The H2 Wiki


Why Flexport’s CEO is wrong about the bottleneck at California’s ports

– by Tom Ellis – 25th October 2021

I recently read Eliyahu Goldratt’s The Goal, a “business novel” about how to increase capacity in a manufacturing plant, whose lessons can be applied more widely, to logistics and beyond. Over the last few days, themes from the book have been playing out in public, particularly on Twitter, as a consequence of the gridlock at California’s Los Angeles and Long Beach ports. The CEO (Ryan Petersen) of Flexport (a freight forwarding and customs brokerage startup) has tweeted voluminously about the logistical “bottlenecks” he has discovered on a couple of intelligence-gathering expeditions, firstly to talk to dock workers and secondly to observe dock operations from the water.

“Bottlenecks” (or “constraints”) are exactly what The Goal is about. Given that I’ve read a book and I have critical thinking capacities I deem myself suitably qualified to critique the CEO of a billion-dollar logistics company. Let’s go!

[Disclaimer: I don’t know anything about operations at Flexport. I don’t know whether Petersen has taken the right or wrong actions or decisions regarding his business and California’s ports. On the other hand I am convinced the way that he explained the situation on Twitter doesn’t make sense. That’s what this article is about.]

In the first thread (about talking to the dock workers) Petersen explains that, although the docks have started operating 24/7, trucks are not arriving 24/7 to take away offloaded containers. In the second thread (describing how he observed dock operations from the water) he claims that

everyone now agrees that the bottleneck is yard space at the container terminals

What’s his rationale for the claim that the bottleneck is yard space? Well, it seems that there are currently a lot more imports from China than exports to China so the trucks that took full containers from the port are bringing back empty ones. Apparently you can’t get empty containers off your dockyard by loading them onto ships because shipping companies are not being paid enough to take empty containers or because it’s dangerous to carry empty containers. Stacks of empty containers are building up at ports leaving too little space for ships to be unloaded. The problem compounds when truckers bringing empty containers back to the port find no space to leave them, so they can’t exchange their container for a full one, so full containers stay at the port. An overload of containers is jamming up port operations, empty containers cannot be removed by ship nor can full containers be removed by truck!

Petersen concludes (and in a moment I’ll explain why I think he is incorrect) that the bottleneck is yard space. His proposed solution is for local and national government to act to allow trucking companies to leave empty containers elsewhere, on large empty sites around the locality (one imagines disused airfields etc.) leaving space in the dockyard for unloading containers and space on trucks to take full containers from the port.

In his analysis Petersen explicitly recommends The Goal as “the most important business book ever written”. I wouldn’t dare disagree with him. But I think he has applied the lessons of The Goal incorrectly. Let’s look at some of his Twitter thread.

When you’re designing an operation you must choose your bottleneck. If the bottleneck appears somewhere that you didn’t choose it, you aren’t running an operation. It’s running you. [Source]

Sounds edgy, thanks Tyler Durden.

You should always choose the most capital intensive part of the line to be your bottleneck. In a port that’s the ship to shore cranes. The cranes should never be unable to run because they’re waiting for another part of the operation to catch up. [Source]

This sounds strange to me. You rarely get to choose your bottleneck. The bottleneck is an emergent phenomenon. Perhaps some highly skilled process engineers can design a bottleneck, but generally a bottleneck is something you identify and manage. Furthermore, however much skill you put into design you only ever design the part of the system you control. The bottleneck may come from another part of the system entirely, one you have limited control over or insight into, such as a partner or the market itself.

Another way of putting Petersen’s statement “you should always choose the most capital intensive part of the line to be your bottleneck” is “you should always choose the bottleneck to be the most expensive part to run”. But this confuses cause and effect! You should indeed only make large capital outlays if they result in improving the flow through the bottleneck; any other outlay is wasted. The factory that features in The Goal has outlaid enormous amounts of capital on robots, but they are not the bottleneck and no amount of trying to reconfigure the system will make them the bottleneck. The capital outlay on the robots was simply wasted.

Indeed in an efficient operation the bottleneck should always be the most capital intensive component, but that’s because you chose to invest capital in the bottleneck, not because you chose to bottleneck your capital.

The bottleneck right now is not the cranes. It’s yard space at the container terminals. And it’s empty chassis to come clear those containers out. [Source]

I don’t agree with either claimed bottleneck. I’ll tell you what I think the bottleneck is presently.

In operations when a bottleneck appears somewhere that you didn’t design for it to appear, you must OVERWHELM THE BOTTLENECK! [Source]

“Overwhelm the bottleneck”? “Overwhelm” is not a good word. What does he mean? “Overwhelm” does not occur in Goldratt’s Five Focusing Steps (which use the term “constraint” rather than “bottleneck”):

  1. Identify the constraint

    There is normally only one “constraint” (or “bottleneck”) in a system. It is the unique part of your process which constrains the useful operation of every other component in the system.

  2. Exploit the constraint

    Make sure you are getting full capacity from the constraint. For example, if it’s a machine, make sure it has no more idle time than absolutely necessary.

  3. Subordinate everything to the constraint

    There’s no point expending effort on increasing the capacity of other components except to the extent that their operation ensures the constraint is kept fully exploited. “Subordinating the constraint” means making sure every other part of the system works neither less nor more than is required to achieve that end.

  4. Elevate the constraint

    “Elevating the constraint” means increasing the capacity of the constraint, which, given that everything is already subordinated to the constraint, will immediately increase the capacity of the overall system.

  5. If elevating the constraint makes something else the constraint, go back to step 1, but prevent inertia from becoming the constraint.

These five focusing steps are the basis of Goldratt’s Theory of Constraints. I have never come across a principle that states “overwhelm the bottleneck”. In fact the only Theory of Constraints related references that I found to to “overwhelm the constraint” or “overwhelm the bottleneck” suggest that “overwhelming the bottleneck” is a negative thing. “Overwhelming the bottleneck” is what happens when you don’t subordinate everything to the bottleneck and the bottleneck has difficulty running at full capacity because of that.

But perhaps I’m being too harsh. Perhaps it’s a simple slip of the tweet and by “overwhelm the bottleneck” Petersen meant “exploit the bottleneck”. But failure already occurred at step 1: identify the constraint. So why do I disagree that yard space and empty trucks are the bottleneck? Because although finding more yard space for containers will allow more containers to be taken off ships – and therefore more goods to come into the country – it will not increase the steady state capacity of the system.

In the absence of ships taking containers away from the US there will be an increase in US “container inventory”. The Goal goes to great pains to explain why, although you always need small buffers to minimise disruptive shocks, especially in front of the constraint, inventory should be reduced not increased! The most obvious reason to reduce inventory in this case is that empty containers should be sent where they are most needed: China, where they are filled with goods destined for the US. Empty containers should be piling up where they are needed (China) not where they are redundant (the US). In response to Petersen’s claim that the nation needs more space for empty containers his followers contacted their elected representatives and subsequently more space was found for empty containers. That space will fill up without solving the root problem. As soon as there are too few empty containers in China the problem with manifest again.

The actual constraint

So what do I think the actual constraint is and what should be done about it? Let’s go through the Five Focusing Steps:

  1. Identify the constraint: It seems to me that the constraint is the shipping of empty containers to where they are actually needed, China and other origins of US imports.

  2. Exploit the constraint: make sure that no ship leaves an overloaded port without as many containers as it can carry. If that requires increasing the fee paid to shipping companies then so be it. If transporting empty containers is dangerous then find a way around it. Maybe you can fill empty containers with scrap metal or sand.

  3. Subordinate everything to the constraint: only allow one container off a ship for each container that you have just put on a ship. Only allow a truck to bring a container to a port if there is a ship waiting to take it.

  4. Elevate the constraint: perhaps unladen ships that would not have otherwise been sent to US ports can be diverted there to pick up empty containers and take them to China.

  5. At this point, if the constraint has been broken then it’s back to business-as-usual for the ports, Flexport and international commerce. There are surely some new constraints but I’m not qualified to say what they are. I haven’t read the right book.

How to burden the constraint even more

The Long Beach port chief has admonished consumers to “shop early” for the holiday season. Spiking demand exactly when the system seems unable to cope seems like the opposite of the best thing to do. On the other hand perhaps he knows exactly how sticky consumer behaviour is and that customers won’t actually start buying for a few weeks, by which time he’s confident the slowdown will have been resolved. Perhaps.


I don’t actually know the right way for Flexport to solve the difficulties it faces; I don’t know enough about the details of its business. But I do know that, despite recommending The Goal as “the most important business book ever written”, what Petersen publicly wrote on Twitter does not align with the lessons of that book.

Addendum on marketing

Intriguingly, each of Petersen’s tweets starts with the same structure:

I’m not sure if this is standard marketing copy or something Flexport devised themselves.