Editor’s Note — This article is sponsored by SOLO (and just so you know, I also do some advising work for them, which is totally unrelated to this piece, but just FYI.) As with all sponsored content in Fintech Takes, this article was written, edited, and published by me, Alex Johnson. I hope you enjoy it!
Here’s a question: If you were to sit down and design a credit bureau from scratch, in 2025, what would it look like?
The reason this is an interesting question is that the answer is likely very different from what the credit bureaus that we have today — Equifax, Experian, and TransUnion, being the big three — look like.
So, in this article, I’d like to explore this question in more detail, using three more detailed questions as our guide:
- Why do the credit bureaus work the way they do?
- What are the challenges inherent to the way that credit bureaus work today?
- What characteristics should a modern credit bureau have?
Why do the credit bureaus work the way they do?
In his recent market research report — How Cash Flow Data Can Defuse the Credit Score Time Bomb — Martin Kleinbard summarized the answer to this question very well, writing, “The longstanding equilibrium was more a result of technological happenstance than analytical master planning.”
That’s exactly right.
When you ask questions like, “Why don’t the credit bureaus include income or bank transaction data?” Or, “Why don’t the credit bureaus provide consumers with the same level of granular control over sharing their data that the open banking data aggregators do?” Or, “Why don’t the credit bureaus pay furnishers for the data they provide?” The answers almost always trace back to the same root problem — the credit bureaus, as we know them today, were designed in the 1980s.
If you wanted to create a standardized and highly reliable dataset for lenders to use in their risk evaluations back then, you would have picked credit repayment data. It was easily output by the electronic loan management systems of the day, and it was narrow enough in scope that it could be parsed by lenders relatively efficiently.
If you had proposed back then that consumers maintain granular, real-time control over their data and on-demand visibility into how it’s being used, everyone would have looked at you like you had lost your mind. We simply did not have the technology in place in the 1980s to empower consumers to be active and informed participants in the emerging financial data economy.
And if a lender had the foresight to ask the credit bureaus to be compensated for the data they were furnishing back then, the credit bureaus would have found the request both bizarre and slightly unseemly.
Put simply, the way that the credit bureaus work is a historical accident. It’s not the result of careful, forward-looking planning and design.
What are the challenges inherent to the way that credit bureaus work today?
There are a few!
And they have become much more obvious over the last 10 years or so.
First, the scope of the data that the credit bureaus have is too limited. We don’t have the same technical restrictions on the collection, storage, and analysis of data that we had in the 1980s. And it turns out that when you set those historical constraints aside, you can find data that is enormously valuable for assessing credit risk.
Cash flow data, sourced from applicants’ bank accounts, is the most obvious example. Larry Rosenberger, the former CEO of FICO and one of the creators of the original FICO Score, once said, “Cash flow data is gold. The only reason it wasn’t used at the beginning of FICO is because it wasn’t widely available.”
Second, credit bureau data is not, uniformly, trustworthy. While the credit bureaus used to maintain relatively strict requirements for lenders to both furnish and use data (the give-to-get model), that rigor has been lost in recent decades. The result is that there are quite a few companies today that furnish data without the natural incentive to keep the quality and predictive value of the data high. Credit builder products, which I have been screaming about in the newsletter for years, are the best example.
Third, and related to that prior point, the lack of financial incentive for furnishers makes it more difficult for the credit bureaus to collect data from all corners of the lending ecosystem, which further contributes to the unreliability of the data. The big pay-in-4 BNPL providers are the best example. In 2025, these companies (rightly) view their first-party customer data as a proprietary asset. All things being equal, they see no reason to spend money (furnishing data to Consumer Reporting Agencies) to give their valuable, first-party repayment data to their competitors.
Fourth, the credit bureaus do a terrible job enabling consumers to monitor their data, correct errors, and regulate how and when their data is used by lenders and other ecosystem participants. If you’ve ever tried to freeze your credit report or your kids’ credit reports, you know exactly what I’m talking about. If consumers’ financial data belongs to consumers, which I think most people would agree with, then the credit bureaus’ model for consumer empowerment leaves a lot to be desired.
What characteristics should a modern credit bureau have?
So, if we were going to solve for these problems, how would we do it? What characteristics would be important to focus on in designing a modern credit bureau?
I’ll offer five suggestions for our Credit Bureau of 2025:
1.) Collect more data. As we’ve discussed, there is a staggering amount of data outside the reach of the credit bureaus that is highly predictive of credit risk and compatible with the regulatory requirements that govern the lending industry. We will want the ability to ingest as much of that data as possible, which means we will want to replace the antiquated Metro 2 data specification with a more modern and flexible alternative.
2.) Focus on data quality and lineage. Not all data is created equal. We will want to understand the lineage of the data that is being furnished (Where is it coming from? What steps is the furnisher taking to ensure the quality of the data?) and make that information available to lenders so that they can make qualitative decisions about which data to use and how much analytic ‘weight’ to give it.
3.) Incentivize participation. It is in the best interests of consumers and lenders to furnish data. It creates a more accessible and competitive ecosystem. However, as the BNPL companies’ reluctance to furnish data to the credit bureaus and JPMorgan Chase’s demand to be paid for access to its open banking APIs demonstrates, financial services companies are highly reluctant to give their data away for free, even if doing so does benefit consumers and the overall ecosystem in the long run. I think there’s a strong case that we will want to pay data providers reasonable fees to align economic incentives and lead to better data and better data-sharing infrastructure over time. (BTW, it seems like the current trajectory of open banking in the U.S. is heading in this direction.)
4.) Move away from “One Score to Rule Them All”. As SOLO CEO Georgina Merhom explained in episode one of our new Source of Truth podcast series, over-standardization on a single credit score creates a lot of problems. A general-purpose credit score is a very useful tool, but it’s just one tool in the toolbox. Moving forward, I think we will want to move towards multiple credit scores, built around different datasets and designed to be used in different contexts (products, customer segments, etc.), rather than trying to force the industry to over-standardize on a single score.
5.) Empower consumers. The arc of history in the financial data economy is long, but it bends toward consumer control. The Fair Credit Reporting Act (FCRA) was groundbreaking in consumer privacy and data rights when it was passed in 1970; however, today, technology exists to extend consumer control over their personal data much further, and the need for that level of control will only become more critical in the decades to come. For our Credit Bureau of 2025, we will want to skate to where the puck is going by building out robust consumer permissioning, access management, and data quality and dispute management capabilities that go far beyond what today’s regulations require.
