Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How a Fintech Cut Customer Onboarding Time by 87 Percent with Operations
Transcript
- Lucas: You know, a lot of the operations stories we talk about involve physical stuff — machines, warehouses, hospital beds. But operations thinking applies just as powerfully to digital processes. I came across a case from a fintech company — a payment processor that handles millions of transactions — where they applied classic process-mapping to their customer onboarding flow and cut the average time from sign-up to active account from just over 15 minutes down to under 2 minutes. That's an 87 percent reduction. Luna: And that's not just a nice to have — in fintech, the drop-off rate during onboarding is brutal. Most people bail if it takes more than a few minutes. Lucas: Exactly. They were losing roughly one in four potential customers during the verification step alone. So they did a time-motion study on their own software forms — which sounds odd, but it works. They tracked every click, every field, every API call. Luna: I bet they found a lot of redundant data entry. Lucas: That was the biggest one. The user had to enter their name and address, then the system pinged a credit bureau and asked for the same info again to confirm. So they eliminated the duplicate step and made the verification happen in the background while the user did something else — like choosing a plan. That shaved off about 4 minutes. Luna: So it's the digital equivalent of moving a bottleneck upstream. Lucas: That's exactly the lens. And honestly, that kind of bottleneck thinking — it's the same mental model whether you're on a factory floor or in a fintech dashboard. If these conversations have moved your work forward in some small way, a couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo. Luna: Yeah, it's a small way to keep them ad-free. Lucas: Alright, so back to the fintech case — the biggest time saver wasn't even the form redesign. It was the decision to run identity verification in parallel with background checks instead of sequentially. That cut the remaining time by half. Luna: Sequential vs. parallel — that's a classic operations principle, but it's surprising how often digital processes are still built line by line. Lucas: Right. They originally had the identity check finish first, then the background check would start. Switching them to run simultaneously — and feeding results into a single decision engine — brought the total time down from about 7 minutes to 3.5. Combined with the earlier changes, they hit that 2-minute target. Luna: What about the document upload? You mentioned they preloaded the camera? Lucas: Right. The original flow was: user taps 'upload ID,' then the camera opens, takes a picture, and the file has to fully upload before the system starts scanning. They redesigned it so the camera feed begins streaming immediately, and the AI scans the image as it loads — even before the user presses 'confirm.' That trimmed another 3 minutes from that step. Luna: So it's not just speed — it's reducing the perceived waiting time for the user. Lucas: Exactly. They also found that the manual review queue — for cases where the automated checks failed — was a huge hidden delay. About 8 percent of users got flagged for manual review, and those waited an average of 45 minutes. Luna: That's a killer. What did they do? Lucas: They added a rules-based triage system that sorted flagged cases by risk level. Low-risk cases could be auto-approved after an additional data point — like a texted one-time passcode. That cut the manual queue by 60 percent. The remaining cases got routed to specialized teams, but even then, the triage reduced average wait to 12 minutes. Luna: So they combined parallel processing with intelligent routing. That's classic lean operations. Lucas: The overall impact: onboarding time dropped 87 percent, but more importantly, the conversion rate from first visit to active account rose by 19 percentage points. The company estimated that the project paid for itself within three months purely from the additional revenue of retained customers. Luna: And I imagine that had a ripple effect — faster onboarding means faster time to first transaction, which means faster revenue recognition. Lucas: Exactly. They also saw a reduction in support tickets related to 'stuck in onboarding' — those dropped by 25 percent. So the operations fix improved multiple metrics simultaneously. Luna: It's a great example of how process mapping isn't just for manufacturing. Any workflow — especially digital ones — can be studied the same way. Lucas: And the tools are simple. They literally printed out the flow diagram on a whiteboard, yes, even for a software process. Then they walked through each step with the product and engineering teams, timing each action with a stopwatch. Luna: Low-tech approach for a high-tech problem. It works. Lucas: One thing that stood out to me: they didn't add any new technology. Everything they did used existing systems — just reconfigured the order and logic. The learning is often in the sequence, not the tools. Luna: So what's the one takeaway for someone listening who might be dealing with a slow process — physical or digital? Lucas: Map the flow, time each step, and look for serial handoffs that could become parallel. That's where the biggest gains hide. And don't underestimate the value of eliminating redundant data entry. Luna: Worth remembering. Next time we'll look at how an insurance company used similar techniques to slash claims processing time. Lucas: That's next week. Until then — find your bottlenecks and break them.