Jon Jacobson is CEO of privacy-preserving data collaboration platform business Omnisient.
getty
Tens of millions of people remain invisible to banks and lenders, rendering them ineligible for life-changing credit products ranging from mortgages to small business loans. In the U.S. alone, an estimated 45 million consumers have thin or non-existent credit files.
At the heart of this persistent problem is a data gap. Lenders rely on predictive signals to assess risk, but traditional models draw heavily on data generated within the credit system itself—information that millions of people simply do not produce, because they’ve never been granted credit. If access to credit relies on having taken out credit before, it creates a circular problem.
This represents both a financial inclusion challenge and a growth constraint for lenders. Institutions are unable to extend credit, not necessarily because of elevated risk, but because of insufficient visibility into that risk.
One of the most powerful sources of behavioral signals now being used by leading banks to assess risk, particularly where no credit history exists, is SKU-level grocery data. Grocery shopping data is universal, frequent, recent and rich, especially at the SKU level.
Patterns in affordability, financial resilience and life stage can provide meaningful context that complements traditional credit data, particularly for consumers with limited credit histories.
Importantly, these signals are not derived from individual transactions in isolation.
Their value lies in aggregated patterns and statistically validated relationships across large populations, patterns that can now be identified and modeled at scale using AI.
Grocery purchase data is, therefore, not a replacement for traditional credit data, but a highly predictive source of additional signals that can improve model accuracy when used responsibly.
The challenge is not whether these signals exist, but how to use them safely.
Consumers are understandably cautious about sharing detailed purchase histories, and banks face strict regulatory requirements around privacy, fairness and data governance. Simply moving or sharing raw transaction data is neither practical nor desirable.
The solution lies in separating insight from identity.
Through privacy-preserving data collaboration within data clean rooms, lenders can overlap their data with anonymized shopping data accessed directly from a grocery retailer and analyze it in a neutral and controlled environment without transferring or exposing the underlying data. Data is never exchanged, only anonymized insights.
In this model, detailed transaction-level data remains protected at source. AI models are applied within a secure environment to extract predictive signals, and only those derived insights—not the raw data—are made available to the lender.
This means banks can benefit from the richness of SKU-level data without ever taking custody of a retailer’s data set or handing over their own dataset to a third party, significantly reducing both privacy risk and regulatory exposure for all.
In practice, this approach is already delivering measurable results.
In one case I've seen, the introduction of grocery shopping data as alternative data for credit scoring improved the predictive model’s accuracy by more than 40% in assessing the likelihood of loan repayment among applicants with little to no credit history.
In practical terms, this allows banks to more accurately distinguish between applicants who are likely to repay and those who are not, even when no credit history is available.
With this enhanced predictive capability, banks can confidently expand into segments that were previously considered too risky, growing their credit portfolios while maintaining or improving risk performance.
For consumers, the implications are significant. Access to credit can enable financial resilience, support setting up a small business or provide a pathway to home ownership. These are opportunities that remain out of reach when individuals cannot be properly assessed.
New approaches to credit risk don’t mean lowering standards. They mean making better-informed decisions with supplemental data that provides a more complete picture of how people actually manage their money.
Used properly, with the right privacy safeguards in place, SKU-level grocery data can be effectively used by lenders as an alternative data source for credit scoring to expand access to credit without taking on additional risk, and give millions more access to affordable financial services.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。