Small Business Credit: Still Broken, Still a Huge Opportunity

The small business credit gap in the United States is not a new problem. Banks started pulling back from sub-$250K business loans after 2008, and they never came back. The economics didn’t work for them — the cost of underwriting a $150K loan is not dramatically different from the cost of underwriting a $5M loan, and the margin on the smaller loan is a fraction of the larger one. So banks moved upstream, and small businesses were left with credit cards, merchant cash advances at triple-digit APRs, and the occasional SBA loan that takes three months to process.

That gap is somewhere between $300B and $400B annually in unmet credit demand. It has not closed. If anything, it has widened as banks have continued consolidating and community banks have continued disappearing. The fintech companies that have attacked this market have made genuine progress, and several have built durable businesses. But the market is still large enough that the opportunity is not crowded.

Why Traditional Underwriting Fails Small Businesses

The bank underwriting model was built for standardized financial documentation: audited financials, tax returns, DSCR calculations on historical cash flows. That works fine for a restaurant with three years of audited P&Ls. It doesn’t work for a two-year-old landscaping company whose books are in QuickBooks and whose owner pays themselves irregularly.

The data that accurately predicts small business creditworthiness is often not the data in traditional credit files. It’s cash flow patterns in the business bank account. It’s payment history to suppliers. It’s receivables aging. It’s customer review quality as a proxy for business health. None of this shows up in a credit bureau pull or a tax return review.

The Approaches That Are Working

We’ve seen three distinct models demonstrate genuine loss performance at meaningful scale:

Revenue-based financing. Structuring repayment as a percentage of daily or weekly revenue rather than a fixed monthly payment addresses the cash flow volatility problem that kills small businesses on traditional term loans. When revenue is lower, repayment is lower. When revenue recovers, repayment catches up. For businesses with seasonal or lumpy revenue, this structure dramatically reduces the risk of payment default during slow periods that have nothing to do with credit quality.

Embedded lending at the point of workflow. The lenders with the best data and the best conversion rates are not cold-acquiring customers. They’re embedded in the software the business is already using. A payment processor that offers working capital to merchants using its platform has 18 months of transaction history, knows the business’s seasonality, and can see cash flow in real time. That data advantage translates into better underwriting and better approval rates for creditworthy borrowers who would be rejected by a traditional lender.

Industry-specific credit products. A generic small business lender has to build underwriting models that work across every industry simultaneously. That’s hard. A lender focused on, say, medical practices, or construction contractors, or e-commerce sellers on a specific platform can build underwriting models tuned to the specific cash flow patterns, receivables structure, and risk factors of that industry. Vertical-specific lenders consistently outperform generalists on loss rates in comparable credit quality bands.

The companies we’ve invested in who are winning in SMB lending share one thing: they have better data than the alternatives, and they’ve built the underwriting to use it. That’s not a technology story. It’s a data access story combined with the credit judgment to know what signals matter.

The Open Banking Effect

The slow rollout of open banking data access in the US is directly relevant here. When small businesses can permissioned-share bank account data with a lender in real time, the information asymmetry that makes small business underwriting hard largely disappears. The lender can see actual cash flows, not the lagged picture from tax returns. They can see current account balances, not a snapshot from six months ago.

Several SMB lenders we follow are building underwriting models that rely heavily on real-time bank data rather than traditional credit signals. Early loss performance on these models is promising, particularly for the segment of small businesses that have good underlying cash flows but poor credit bureau scores because of historical events or thin file history.

Where the Risk Sits

The risk in SMB lending is not the underwriting in a normal environment. It’s the underwriting in a downturn. Small businesses are disproportionately impacted by recessions, and the alternative data signals that work well in stable conditions can diverge rapidly in a credit cycle turn. The companies that have built through-the-cycle loss models — and priced accordingly — will perform differently from the ones that optimized purely for growth in a benign environment.

We look hard at loss vintage analysis when evaluating SMB lenders. What did 2020 look like? What did 2022 look like, when inflation hit small business margins hard? The companies with honest answers to those questions have earned the confidence to deploy more capital. The ones who haven’t been through a cycle yet are taking risk they may not have fully modeled.

The opportunity in small business credit is real and durable. The banks are not coming back for this market. The fintech companies that build with real credit discipline — and sustainable cost of capital — have a chance to serve 33 million small businesses that have been underserved for fifteen years.