Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / How Fintechs Use Real-Time Gig Work Data for Dynamic Income Verification
Transcript
- Lucas: So there's this number that keeps sticking with me — over sixty million Americans now earn at least some income through gig platforms, and yet most lenders still underwrite them like they're freelancers from the nineties. Luna: You mean the classic two years of tax returns and a prayer? Lucas: Exactly. And that's the gap fintechs are starting to attack. Today I want to dig into how some lenders are using real-time gig work data to verify income dynamically — not just at application time, but continuously. So instead of a snapshot from last April, they're seeing earnings as they happen. Luna: That sounds powerful, but also a little Big Brother. How does it actually work in practice? Lucas: Good question. The typical setup is via API connections to gig platforms. Think Uber, DoorDash, Upwork, TaskRabbit. With the worker's consent, the lender pulls earnings data directly — gross earnings, net pay, number of trips, even tips broken out by week. Some platforms have dedicated developer portals for this. Luna: So it's not just a bank statement screenshot that anyone could fake? Lucas: Right. This is verified, structured data straight from the source. And the key insight is that lenders aren't just looking at the average — they're looking at the volatility. A gig worker might earn three thousand dollars one week and eight hundred the next. A traditional underwriter would see that as risk. A dynamic model can actually price for it. Luna: So they're using the real-time data to better assess risk, not just to say yes or no. Lucas: Precisely. And that opens the door to products that were basically unavailable to gig workers before. I'm thinking auto loans, small personal lines, even mortgages in some cases. Luna: Wait, mortgages? That's a big jump. I know some lenders are experimenting with that, but how do you underwrite a mortgage on income that fluctuates week to week? Lucas: That's exactly what a fintech called Upstart tried to crack a few years ago, and more recently we've seen players like Figure and some credit unions piloting programs. The approach is to use, say, ninety days of real-time earnings data, then apply algorithms that look at patterns — seasonality, weekly cycles, even the impact of local events like a snowstorm that might spike delivery demand. Luna: So they're building a predictive model of future income, not just averaging the past. Lucas: Yes. And that's where the dynamic part really kicks in. After the loan is originated, they can continue to monitor earnings. If a driver's income dips for two weeks, the lender might proactively offer a payment deferral or a revised plan — instead of waiting for a missed payment. Luna: That's a huge shift from the traditional 'set it and forget it' loan servicing model. Lucas: It is. And it creates this interesting alignment: the lender has a strong incentive to keep the borrower earning. Some are even partnering with gig platforms to offer financial coaching or tips on when to drive to maximize income. Luna: If today's conversation is helping you think differently about fintech and lending, that's exactly what we're here for. This show stays ad-free because of listener support — so if you found something useful, you can buy us a coffee at buy me a coffee dot com slash fexingo. It genuinely keeps these conversations going. Lucas: Yeah, and it's a meaningful way to keep the insights flowing without any corporate sponsors. So, back to the data — I want to talk about a specific case that really shows the potential. Luna: Please, I'm all ears. Lucas: So there's a driver in Austin — let's call him Marcus. He'd been driving for Uber and Lyft for about two and a half years, but he wanted to upgrade to a newer car. His credit score was decent, around 680, but his tax returns showed only twenty-eight thousand dollars in net income. A traditional auto lender would have put him in a subprime bucket with double-digit rates, if they approved him at all. Luna: Classic story. So what did the fintech do differently? Lucas: This lender — I'm not going to name them because they're still piloting — used a real-time data aggregator to connect to Marcus's Uber and Lyft accounts. Over the next ninety days, they saw his actual earnings: about forty-five thousand dollars gross, with clear weekly patterns. His best weeks were around South by Southwest, and he tended to work longer shifts on weekends. Luna: So they could see he wasn't just a random earner — he had a solid, repeatable income stream. Lucas: Exactly. Based on that, they offered him a loan at a rate nearly four percentage points lower than what he'd been quoted elsewhere. And critically, they didn't just use the average — they modeled his worst month and made sure the payment was still comfortable. Luna: That's a great example of using data to find the real risk profile instead of defaulting to a broad bucket. Lucas: And the story doesn't end at origination. Six months in, Marcus hit a slow patch — maybe a new competitor in his area. The lender's monitoring flagged a thirty percent drop in earnings over two weeks. Instead of a late payment, they sent an alert and offered to temporarily reduce his payment by half for that month. Luna: Wow. That's proactive, but it also raises the question — how much privacy are borrowers giving up for that kind of safety net? Lucas: It's a legitimate trade-off. The data is shared with consent, but the continuous monitoring is a different beast than a one-time credit pull. Regulators are starting to look at this under fair lending and data privacy rules. The CFPB has been quiet so far, but I think we'll see more guidance in the next year or two. Luna: So there's a balance to strike between financial inclusion and surveillance. Lucas: Exactly. And that's where the smart fintechs are differentiating themselves. They're being transparent about what data they collect, how it's used, and giving borrowers control — like the ability to pause data sharing or see a summary of what the lender is viewing. Luna: That's reassuring. But what about the other side — the gig platforms themselves? Are they all on board with sharing this data? Lucas: Mixed bag. Uber and Lyft have been relatively open through their developer APIs. DoorDash has been more cautious, probably worried about driver privacy and potential liability. Upwork, interestingly, has been proactive — they see it as a way to attract higher-quality freelancers who can get better credit. Luna: Makes sense. If you're a platform, enabling your workers to get loans might increase loyalty and reduce churn. Lucas: Right. So there's a real ecosystem forming here — fintechs, platforms, data aggregators, and even credit bureaus. I'm seeing new startups focused specifically on this niche, like Level and Grain, though they're small. Luna: And what's the biggest technical challenge? I'd imagine standardizing data across so many platforms is a nightmare. Lucas: You'd be right. Each platform has a different API structure, different data fields, different update frequencies. Some update earnings in real time, others only daily. And then you have to deal with gaps — like when a driver takes two weeks off. The algorithms need to distinguish between a vacation and a decline in business. Luna: So it's not just about connecting to the data; it's about interpreting it in context. Lucas: Yes. And that's where machine learning comes in. Some lenders are training models on historical gig worker data to predict which patterns are temporary and which are structural. The goal is to avoid false alarms that might cause unnecessary stress or, on the flip side, catch real problems early. Luna: That's sophisticated. So where do you see this going in the next few years? Lucas: I think we're moving toward a world where income verification becomes as continuous as the gig economy itself. The old model of annual tax returns will fade for this segment. We'll likely see more mainstream banks adopt similar approaches, not just fintechs. And maybe, eventually, this dynamic data will feed into credit scores themselves, not just loan decisions. Luna: That would be a fundamental shift in how we think about creditworthiness. Lucas: It would. And it raises a big question — if we can do this for gig workers, why not for salaried employees? Their income is usually more stable, so the data might be even more predictive. But that's a conversation for another day. Luna: Funny you say that. I recall you mentioning a similar idea in episode 148 about payroll data. Lucas: You caught me. The principles are the same, but the gig economy adds that extra layer of volatility that makes it a more interesting puzzle. Luna: Fair point. So for a listener who's a gig worker or a lender listening in, what's the one takeaway? Lucas: I'd say this: real-time income data isn't just about better risk management — it's about unlocking access to credit for a huge population that the old system systematically excluded. And if you're a gig worker, don't assume a traditional lender is your only option. Look for fintechs that understand your income model. That's the future.