Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / How Fintech Is Solving the Small Business Lending Gap
Transcript
- Lucas: If you're a small business owner and you walked into a bank branch in 2025 asking for a fifty-thousand-dollar loan, the bank would want three years of tax returns, a personal guarantee, and probably a lien on your house, and even then you'd wait two to three weeks for an answer. Luna: And that's if you had perfect credit. If you were a newer business or had a few late payments, forget it. Lucas: Exactly. Now contrast that with a coffee roaster in Portland, Oregon—three years old, maybe a couple million in annual revenue through Square's payment system. They applied for a loan through Square Capital last year, and within ten minutes they had fifty thousand dollars deposited into their account, with repayment set as a fixed percentage of daily card sales. Luna: Ten minutes. That's the gap fintech is trying to close. And if these conversations are useful for what you're building or running, it's worth mentioning that we keep this show entirely ad-free by design. No sponsor reads, no ad breaks. Lucas: Yeah, that's intentional. We think the content should stand on its own. If you find value in episodes like this and want to support that choice, there's a simple way to do it — buy me a coffee dot com slash fexingo. No pressure, just an option for folks who want to keep this going. Luna: Exactly. So back to that coffee roaster — how did Square know they were good for fifty grand without asking for a single tax return? Lucas: They looked at the transaction data. Square processed every single one of that roaster's card sales for three years. They could see daily revenue trends, seasonal dips, repeat customer frequency, even average ticket size. That data is way more granular than what a bank gets from a credit report. Luna: So the underwriting is based on cash flow in real time, not on a static score. Lucas: Right. And the repayment structure mirrors that cash flow. Square takes a fixed percentage — typically between eight and twelve percent of daily card sales — so if the roaster has a slow Tuesday, they pay less. If they have a huge holiday rush, they pay more. There's no fixed monthly payment that could break them. Luna: That's a huge shift. Traditional bank loans have fixed principal and interest, which doesn't care if your business is seasonal. Lucas: Exactly. And the numbers suggest this model works. Square's internal data from its public filings a couple of years ago showed default rates around four to five percent on these loans — comparable to, or even better than, many small business loan portfolios at big banks. And because Square can adjust the percentage in real time, they can reduce the risk of the borrower getting underwater. Luna: But isn't that only possible because Square owns the payment flow? What about a business that uses multiple processors? Lucas: That's the big limitation. Square Capital works best for businesses that process all their card sales through Square. If a restaurant uses Square for in-store but also takes online orders through a separate gateway, Square only sees part of the picture. That's where newer entrants are trying to build a more complete view by aggregating data from multiple processors or using bank account transaction data via Plaid. Luna: So there's a data moat. The more you process through one platform, the better the loan terms you can get. Lucas: Exactly. And that creates an incentive for businesses to consolidate their payment processing. But it also raises questions: what if you don't want to be locked in? And what about privacy? Some regulators have started asking whether these lending models are effectively using transaction data in ways customers didn't fully consent to. Luna: The CFPB has been looking into this, right? Lucas: Yeah, the Consumer Financial Protection Bureau issued a guidance early last year on data aggregation and consent for small business lending. They're trying to make sure that fintechs are transparent about what data they're using and that borrowers can access their own data easily. Luna: Let's talk about the specific mechanics. How does Square actually decide the loan amount and the percentage? Lucas: It's an algorithm trained on millions of businesses. Key inputs include: average monthly processing volume, length of processing history, industry type, and a proprietary risk score that incorporates things like chargeback rates and seasonal volatility. They also look at how consistent the revenue is month over month. A coffee shop with steady daily sales gets a better rate than a landscaping business that spikes in summer and drops in winter. Luna: So the algorithm is effectively rating the business's revenue stability. Lucas: Yes. And it's dynamic. If the business's volume drops for a few months, the loan payments automatically adjust downward. That's a huge advantage over traditional debt. But there's a catch: the effective interest rate can be quite high. Square doesn't call it interest — they call it a 'flat fee' — but if you annualize it, the cost of capital can work out to twenty to thirty percent APR for a loan repaid in twelve months. Luna: That's expensive. But for a business that can't get a bank loan at all, it might be worth it. Lucas: That's exactly the trade-off. And it's not just Square. Stripe Capital operates similarly — they offer loans to businesses that process through Stripe. The difference is Stripe tends to focus on e-commerce and SaaS businesses, while Square is strong in retail and food service. Luna: What about the SBA's new fintech partnership program? I read that they're trying to get these kinds of lenders to participate in government-backed lending. Lucas: Yes, the SBA launched a pilot last year called the 'Fintech Lender Partnership' that allows approved fintechs to originate SBA 7 loans using their own underwriting models, with the SBA guaranteeing up to seventy-five percent of the loan. The idea is to combine the speed of fintech with the lower cost of government-backed lending. Luna: So potentially lower rates than Square Capital's flat fee, but still fast. Lucas: That's the hope. A few companies have been approved so far — including a couple of the payment-based lenders. But it's still early. The challenge is that SBA loans have complex compliance requirements, and the fintechs have to integrate with the SBA's systems, which are not as modern as what they're used to. Luna: Let's go back to that Portland coffee roaster. What happened when they took the loan? Did it work out? Lucas: According to the case study Square published, the roaster used the fifty thousand to buy a second roasting machine and expand their wholesale distribution. Their revenue grew about thirty percent over the next year, and they paid back the loan in about nine months. Square's cut was around six thousand dollars in fees — so about twelve percent of the principal. Luna: So a win-win. But not every story ends that way. Lucas: No, and the critics point out that the model can be pro-cyclical. If a business hits a rough patch and revenue drops, the payment amount drops too — that's good. But if revenue keeps dropping, the loan takes longer to repay, and the fees keep accruing. There's also the risk that the fintech can cut off future credit just when the business needs it most. Luna: And there's the lock-in effect. If you have an outstanding loan with Square, you can't easily switch to another processor without potentially triggering a default. Lucas: Exactly. That's a real concern. Some businesses feel trapped. But overall, the data suggests that access to this kind of credit has been a net positive for millions of small businesses that were previously unserved by banks. Luna: So where do you see this going in the next year or two? Lucas: I think we'll see more platforms open up their lending to third-party data sources. If a business gives permission, a lender could look at their bank account data, their accounting software, their e-commerce platform, and their payment processor all at once. That gives a much richer picture and potentially better rates. Luna: So it becomes a data marketplace, with the business controlling access. Lucas: Right. The technology is already there. The question is whether regulation will encourage that openness or push toward more siloed models. And whether businesses will trust the system enough to share that data. Luna: It's a fascinating shift. Lending used to be about collateral and credit scores. Now it's about data and algorithms. Lucas: And for a coffee roaster in Portland, that difference is the difference between getting a loan in ten minutes versus not getting one at all.