Latest / The Operations Podcast with Fexingo: Process, People, and Profit in Modern Business / How an Airport Baggage System Lost 30000 Suitcases a Year and Fixed It
Transcript
- Lucas: So here's a number that stopped me cold: one major airport was losing thirty thousand checked bags every single year. Not delayed. Not misrouted to the wrong destination. Lost. As in, the system had no idea where they were. Luna: Thirty thousand? That's roughly eighty-two bags a day, every day. How does that even happen? Lucas: That's exactly the question. And the answer turns out to be a classic operations problem hiding in plain sight. We're talking about baggage handling systems — the miles of conveyor belts, scanners, and sorters that move your suitcase from check-in to the plane. At Heathrow's Terminal 5, which handles about twenty million bags a year, the mishandling rate was tiny in percentage terms. But in absolute numbers? Thirty thousand bags vanished annually. Luna: So what was the root cause? Was it scanners failing, or just human error? Lucas: Both, but the deeper issue was a bottleneck at the main sorting carousel. The system uses a massive rotating loop — picture a horizontal conveyor belt shaped like a racetrack, about three hundred meters long. Bags enter the loop from dozens of induction points, and optical scanners read the barcode to divert each bag to the correct chute. The problem was that the loop could only process about thirty-six hundred bags per hour. During peak departure banks — say between six and eight AM, when thirty flights push out — the system was getting flooded with over five thousand bags per hour. That's fourteen hundred bags more than capacity. Luna: So bags start piling up on the conveyor, and the system has nowhere to put them. That's when they start falling off, or getting stuck in dead zones. Lucas: Exactly. The loop would literally clog. Bags would stop moving, get pushed off by subsequent bags, or end up in what they call shadow zones — spots where the scanners lose line of sight. Once a bag disappears from the tracking system, it's effectively lost. Heathrow's operations team spent months tracing these lost bags. They found that seventy percent of them were stuck in just four specific segments of the loop, all within twenty meters of the main induction point. The bottleneck wasn't the entire system — it was a two-hundred-foot stretch of conveyor. Luna: That's surprisingly localized. So the fix wasn't a total rebuild — it was targeted. Lucas: Right. The team implemented what's essentially a predictive routing system. Instead of letting bags enter the loop at whatever rate the check-in counters spit them out, they added a buffer queue — basically a holding area — that releases bags in staggered batches. The algorithm monitors the loop's real-time density and only admits new bags when there's available capacity downstream. It's like a ramp meter on a highway. You hold cars back to keep traffic flowing, even if it means a short wait on the on-ramp. Luna: And did it work? Lucas: Within six months, mishandled bags dropped by eighty percent. The thirty-thousand lost bags became about six thousand. And the average time from check-in to loading actually decreased by twelve percent, because the system stopped thrashing. The bags that do enter the loop move through without interruption. Luna: So the lesson is that adding more capacity isn't always the answer. Sometimes you just need to control the flow rate. Lucas: That's the textbook lesson from queueing theory. In operations, we talk about the relationship between utilization and wait times. Once a system crosses about eighty percent utilization, queue lengths explode exponentially. Heathrow's loop was running at over one hundred percent during peak hours. The fix was to cap utilization at seventy-five percent by using that buffer queue. You lose a little throughput in the short term — the bags wait an extra ninety seconds in the buffer — but you eliminate the catastrophic failures. It's the same principle Toyota uses with their kanban system, which we've talked about before. Pull-based flow, not push-based. Luna: And the interesting thing is, this isn't just about airports. Any operation that moves physical items through a choke point — a warehouse, a hospital pharmacy, a food processing line — faces the same math. I read a case study last year about a pharmaceutical distribution center that had similar issues with their automated sorting system. They applied the same ramp-metering approach and cut error rates by sixty percent. Lucas: It's a genuinely transferable pattern. But what I find most striking about the Heathrow story is that the root cause was hiding in plain sight. The system had been losing thirty thousand bags a year for nearly a decade. Everyone assumed it was just a volume problem — too many bags, period. But the data showed it was a flow problem, not a capacity problem. The fix cost about two million pounds in software and buffer conveyor — a fraction of what a full system upgrade would have cost. Luna: Which brings up the organizational side. How did they finally get the data to see that? Lucas: They hired a small operations analytics firm that did something simple: they attached RFID tags to ten thousand bags over three months and tracked every single one through the system. That gave them exact positional data — where each bag stopped, for how long, and where it finally went missing. That's how they found the shadow zones. Before that, the only data they had was the aggregate throughput numbers — total bags in, total bags loaded. They never looked at the granular path of individual bags. It's a classic measurement problem: you can't fix what you don't measure at the right resolution. Luna: And that's a lesson for any operations leader. If your KPIs only show big averages, you might miss the real bottleneck. You need time-series data at the individual unit level. Lucas: Exactly. And once they had that data, the solution was almost obvious. They implemented the predictive routing algorithm, which was essentially a few hundred lines of code. It's not rocket science. But it required the willingness to say: the system isn't broken because it's overloaded — it's broken because we're feeding it wrong. Luna: So what's the takeaway for someone running a small operation — say a regional distribution center or a mid-size manufacturer? Lucas: I'd say three things. One: look for the localized bottleneck. It's almost never the whole system. Two: measure at the individual unit level, not just aggregates. Three: consider flow control before capacity expansion. Adding more conveyor, more staff, or more machines often just feeds the bottleneck faster. Instead, regulate the input rate to match the bottleneck's capacity. That's the Heathrow lesson in a nutshell. Luna: And if you want to dig into the queueing theory behind it, there's a great public dataset from Heathrow's operations team — they published the anonymized RFID tracking data after the project. We'll link to it in the show notes. Lucas: We will. And speaking of the show, if these conversations have helped you think about your own operations in a new way — even just once — we'd love it if you considered supporting the podcast. A couple of dollars a month is genuinely what keeps these going. You can find us at buy me a coffee dot com slash fexingo. It makes a real difference, and we're grateful. Luna: Yeah, completely. And it keeps the show ad-free, which is how we both prefer it. Lucas: Alright, back to bags. One final stat that I love: after the fix, Heathrow's Terminal 5 saw a forty percent reduction in late departures attributed to baggage loading. That's not just a customer service win — it's a direct financial benefit. Airlines pay landing fees, and late departures incur penalties and fuel waste. So the baggage system fix paid for itself in under a year just from reduced delay costs. Luna: That's the kind of ROI that gets a CFO's attention. Operational improvements that show up on the bottom line. Lucas: Exactly. And it all started with asking a simple question: where exactly are the bags going? The answer was right there, on a two-hundred-foot stretch of conveyor belt. So next time you're waiting at baggage claim and your bag actually shows up — there's a good chance a little algorithm and a buffer queue made it happen. Luna: And if it doesn't show up — well, maybe they haven't read this episode yet. Lucas: Send them the show notes. I'm Lucas. Luna: I'm Luna. Lucas: And this has been The Operations Podcast.