After years of disruption, companies swore they would build resilience. The latest data suggests the old habits came right back.
A few years ago, the global supply chain delivered a lesson nobody could ignore. Shelves emptied, factories idled for want of a single part, and the elegant just-in-time machine that had quietly run the modern economy revealed how little slack it had. Executives gave speeches about resilience. Consultants sold thick reports about building redundancy. Everyone agreed, loudly, that things had to change. The latest data suggests that almost as soon as the pressure eased, the old habits came right back.
The diagnosis was not complicated. Decades of optimizing for cost had stripped the system of any margin for error. Companies sourced critical parts from a single supplier because it was cheapest, held minimal inventory because storage costs money, and trusted that a smooth world would keep the parts flowing. When the world stopped being smooth, there was nothing to absorb the shock.
Resilience is something everyone wants to have bought and no one wants to keep paying for.
— a logistics executive, speaking candidly
The promised fixes were sensible and widely announced. The trouble, visible only now, is that nearly all of them cost money in the present to prevent a problem in the future — exactly the kind of trade-off that survives a crisis and dies in a budget meeting once the crisis fades.
For a while, genuinely, they did it. Inventories rose, suppliers were diversified, and the language of resilience showed up in earnings calls. Then the disruptions faded, the cost pressure returned, and the metrics quietly drifted back toward where they had started. Inventories have thinned again. The expensive second suppliers, used rarely, have been pruned. The reshoring that made for good press has, in many cases, stalled once the spreadsheet caught up with the speeches.
This is not because anyone forgot what happened. It is because the system rewards the cost saving you can see this quarter over the disaster you might avoid in some future one. Resilience is insurance, and insurance always feels overpriced right up until the moment you need it.
The uncomfortable conclusion is that this is not a story of bad memory but of bad incentives, and incentives do not change because of a hard year. As long as the manager who trims the buffer gets rewarded now and the cost of that trim only appears during a shock that may arrive on someone else's watch, the slack will keep being optimized away. The lesson was learned. It was simply not the kind of lesson the system is built to keep.
None of which means nothing changed. Some firms, especially those burned worst, have kept their buffers and meant it. But the broad retreat to old habits suggests that the next disruption will find a system not so different from the one the last one caught off guard — and that the speeches, when they come again, will sound familiar.
Companies promised supply-chain resilience after the last shock, then quietly let buffers, backup suppliers, and reshoring slide once cost pressure returned. The problem is not memory but incentives that reward visible savings over invisible insurance.