3 years instead of shrinking by 2%. 7% annual dilution is of course a significant and insidious penalty to shareholder returns that needs to be explicitly account for. At 35x year 5 free cash flow/share + accumulated cash, I impute a 4% return from today’s price ($370).

………………………………………………….

The bulk of the credit bureaus’ and

FICO’s revenue is b2b, derived from banks, lenders, and other enterprises who

use the data, analytics, and scores to evaluate credit risk. A much smaller portion comes from consumers,

who pay a monthly subscription fee to access their payment histories, credit scores,

pointers on improving their scores, or identity theft protection. Alternatively, under affinity arrangements,

this information is white-labeled by banks and fintechs, who use it as a lead

gen tool for credit offerings (“we’ll give you a free credit report from

TransUnion and in return you let us market credit card offers to you”).

From 2002 to 2010, the credit bureaus launched a bunch of DTC membership sites (including freecreditreport.com, checkmyscore.com, and creditreport.com), where a vigilant consumer might pay $10-$20/month to obtain her credit score and payment history before applying for a loan. What she may not have known without reading the fine print was that the credit score furnished to her was not the official FICO Score that lenders would use to evaluate creditworthiness, but rather an “educational” Experian Credit Score or a TransUnion Credit Score. Moreover, these sites often employed dubious and aggressive marketing tactics, representing themselves as “free” to stealthily lasso consumers into subscriptions.

As these sites came under heavy regulatory fire after the financial crisis, Experian put its “free” brands into run-off and invested more heavily behind its paid site, Experian.com[13]. But these sites are still around, promoted under pretext that they lend transparency and consumer allow to take control of their credit destiny…and yet:

Source: American Banker

Considering how much more directly entwined

large financial institutions like JP Morgan and Bank of America are with the

financial lives of consumers, it’s kind of nuts that these 3 credit bureaus

accounted for 1/3 of the complaints received by the CFPB in 2018.

Anyways, in 2013, with goal of

cleaning up a marketplace rife with confusion and sketchiness, FICO launched

the Open Access program, where all the major banks could disclose FICO Scores

to their customers for free. Rather than

pay these “free” sites $15/month for an educational credit score that may not

bear resemblance to the score a lender would use to inform its underwriting,

bank customers could now see their real FICO Score displayed on their monthly

bank statements. 

Starting around 2015, FICO monetized direct consumer access to FICO Scores by: 1/ offering the FICO score and the credit reports from EET on its own website, myFICO.com[14]; and 2/ striking resale agreements with Experian and Discover. By being the only bureau with legit FICO Scores, Experian could differentiate its DTC premium subscription service from those of Equifax and TransUnion (Experian’s proprietary credit score is the “FICO Score” of the UK market). Discover, meanwhile, could offer free FICO Scores to attract new customers. 

With their paid DTC business under

pressure, TransUnion entered into white label affinity agreements with

fast-growing fintechs like Credit Karma, who offered free credit scores to

consumers in return for being able to market financial products to them. Equifax followed shortly after. Experian held out and this has cost them

dearly. TransUnion has more than doubled

its Consumer revenue over the last 5 years (with 70% of revenue coming from

such affinity deals, TransUnion’s “Consumer” business is more accurately

characterized as b2b). Equifax Consumer

enjoyed low/mid-teens growth in the years leading up the breach debacle. Experian’s Consumer revenue, meanwhile, has gone

nowhere. The FICO agreement only helped

offset double digit declines at its “free sites” (which had come under

regulatory scrutiny for their misleading ads) and weakness in its affinity

channel (banks were pulling back on marketing). 

It has not been enough to dramatically lift Experian’s DTC fortunes.  

And frankly, DTC hasn’t been great for FICO either. In 2015/2016, management pitched B2C as a promising growth vector for Scores, but it hasn’t really amounted to much. DTC was described as a “very small” part of the company in 2017 and in the years since, myFICO has only growth by ~mid-single digits/year. The b2c business in aggregate grew by just 3% in 2017, implying that the resale (“affinity”) portion of B2C grew even slower than myFICO.

I fear that the “FICO Score”, in its ubiquity, has disassociated from FICO the company and become a generic name for any 3 digit number that proxies credit-worthiness. While banks are keenly aware of the difference between “educational” credit scores from Equifax and TransUnion and a real FICO Score from FICO, consumers often are not, which makes credit scoring more competitive in the consumer domain than it ought to be. Experian and FICO have been denigrating these educational scores as phony BS replicas for some time, but whatever. TransUnion’s Consumer segment has enjoyed massive growth over the last 6 years – with revenue growing from $200mn to $500mn – by licensing out its educational score to fintechs and banks. 

Footnotes

[1] In the UK during the early 1800s,

merchant members of what would later be known as the Manchester Guardian

Society pooled together the names of people who failed to pay their debts

[2] For a more thorough look at the

origins of Experian, I recommend reading this abbreviated history.

[3] Serasa was formed in 1968 by an

association of Brazilian banks. Its market share in Brazil is ~60% vs. #2

Equifax at 17%

[4] DataCredito, 60% market share

[5] including

LowerMyBills.com[a], PriceGrabber.com, ClarityBlue, Baker Hill, QAS, FootFall

and CheetahMail

[a]

LowerMyBills.com, which sourced ¾ of its business from subprime lenders, was

hit particularly hard during the last recession.

[6] Experian

sold its comparison shopping and online lead gen businesses to Ybrant Digital

in 2011 for $175mn; jettisoned Baker Hill (community bank software), FootFall

(retail data), Hitwise (online marketing intelligence), and various other

smaller businesses in Estonia, Morocco, and The Netherlands in 2015; and its

sold email and cross-marketing businesses to Vector Capital in 2017 for

$400mn. 

[7] mostly notably the acquisition of

the Credit Bureau of Cook County in 1969    

[8] Equifax first got into Brazil in

1998 by acquiring a large commercial data provider and in the ensuing years

tried but failed to build its own consumer database. After Experian acquired Serasa, Equifax

established a marketing relationship with ACSP, which was a non-profit run by

Brazil’s Chamber of Commerce and ran the second largest consumer credit agency

in the country. In November 2010, ACSP

spun out its consumer data division as a company called Boa Vista (BVs) and in

2011, Equifax merged its Brazilian business (which consisted of commercial data

and a little consumer data) into BVS, taking a 15% equity stake in the

latter. 

[9] In 1996, Equifax divested its

healthcare business; In mid-1997, Equifax spun-off its insurance services

operations; in 2000, it sold risk management collections; and in 2001, it

spun-off Payment Services as a company called Certegy, which Fidelity National

Information Services would acquire in 2005 and then sell to Varian Equity

Advisors in September 2018.

[10] Experian also divested large

portions of its Marketing Services business in fy16 and fy17 (ending March).

[11] After backing out some minor

revenue charges related to one-time settlements with commercial customers

[12] FICO has been refactoring its

apps for the cloud since 2012. SaaS

comprises 40% of software revenue and is growing at ~15%/year. 

[13] In 2014-2015, the company

migrated subscribers from the free brands to Experian.com

[14] approximately 65%-70% of myFICO revenue would then be remitted to the credit bureaus who own the IP for the credit reports.

Disclosure: At the time this report was posted, Forage Capital did not own shares of FICO, TRU, EFX, or Experian. This may have changed at any time since.