Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How a Warehouse Cut Picking Errors by 90 Percent with Voice Technology
Transcript
- Lucas: Last year, a mid-sized distribution center in Ohio was sitting on an eight percent picking error rate. That meant one in every twelve items shipped was wrong — wrong item, wrong quantity, wrong location. Luna: That's wild. I mean, for a warehouse moving tens of thousands of orders a day, that's hundreds of customer complaints, returns, re-ships. Lucas: Exactly. And the hidden cost is even bigger — each error costs roughly twenty-two dollars to fix, between return shipping, restocking, and customer service time. So this facility, call it a hundred thousand picks a day, was bleeding over seventeen thousand dollars daily in mistakes. Luna: So what did they do? Throw more scanners at the problem? More training? Lucas: They did something simpler and smarter. They switched from handheld barcode scanners to voice-directed picking. Workers wear a headset and a waist-mounted computer. The system tells them, in natural language, where to go and what to pick. They confirm by speaking back a check digit from the location label. Luna: So no more stopping to scan. Hands-free, eyes-free. That speeds things up, but does it actually reduce errors? Lucas: Dramatically. Within three months, their error rate dropped from eight percent to under one percent. A ninety percent reduction in errors. The voice system eliminated the two biggest sources of mistakes — misreading a label and keying the wrong quantity on a terminal. Luna: I've heard of voice picking before, but it always seemed like one of those older technologies that got pushed aside by automation — robots, conveyor sorters. Lucas: It is older. Vocollect, the company that pioneered this, was bought by Honeywell ages ago. But it's having a comeback, especially in e-commerce fulfillment where order profiles are chaotic — single-item orders, mixed SKUs, constant new products. Voice adapts faster than fixed automation. Luna: And it's much cheaper to deploy. No conveyor belts, no robotic arms. Just a headset and software. Lucas: Right. The Ohio facility spent roughly two hundred thousand dollars on the voice system. Compare that to a multi million dollar automated sortation system. Their payback period was under five months, purely from error reduction. Luna: But I imagine there was some resistance from the floor. Workers who've been scanning for fifteen years don't want to change. Lucas: A lot. The biggest pushback came from the veteran pickers. They said wearing a headset all day would be uncomfortable, that talking to a computer felt weird. The operations team handled it by running a pilot with six volunteers, letting them test it for a week. After three days, all six refused to go back to scanners. Luna: Why? What changed their minds? Lucas: The hands-free aspect. Pickers said they felt less fatigued because they weren't constantly raising and lowering a scanner. They also appreciated not having to remember location numbers — the system guided them step by step. Productivity went up about fifteen percent too, because they were moving continuously. Luna: So it's a win-win. But are there any downsides? I've read that voice systems can struggle with ambient noise — fans, forklifts, beeping. Lucas: That used to be a bigger issue. Modern voice engines are much better at filtering background noise. The Ohio facility runs twenty-four hours with forklifts, conveyors, and blowers, and the error rate stayed below one percent. The system learns each worker's voice profile, so it adapts to accents, breathing patterns, even a cold. Luna: That's impressive. So what's the catch? Why hasn't every warehouse adopted this? Lucas: Two reasons. First, integration with existing warehouse management systems can be tricky. Some older WMS platforms don't play nicely with voice middleware. Second, change management. It's not a technology problem — it's a people problem. Getting supervisors to trust the data, retraining trainers, rewriting standard operating procedures. Luna: That resonates. A lot of operational improvements are really about culture, not tech. Lucas: Exactly. The Ohio facility's operations manager told me the hardest part wasn't installing the system — it was convincing the shift leads that they didn't need to micromanage pick routes anymore. The voice system does that automatically. Luna: So the leads had to redefine their own jobs. That's a big ask. Lucas: It is. But once they did, the leads started focusing on coaching and problem-solving instead of policing. That's when the real improvements happened. Luna: I want to zoom out for a second. Voice picking has been around since the nineties. Why is it only getting traction now? Lucas: Two catalysts. One: e-commerce exploded, and with it, the cost of errors. A wrong item shipped to a consumer triggers a free return, a refund, and a lost customer. The tolerance for mistakes is near zero. Two: the hardware got cheaper and more durable. A voice terminal today costs about a third of what it did in 2010, and battery life is a full shift. Luna: So the economics finally make sense for mid-sized operators, not just the Amazons of the world. Lucas: Precisely. The Ohio facility is a third-party logistics provider doing about fifty million in revenue. They're not a tech giant. But they saw the math and went for it. Luna: And the results speak for themselves. Ninety percent fewer errors, fifteen percent more picks per hour, five-month payback. That's a case study I'd use in any operations class. Lucas: The thing I find most telling is that they didn't buy a shiny new robot. They invested in a technology that amplifies human capability instead of replacing it. The pickers are still the core of the operation — they just have a better tool. Luna: Yeah. And honestly, if conversations like this have moved your thinking forward in some small way, that's what keeps us going. If today's episode was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. Lucas: And a quick coffee is enough. It really helps us keep this show ad-free and focused on the kinds of operational deep dives we both love. Luna: Exactly. So, back to voice picking — what's next for the Ohio facility? Are they adding more voice zones? Lucas: They are. After the picking success, they're rolling out voice for putaway and cycle counting. The same technology, just different workflows. Early results show cycle count accuracy improving from eighty-five percent to ninety-nine percent. Luna: So the same principle — reduce cognitive load, reduce errors — applies across the warehouse. Lucas: Exactly. It's not about the technology itself. It's about designing work so the human brain can focus on judgment calls, not memorizing location numbers. Luna: That's a lesson that extends way beyond warehousing. Any operation where people have to move and remember — hospitals, parts distribution, even kitchens — could benefit from the same thinking. Lucas: Right. And that's the kind of operational principle worth holding onto.