Beyond Credit Scores: Using the Power of Alternative Data to Assess Commercial Credit Risk

Commercial lending is a bit like solving a puzzle. To understand the bigger picture of a business, namely hidden risks that might lead to default or indicate fraud, you must piece together the right data. For this, lenders are increasingly relying on alternative data sources, other types of data that are not included in traditional credit files, in addition to standard business credit data. 

Recently, Senior Vice President of the Equifax commercial business, Sal Hazday, sat down with John Fenstermaker, Chief Data and Analytics Officer at Equifax, to discuss this trend and explore how businesses are using alternative data to overcome challenges in their identity validation, credit risk, and fraud workflows.  

Hazday: What challenges are businesses trying to overcome that they can’t solve with traditional data alone? 

Fenstermaker: Let’s be clear: traditional credit data is still the gold standard for evaluating credit risk. But given today’s volatile economy and widespread access to advanced AI technologies, there is a widening “gray area” between credit risk and fraud that’s driving a need for expanded insights. Specifically, we’re seeing companies increasingly utilize alternative data for: 

• Business validation: Being able to verify a business entity using a trusted source, something that's not customer-supplied, is critical not only for credit risk but also for fraud risk. Standard commercial credit data might not be available when dealing with small businesses, which are often self-funded by a sole proprietor. Companies are feeding alternative data into an automated validation process to help fill these gaps and help them more quickly and confidently confirm that customers are who they say they are. 
• Machine learning model optimization: When combined with traditional credit data, machine learning models can achieve a higher level of accuracy and risk separation than traditional models alone. Feeding these models new streams of alternative data increases model accuracy and performance, while reducing fraud and default rates. For example, a commercial credit risk model may benefit from adding cash flow data to more accurately predict the risk of payment default or business failure. 

• Reducing underwriting friction: When the bulk of lending applications flow from point-of-sale transactions (in-person or digital), it’s difficult to push back on the customer to request additional information. They tend to abandon the transaction, and conversion rates tank. Folding alternative data into a waterfall decisioning process gives lenders a deeper understanding of customers and a more accurate risk ranking, at the exact moments when more insight is needed, without having to ask the customer. 

Hazday: What are you hearing about AI deep fakes and document validation, and how are commercial lenders using alternative data to help combat fraud? 

Fenstermaker: First-party fraud and deep fake documents are a huge problem for lenders today. As the name suggests, deep-fake documents (bank statements, business licenses, etc.) look incredibly authentic. Everything “looks right,” even the bank logos and Secretary of State seals. But these customers have no intention of ever paying back the loan. 

Lenders need alternative sources of data that can serve as a “source of truth,” something they can verify customer information and documentation against that is not provided by the customer. 

Some are using cash flow data that can be matched against PDF bank statements. Others are using merchant data to validate that a small business is active and that its revenue matches what’s stated on the application. Meanwhile, some are performing reverse look-ups and other forms of manual online research to validate customer-provided data.  

Hazday: What is Equifax doing today with alternative data? 

Fenstemaker: We’re constantly experimenting, innovating, and building out our alternative data solutions to help lenders solve their most pressing challenges. 

• Our Merchant Data Network is a powerful resource that’s growing every day. It provides alternative insights into businesses, including the prior month’s sales, trended sales for the last five years, current chargeback rates, chargeback trends over time, and much more. This data is extremely predictive in determining business legitimacy and assessing credit risk. Based on Equifax analysis, when we have tradeline information for a business, it's incrementally predictive, providing up to a 10 percent performance lift. But even when we lack trade line information, we're seeing as much as a 100 percent lift in credit scores and decision strategies. In general, using the Merchant Data Network can boost automated approval rates by up to 5 percent on that population.¹

• The Commercial Financial Network is another robust source of alternative data. Members get direct access to the commercial payment performance data of other members, including unparalleled coverage of equipment financing tradelines, along with unique data about business identity and ownership, business registrations, and company firmographics. It’s a gold mine of alternative insights that can be used for business verification, more secure and accurate underwriting, improved fraud detection, and more.  

• We also have our flagship commercial data resource: OneScore for Commercial. It combines our most potent data and technology into one of the best commercial delinquency scores on the market. Powered by patented machine-learning technology, it compiles data from the Commercial Financial Network, trended commercial data, public records, firmographic data, and more into a configurable score with 12 industry-specific scorecards stemming from NAICS and SIC codes. Since it uses both traditional credit and alternative data, Equifax analysis shows it can score up to 50 percent more applications, including customers that lack business credit. In addition to the commercial-only version of OneScore, there is also a version that combines data on the business with data on the guarantor, and this version outperforms the high-performing, commercial-only version by as much as 20%.¹

Hazday: What types of alternative data resources can we look forward to in the future? 

Fenstermaker: Payroll data is on our radar. Not only is it a quality alternative source of data that can help confirm and augment customer-supplied information, but ideally it could be easily integrated into an automated process to improve decisioning performance without introducing added friction. 

Inquiry data, which are already used to assess risk, is another area worth exploring further. More specifically, how can we better mine our inquiry data to identify complex patterns and trends that are indicative of fraud or credit risk? AI will unlock new opportunities to extract more value from all our data assets, and these new AI-powered insights will power the next generation of risk solutions. 

At the end of the day, commercial lenders want to help small businesses gain greater access to the financial services they need. Yet, they also need to protect their bottom line against losses. Incorporating alternative data into their business verification, underwriting, and fraud mitigation processes is clearly helping them achieve both goals by providing more complete views of their customers. 

Discover how Equifax can help commercial lenders with innovative data-driven solutions. 

1. Individual results may vary.