Ocean freight is a good way to look at your own operating system.

About half of container ships are arriving on time. The late ones are close to a week behind. Congestion is tying up roughly one ship in eight, a four-year high. Rail boxes at Los Angeles and Long Beach sat nearly seven days — the longest dwell in a year. Factories in China are shut for Golden Week. Sailings got pulled. The first ships after the holiday are the ones that will roll cargo.

The easy read is that we need more capacity. That's the wrong read.

A ship sitting in a queue isn't capacity. A container on a ramp isn't stock you can sell. Adding ships into a clogged port just makes the line longer. It doesn't make the customer whole.

I keep seeing the same pattern inside companies. The plan looks fine in a quiet quarter. Then variance shows up, and suddenly the single path is the whole business.

Four moves. Not a new program.

Name the fragility. Find the one port, the one sailing string, the one promised date that carries the quarter. If that path breaks, who finds out last — you, or the customer?

Protect the core. The orders you already promised. The cash that has to clear. The few suppliers who actually show up. Put a wall around those. Everything else can take a punch.

Put one small option next to it. A second origin. A smaller lot that can move another way if the main string rolls. Write the kill rule before you book it: if it hasn't helped the customer by a date you pick, stop it. That's a bet. A second full network you can't afford isn't.

Run the cadence. What did we protect? What did we try? What did we kill? What do we try next? A quarter, not a task force.

Put the cost on the person who made the promise. Storage. A missed install. A customer waiting. If the cost lands on someone else, the booking stays optimistic. Decision rights belong near the work — so do the consequences.

Call the accounts that will hurt if a box slips a week. Offer a date you can keep, or a smaller shipment you can actually move. Customers before extra capacity. A kept promise beats a full ship that arrives late.

Drop the lane that only works in a calm month. Drop the meeting that reports scheduled sailings as if they were loaded ones. That's how the operating system stays honest.

Stress doesn't grade how elegant the plan was. It shows you what still works when the schedule breaks. The firms that come out of this stronger won't be the ones with the most ships on paper. They'll be the ones that named the single point of failure, kept the core promise, and killed one thing the queue taught them to kill.