Editor’s Note — This article is sponsored by Nova Credit. As with all sponsored content in Fintech Takes, this article was written, edited, and published by me, Alex Johnson. I hope you enjoy it!
There comes a moment, in the development of any new technology, where it makes sense to stop thinking in terms of use cases and start thinking in terms of infrastructure.
The challenge is in identifying this moment.
You don’t want to make the switch from use cases to infrastructure too early. New technologies need to prove their value before they deserve the organization-wide attention and investment reserved for core infrastructure. And it can take a while for new technologies to prove themselves, especially in financial services, where costs are often felt indirectly and the consequences of excessive risk taking are only realized months or years afterwards.
Still, at a certain point, it makes sense to start treating new technology as infrastructure.
Take customer relationship management (CRM) software as an example.
We’ve been using computers to help salespeople keep track of prospects and manage deals as they progress through the pipeline for 40 years. However, it wasn’t until the development and widespread adoption of SaaS-based CRM platforms like Salesforce, that companies began to realize that CRMs aren’t a technology for discreet sales use cases, they’re infrastructure — a system of record — that underpins the entire business. Once this realization hit, companies started asking very different questions. It became less about features and functionality, and more about architecture (What is the data model?), integrations (How do we tie our CRM and ERP systems together?), and operational efficiency (How do we reduce duplicate work? How do we optimize the number seat licenses that we need?)
In consumer lending, we’ve been using cash flow data to assess risk for a very long time, and, in its modern, system-to-system form (often facilitated by open banking) for more than a decade. However, it’s only within the last year or so that I have observed lenders beginning to think of cash flow data, not in terms of use cases, but in terms of infrastructure. Specifically, I’m beginning to hear the term “cash flow intelligence” a lot more, so I thought it might be worth defining that term more precisely and explaining why it’s becoming a more popular way to think about this category of financial technology.
Let’s do so in the form of a quick Q&A.
What is cash flow intelligence?
Cash flow intelligence is the practice of turning bank transaction data into continuous, timely insights about a consumer’s financial health, and applying that insight across the entire credit risk lifecycle.
Why are lenders starting to think about the use of cash flow data holistically, rather than for specific use cases?
Probably because it has been so successful when applied to individual use cases.
Underwriting is obviously the most common use case, and all of the available research (such as this report from FinRegLab) confirms what our intuition already tells us: Cash flow data provides a significant predictive lift, on top of traditional credit bureau data, for accurately assessing consumer credit risk, across the entire risk spectrum.
Put more simply, it works, which is why lenders are incorporating it into their underwriting processes, either as a second-look for applicants who would otherwise be declined, or (if the lender also has depository relationships) by leveraging “on-us” bank transaction data for all applicants where it is available, at the beginning of the underwriting process.
But underwriting is just the beginning.
Given its orthogonal (i.e., not duplicative) predictive power, relative to traditional credit bureau data, lenders are beginning to use cash flow data everywhere that they use traditional credit bureau data. Before underwriting, when lenders’ need to efficiently find and pre-qualify prospective borrowers. And after underwriting, using borrower’s real-time cash flow data to proactively identify problems to mitigate or opportunities to cross-sell into.
Hence the need for infrastructure; an intelligence layer that sits underneath the entire consumer credit risk lifecycle and that can power any number of different use cases across that lifecycle, in the most efficient and effective manner possible.
What infrastructure questions should lenders be asking about cash flow data?
Here are the four questions I hear most frequently (with a few of my own thoughts on how to answer them).
- If we’re using “off-us” cash flow data (permissioned by the consumer) across lots of different use cases, how do we optimize the user experience? You obviously would prefer not to be constantly asking consumers to re-permission access to their data, as that is a bad experience that introduces unnecessary friction. You want to think about how you can combine multiple requests for permission into one, and how you can pass verified consumer permission between systems and even between different companies that you may partner with across the lifecycle.
- How do we optimize for our own costs? This is an important question because accessing cash flow data doesn’t just introduce friction into the user experience, it also costs money. Multi-aggregator systems (i.e., the ability to dynamically switch between Plaid, MX, Akoya, etc.) can be used to optimize costs, as well as reliability, and savvy lenders generally look to reuse data, as long as it’s fresh and permissioned, across use cases rather than pulling it new every time.
- How do we ensure that cash flow insights (including on-us insights) are shared across the company instead of being trapped in silos? At an analytics level, custom features and models can produce better results for individual lenders that are using cash flow data in specific use cases. Shared libraries and decision governance infrastructure allow insights from on-us data to be built into these features and models and be reused across the company. Build the infrastructure to facilitate sharing.
- How do we navigate a perpetually uncertain regulatory environment? This is the most critical question to ask on an infrastructure level because the more you do with cash flow data, the greater your overall exposure to compliance risk. The key is to adhere to the principles of existing laws (e.g., Dodd-Frank’s mandate for consumer control) in the absence of specific rules and regulatory guidance and to exercise caution when wrestling with any genuinely new compliance questions (e.g., How do fair lending laws apply to bank transaction data?)
There’s more architecture to sketch out here. I’ll be digging into all of this, and more, at the Cash Flow Intelligence Summit on September 10th.
