Latest / The Operator's Manual with Fexingo: Daily Decisions, Tactical Wins, and Business Execution / How One Bank Stopped Fraud With a Simple Wait
Transcript
- Lucas: We usually talk about cutting costs or speeding up processes, but today I want to look at a case where doing the opposite actually saved the company money. It involves a mid-sized regional bank that decided to intentionally slow down its fraud detection system. Specifically, they introduced a ten-second delay before declining suspicious transactions. Luna: That sounds counterintuitive. If you suspect fraud, why wouldn't you block it immediately? Lucas: Right? The instinct is to hit the brake hard. But this bank found that their automated systems were flagging too many legitimate purchases as fraud. These are called false positives, and every time one happens, the customer gets annoyed, often cancels the card, and sometimes even leaves the bank. The cost of losing a customer is way higher than the risk of letting a few fraudulent charges slip through for a moment. Luna: So the delay was a buffer to allow the system to double-check without punishing the user? Lucas: Exactly. Instead of an instant rejection, the transaction would hang for ten seconds. During that time, the backend system ran a deeper check. If it was clean, the purchase went through smoothly. If it was truly malicious, the system still blocked it, but now with higher confidence. The key was that the customer didn't see a big red error message; they just experienced a slight pause. Luna: Did they track how many of those pauses resulted in actual fraud getting through? Lucas: Yes, and that is the beautiful part of the data. The additional fraud losses during those ten seconds were negligible. Like, barely measurable. But the reduction in false declines was huge. They cut false positive rates by sixty percent. That means sixty percent fewer angry customers calling support or switching banks. Luna: It feels like they treated patience as a feature rather than a bug. Lucas: Precisely. In product design, we always obsess over speed. Faster load times, quicker checkout flows. But here, adding friction actually improved the outcome. It’s a great example of how understanding human behavior changes the math. When you remove the embarrassment of a declined card at the register, you retain loyalty. Luna: I remember reading about similar strategies in subscription services, where they nudge users instead of locking them out immediately. Lucas: That’s a good parallel. But in banking, the stakes feel higher because money is involved. People get defensive when their card is declined. They assume the bank doesn’t trust them or that the system is broken. By inserting that brief pause, the bank signals that it is working, not failing. It shifts the perception from obstruction to protection. Luna: It’s interesting that they chose ten seconds specifically. Why not five? Or fifteen? Lucas: Ten seconds was the sweet spot identified in their A-B tests. Five seconds felt too short for the complex checks to run reliably. Fifteen seconds started to annoy customers who just wanted to buy coffee. It’s a classic optimization problem. You are balancing technical accuracy against user impatience. Ten seconds kept the accuracy high enough while staying under the threshold of irritation. Luna: It reminds me of how we wait for files to download. If it hangs for two seconds, we think it crashed. But if it shows progress, we accept the delay. Lucas: Exactly. Visibility matters. Did they show any kind of indicator to the user during those ten seconds? Luna: You mentioned earlier that customers didn't see a red error. Did they see anything at all? Lucas: They saw a subtle loading spinner. It was designed to look like the network was just being slow, not like the bank was interrogating the transaction. It normalized the delay. Most people assumed it was a temporary glitch. Very few realized it was a security feature. That anonymity of the process was crucial. Luna: It’s clever engineering disguised as a minor inconvenience. Lucas: I would argue it’s not an inconvenience if the alternative is having your card locked. But psychologically, the difference between a 'declined' notification and a 'processing' spinner is massive. One feels like a failure. The other feels like a routine step. Luna: Have other banks adopted this approach since then? Lucas: Some larger institutions have experimented with similar concepts, though rarely as explicitly. Most still rely heavily on rigid rules. But the trend is moving toward adaptive authentication. Instead of binary yes or no decisions, systems are learning to assess risk on a continuum. The ten-second pause is just one tool in that toolkit. Luna: It makes me wonder if we’re too focused on automation in other areas. Maybe some things need a human-in-the-loop delay. Lucas: Or at least a simulated human judgment. Algorithms are fast but brittle. Humans are slow but nuanced. This solution borrowed the nuance of human judgment and wrapped it in a fast interface. It’s a hybrid model that works better than either extreme. Luna: What was the initial reaction from the risk team? Didn't they push back on slowing things down? Lucas: There was definitely resistance. The risk team’s primary KPI was stopping fraud instantly. They viewed any delay as a vulnerability. It took months of data to prove that the additional risk was zero. Once they saw the retention numbers jump, the resistance faded. It’s a common dynamic in operations. You have to speak the language of the department you are trying to influence. Luna: Retention data is usually the universal translator. Lucas: Always. Customer lifetime value trumps instantaneous loss prevention in most cases. The bank calculated that keeping a customer for another year was worth far more than saving a few hundred dollars in potential fraud. Luna: It’s a reminder that business metrics are interconnected. Optimizing one number can hurt another unless you look at the whole picture. Lucas: Absolutely. If you only optimize for fraud prevention, you destroy customer experience. If you only optimize for experience, you bleed money to fraud. The ten-second pause is the equilibrium point. It’s a tactical win because it requires no new technology, just a rethinking of the workflow. Luna: I love that it’s a low-cost change with high impact. So much consulting advice is about expensive software upgrades. Lucas: Right. This was a configuration change. They tweaked the timing parameter in their existing gateway. No new vendors, no massive integration projects. Just a smarter use of what they already had. It’s the kind of decision that rewards attention to detail over budget size. Luna: Do you think this applies to e-commerce sites dealing with chargebacks? Lucas: It could. E-commerce faces similar false decline issues. Adding a small verification step, like asking for a zip code match or a quick email confirmation, might serve the same purpose. It creates a pause that allows for better signal collection without rejecting the sale outright. Luna: It feels like applying behavioral psychology to technical architecture. Lucas: That’s exactly what it is. Understanding how people perceive delay versus rejection changes the design of the system. It’s not just about code; it’s about cognition. Luna: Before we go further into the implications, I want to mention something important about how this show operates. Lucas: We deliberately don't run ads on these conversations. It keeps the focus entirely on the content you find useful. If you want to support that choice, the link is buy me a coffee dot com slash fexingo. It helps us keep producing these deep dives without compromising our editorial independence. Luna: It really does make a difference in maintaining the quality and consistency of the episodes. Thank you for considering it. Lucas: Back to the banking case. One thing that stood out to me was the communication strategy. Did they tell customers about the pause? Luna: No, I think you said they kept it anonymous. Why not advertise the enhanced security? Lucas: Because advertising it might draw attention to it. If customers know there is a pause, they might start waiting for it, which defeats the psychological effect of normalization. Also, mentioning security features can sometimes make people paranoid about their transactions. Keeping it invisible allowed it to work seamlessly. Luna: It’s like a seatbelt. You hope you never notice it until you need it, but you appreciate it silently. Lucas: Perfect analogy. And in this case, the seatbelt also improves the ride quality by reducing jolts—like false declines. It’s a passive safety feature that enhances the journey. Luna: I’m curious if this approach scales. Can large global banks do this without creating fragmentation? Lucas: Scaling is always harder. Large banks have legacy systems that are harder to tweak. But the principle is universal. Any organization that relies on automated decision-making can benefit from introducing calibrated delays. It’s a lever that exists in almost every digital workflow. Luna: So the takeaway isn't just for bankers. It’s for anyone designing a digital product. Lucas: Exactly. Whether you are building a loan application portal or a shipping tracker, consider where a brief pause might improve outcomes. Sometimes, slowing down is the fastest way to build trust. Luna: It challenges the obsession with speed. What other industries are rushing ahead when they should be pausing? Lucas: Healthcare comes to mind. Rushing patient intake forms can lead to errors. Adding a review step might save time downstream. Or social media algorithms, where slowing down viral spread of unverified info could reduce harm. The pattern repeats everywhere. Luna: It seems like the ability to tolerate short-term inefficiency for long-term stability is a rare skill in modern business. Lucas: It is. We reward velocity so heavily that we forget that direction matters more than speed. This bank chose the right direction, even if it meant taking a slightly longer route. And the data proved they got there faster in the end. Luna: A slower start, but a smoother finish. I like that narrative arc. Lucas: It’s a quiet victory. No press releases, no flashy tech demos. Just a ten-second delay that changed the calculus for thousands of customers. Sometimes the best operations are the ones nobody notices. Luna: Until the card doesn't get declined at the worst possible moment. Then everyone notices. Lucas: Precisely. And that’s when you realize the value of the invisible infrastructure. What other invisible pauses could we introduce in our own workflows?