What this covers

This is a conversation in which Dev Kantesaria walks through Fair Isaac, starting with its history. Standardized credit scoring let lenders set cutoff scores, compare loans across locations, and price interest by risk, which in turn opened the subprime market. The 1981 PreScore was the first generic score, and in 1995 Fannie Mae and Freddie Mac required FICO scores for the mortgages they handled. Dev stresses that FICO builds mathematical models and holds no consumer data. It has three layers of customers: the credit bureaus, which pay about 75% of scores revenue, the lenders who make decisions, and the consumers who ultimately bear the cost.

The middle of the discussion tests that position. Dev argues the model itself is not the moat, since two MIT students could build one, and that the standard's reach is what counts. The conversation covers the special price increases begun in 2018 after roughly 25 years without raising scores prices, and why Synchrony's move to VantageScore saved little while costing securitization pricing. It also takes up the FHFA's cut in required bureau reports from three to two, the roughly 50 basis point penalty for non-standard providers in securitization, and the cyclical dip in scores revenue during 2007 to 2009.

The talk then turns to the rest of the company: the Falcon fraud product and its data network, the capital-light software segment, and buybacks that retired a fifth of the shares in five years. Two points are contested. One is that score predictiveness has plateaued for about 15 years. The other is whether AI could improve on it without introducing discrimination.

Sharpest takeaway

FICO is a near-ideal business because its scores product is a de facto industry standard whose network effects, switching costs, and tiny cost relative to loan decisions give it durable pricing power and very high margins, while its software segment is a younger, competitive business that is improving.

  • Scores are used in over 90% of US lending decisions and cost less than $1 against loans of hundreds of thousands of dollars, so price increases carry little perceived cost to the end customer.
  • Switching away from FICO (for example Synchrony to VantageScore) saves little money but raises securitization costs, so almost no lender has followed.
  • Price increases since 2018 added about $60 million of high-margin revenue, growing scores revenue despite flat or declining volumes.

The argument · threads2 threads · 58 claims
0.71

FICO's position rests on data-driven scoring and a three-layer customer base that pays for it.

2 pointscentrality 4/5
  • Scoring models improved mainly because combining statistical methods with large datasets of consumer characteristics and default history allowed the models to be refined, not because the underlying math was novel.

    “What they really needed to refine their models were large datasets. So combining the mathematic models with large datasets, consumer characteristics, default history”

  • FICO has three layers of customers: the credit bureaus that pay FICO directly for scores (about 75% of scores revenue), the banks and lenders that make credit decisions, and the consumers who ultimately bear the cost.

    “FICO scores actually have three layers of customers. ... the direct revenue for scores comes from the credit bureaus, which represent about 75% of scores' revenues, but the end customer is, in fact, the banks and financial institutions, which are making credit decisions.”

0.57

FICO's software leadership and a recognized business shift made it a justified purchase.

2 pointscentrality 3/5
  • Fraud detection and customer management together make up almost 50% of software revenue, and FICO is the clear market leader in both areas.

    “together, these two sets of solutions represent almost 50% of total software revenue.”

  • Dev's firm studies companies like historians and monitors them for years, and it bought FICO in 2018 when it recognized a change in the business that others had not yet noticed, even though it had worried about organic growth earlier.

    “We like to be great historians of companies. ... And then we realized that there was a change happening that many others in the marketplace may not have realized.”