Last month I flew from Madrid to Mexico City to participate as a speaker in the 2015 logistics summit. When I got to the customs area at the airport in Mexico City I had to wait some 45 minutes to clear customs. Admittedly, after a 12-hour flight, the last thing you want to do at the airport is to waste almost one hour doing nothing but keeping the line. So I decided to try to do something productive while waiting, and so I had a quick look at how operations are organized at the border facilities. Any Operations 101 course tells you to look for the bottleneck of the system first. That was easy, there was plenty of space, desks (around 20), but not enough officers, as you can see in the picture below.
The corresponding scheme would be something like this, where the yellow represents the first passenger in the queue and he green (red) dots represents busy (idle) officers:
This makes sense, as police officers are by far the most precious resources at the border. Of course, there were enough officers to cope with average passenger demand during the day, but given the huge variability of arrivals (in the form of large waves when international flights arrive), passengers in the line is the main buffer against demand variability, which leads to long waiting times.
The next step was to remember two simple rules to manage bottlenecks, namely 1) the quality of input must be guaranteed, and 2) the bottleneck input buffer must be full at all times. There is where I saw two interesting opportunities to improve.
In the case of quality of input, I observed that some passengers had not filled their customs forms, or, even if they had done so, they had not filled all the mandatory fields. Therefore, they had to do so while at the bottleneck, that is, when the officer told them to do so. That was time consuming (what was the code of my flight?) for both, the passenger and, more important, the officer. This is in contrast to customs in other airports, where non skilled workers filter out incorrect forms before passengers get to the bottleneck.
As for the input being full at all times, note in the picture above that the lady next to the officer has no other passenger/s right behind her. That means that when an officer finishes serving a passenger, the first passenger in the line (the girl in the light grey jumper in the picture) has to identify which officer is idle and go there to be “processed”. But passengers are not very agile at doing so, especially after a 12-hour flight. Therefore it was common to see officers waiving hands at passengers or even shouting things like “Next passenger please”. I measured “set-up time”, that is to say, how much time it took for passengers to go from the beginning of the line to the officer desks: 9 seconds on average. I also measured “service time”, how long it took officers to do their job once the passenger was at the desk: 40 seconds on average. The comparison of he two times struck me as surprising: 22% additional time because of set-up time!
That reminded me of another mantra of operations: “reduce set-up time”. How to do that in this case? A simple way to do it is to add a 1-person or a 2-person buffer right after the person being served. This is common in many airport lines, as shown in the picture below (see, e.g., the guy in the yellow jacket).
The corresponding scheme would be something like this, a snake line followed by several small buffers, as many as bottleneck stations:
But it is notable to me is the fact that that simple modification would reduce those 9 seconds of set-up time to only 2 (I guesstimated this figure).
What would be the impact of this set-up reduction in officers productivity? Some simple calculations lead to a surprising 18% increase in productivity, from current 73 to expected 86 passengers per hour and officer.
Combined this with the first change suggested, and you can expect productivity to boost by more than 20% at a very low or even no cost.
This example illustrates how the usual production techniques can be implemented into service environments, where there is a tremendous opportunity for improvement, mainly because managers in service operations have not realized yet about how much money they are leaving on the table. It also gives some evidence of how easily operations in a service environment can be observed and improved. In this case, 45 minutes of casual analysis may lead to significant annual savings for the same waiting time, or, even better (at least for passengers), dramatically reducing waiting unproductive time to cross the border.