'Why We Have Lead-Times, as Explained by the TSA'
As a consultant in supply chain management, I travel a lot. One of the joys of modern travel is going through airport security. As you may have noticed, getting through security is not always the fastest thing to do. As I wait in those lines, wondering if I am going to miss my flight, I can’t help but fantasize about… material lead-time.
Yes, I can’t shake my operational roots. And really, equating the Transportation Security Administration line with lead-times is a good analogy. All processing networks can be modeled with similar mathematics and they all demonstrate similar dynamics. The line at the security check point really is just like the lead-time for a manufacturing order, or the wait time on hold with a call center, or the packet buffer in an internet router. When I am in a line, I just imagine that we travelers (pun intended) are all work orders in the factory waiting to be processed. Nice little batches of work-in-process on the shop floor, watching our fellow production jobs go, first-in, first-out, into the manufacturing line that is the X-ray machine.
But these thoughts of lead-times and production jobs are not always as happy and comforting as they sound (what, the manufacturing floor is not comforting to you?). As I look at the lines, I think the same thing I think when I see lead-times: Why are they so damn long?
Manufacturing and supply chain professionals have been conditioned (by years of quarterly beatings from our superiors) that lead-times and the associated inventory are bad, bad, bad. Entire philosophies of manufacturing, such as Lean, Six Sigma, the Toyota Production System and Just-In-Time are based largely on the idea that we should reduce inventory and lead-times. Some days, both on the job as an operations professional, and waiting in the TSA line, it feels like all this good knowledge on reducing lead-time isn’t being put into practice.
Here are a few concepts about lead-time dynamics that might help the TSA. For my readers who are manufacturing and supply chain professionals, every time you see “line” or “wait time” below, just think of your lead-time or buffer stock, and you can apply the same learning. Of course, I have to add the huge caveat that I am only addressing the queuing dynamics of the security lines. The TSA personnel are professionals whose primary function is security and protecting us from those who would do harm. In that regard, they have only my appreciation and respect. But about those lines…
Limited Capacity and Variability
Why are there lines at all? Lines occur because of two things: a limited capacity and variability in the process. By “limited capacity” I don’t mean that capacity is less than is needed on average. Eventually, over the course of the day, everyone gets through the security line. So if you take the average over the whole day, there is enough capacity to deal with the number of people that try to get through security. Limited capacity means that there is less capacity than any short term “rush” that occurs.
This is where variability comes in. If everyone who showed up to the security line arrived at exactly the same pace as someone was exiting the security process, then there would be no line and no wait. But there is a great deal of variability in arrival times; sometimes a whole plane-load or more of people will show up almost at the same time, and other times, almost no one gets into the line. There is also variability in process times, or how long the security process takes for some people to complete versus others. I always seem to be in line behind the person who needs 7 bins for all their miscellaneous jackets, shoes, bags, computers, etc.
Having a line buffers that variability and allows utilization of the capacity to be higher. The line is the “backlog” of work that waits, ready to fill the next opening, so that the x-ray machine or the body scanner spends less time idle. This is why so much focus is put on reducing variability in manufacturing and supply chain philosophies: If you can reduce variability, then you can reduce your lead-times or increase your capacity utilization. (To be clear, I am calling utilization the amount of time a resource is in use divided by the amount of time it is available.)
At the security line, reducing variability in arrival times is hard. Flights schedules aren’t going to get adjusted just to smooth out the security lines; and anyway, people make individual choices about how early or late to arrive. Reducing variability in process times, however, can be done. One big source of process time variation in the security line is when a passenger makes a mistake: they forget to take liquids out of their bag, for example. When this happens, the bag involved has to have a complete “do-over” and go back through the x-ray, causing that process to take at least twice as long as it should.
Through years of careful study of human nature, I have come to a startling conclusion: people don’t like to be yelled at. Yet, for some reason, the TSA approach to preventing passenger mistakes is to proactively yell at them while they are attempting to unload their belongings into bins. Shouting “please remove all liquids and gels” at me while I am busy attempting to remove my liquids and gels is not only insulting, I seriously doubt whether it is successful at reducing errors.
I would propose that the TSA person whose job description currently states “stand at the entrance to the body scanner and shout at people in a condescending manner” be re-trained to walk up to people as they get their bins and have an actual normal-volume conversation with each passenger. The TSA officer could go through a quick check list of do’s and don’ts and then each passenger has a chance to ask questions, too. Actually helping people do the right thing is usually a better approach than telling them that they are doing it wrong. In my experience, there is plenty of time for a one-on-one exchange with each passenger between the time the bins are picked up and the time you actually put your stuff into the x-ray scanner.
Adding increased capacity – and living with lower capacity utilization – can also reduce the average line length. There are airports in Asia where the security screen happens right at the gate. You can walk through the terminal without going through security, but when you want to enter your specific gate, there are security personnel, a metal detector and an x-ray machine for the bags. These wait times are short because there is a dedicated security line for that flight. The trade-off is that the airport had to invest in lots more equipment.
If you want to be especially frustrated the next time you are waiting in the line at the airport, remember this fact: the airport considers my time (and the time of everyone waiting with me) to be less important than what it costs to add capacity. Can you imagine if your local grocery store told you, “Please make sure you arrive an hour early for your shopping to allow for enough time to get through check out.” We’d find someplace else to shop. And yet the airline industry finds it acceptable to make us wait rather than provide additional capacity for check in and security.
I’ll sometimes hear the argument that there isn’t enough space for more security lines and equipment. I find it funny that they can’t find the space for equipment, but they can find the space for a huge line of people stretching hundreds of feet. I’m not buying that story.
The length of the line is related to capacity utilization in a non-linear fashion. Without going into the math behind it, the graph below shows the relationship between wait times and utilization in a generalized system. The different curves represent different levels of variability.
There are a few key take-aways from these curves:
There are times when there are a lot of people arriving in the line and times when there are none – that’s the big source of variability. Because the process is not utilized when no one is arriving, it is never possible to run close to 100% utilization, on average, without driving wait times up asymptomatically. In other words, for utilization to stay up near 100% all the time, you would have to have a long enough line to keep the process busy even during the slow arrival times. That would mean making people wait a long time.
I don’t know how the TSA looks at capacity, but if they are calculating the average number of travelers per hour or day, and then providing just enough security capacity for that average rate, then that is not sufficient capacity. Because of variability, the planned capacity utilization has to be well under 100% to avoid huge lines.
A Little Bit of Additional Capacity Goes a Long Way
What’s more, there is a point as utilization increases where that wait time shoots up rapidly. The lesson to remember when we get on the steep part of that curve and wait times start to get long is that a little bit of additional capacity goes a long way to reduce wait times. On the upper part of the curve, moving the utilization number down just a few points by adding a small amount of capacity results in a big change in average wait times. In airports with long security lines, the fix may be a lot cheaper and easier than it appears.
A great example of this can be found in retail stores. Often, when lines start to get long, they will call another person up to man a checkout line temporarily; adding this small amount of once-in-a-while capacity helps keep lines down.
Manage the bottleneck. The security check point is not just one process, but a series of processes. First is the check of the ID and boarding pass. Then you load up your bins. Then the bags go through X-ray, while you go through the scanner. Finally, you pick up your stuff and get dressed again. In any multi-step process, there will be one step that moves a little slower than the others – the bottleneck.
In my observation, the X-ray is typically the bottleneck in the security checkpoint. The X-ray gates the capacity of the whole series of processes. The capacity of the whole system can be increased by increasing the capacity at the bottleneck, and similarly, capacity lost at the bottleneck is lost in the whole system. Therefore, managing the X-ray to get the most output is important to keeping lines shorter.
Getting the most output at the bottleneck requires that the other process steps don’t slow it down. The only time I’ve gotten mad enough to complain on the spot to a TSA supervisor was in my hometown of Austin. I had waited in a long line only to find that three X-ray lines were available with no wait while only one person was checking ID’s. They had artificially created a bottleneck. Simply shifting one person to check ID’s would have dramatically increased the overall capacity of the checkpoint and therefore reduced my wait time in line.
When the bottleneck – the X-ray – becomes the priority, it is easy to observe simple improvements that could keep that resource operating with the best utilization. For example, the TSA should always have bins available, and provide ample space to load and unload bins. I’ve seen exceptionally small bin loading areas that don’t give people enough space and time and therefore force the X-ray machine to wait. Another example is the “bag check” requests made by the X-ray operator. Personnel should be available immediately so the X-ray scanner doesn’t sit idle waiting for someone to look at the suspect bag.
The next time you go through security, I bet you will look for improvements. You’ll see things that could reduce variability, increase capacity, or better manage the bottleneck. I hope you send them to TSA.
I really hope someone from TSA reads this article. I’d like to think that I’ve done my small part in improving the process. If not, I hope my readers in the manufacturing and supply chain world have learned (or at least been reminded of) a few good tricks.
Good luck getting through your next airport security line.
Jeff Wallingford is vice president, Supply Chain Strategy, for Riverwood Solutions.
