When the Warehouse Was Full but the Shelves Were Empty: A Pull-System Turnaround
An equipment assembler had a paradox on its hands. The warehouse was full — aisles of components stacked to the ceiling, working capital tied up everywhere you looked — and yet the line still stopped for shortages of the exact parts it needed that week. Managers reacted the way most do: they ordered more of everything. The warehouse got fuller, the shortages kept happening, and the cash kept disappearing into stock that sat untouched for months.
It was early 2021, and the timing turned a chronic problem into an acute one. Supply was unpredictable, lead times were stretching, and the instinct everywhere was to hoard. But hoarding without a system is just expensive guessing. What this team had was a classic push problem: parts were ordered and made to a forecast and shoved downstream, whether the line was ready for them or not.
Push versus pull, plainly
In a push system, work is produced to a schedule or forecast and pushed to the next stage regardless of actual demand. In a pull system, a downstream step signals when it is ready, and only then does the upstream step replenish what was consumed. Pull is a core idea of Lean: you let real consumption, not a prediction, trigger the next batch. Done well, it lowers inventory and shortages at the same time — which sounds impossible until you see why the two problems share a root cause.
Push optimizes each stage in isolation and lets inventory pile up between them.
Pull links the stages, so nothing is made until the next step actually needs it.
The bridge between them is a simple signal — historically a kanban card — that says "we used some; send more."
How the turnaround happened
The team did not rip everything out overnight. They started small, with the parts that hurt most, and let the system prove itself before expanding.
Separated the vital few from the trivial many. A quick analysis showed a small set of parts drove most of the stoppages. Those got attention first; the rest could wait.
Set replenishment signals. For each chosen part, they defined a clear trigger point — when the bin dropped to a set level, that was the signal to reorder a fixed quantity, no meeting required.
Sized the buffers for real variability. Because supply was shaky, they sized buffers to cover realistic lead-time swings, not best-case timing — pull does not mean zero stock, it means the right stock.
Made consumption visible. A simple board showed what had been pulled and what was inbound, so everyone saw the same picture and stopped panic-ordering.
Expanded only after it held. Once the first parts ran smoothly for a few cycles, they brought more items onto the same system.
The result, and the lesson
Within a couple of months total inventory fell while line stoppages dropped sharply — the cash freed from overstocked items funded the right buffers on the parts that actually held up production. The shift was not about buying less in the abstract; it was about buying in response to real consumption instead of fear. The team stopped reacting to every rumour of disruption and started letting the line tell them what it needed.
The broader lesson holds well beyond a warehouse. Whenever you are producing ahead of real demand — reports nobody reads, features nobody asked for, stock nobody pulls — you are running a push system, and it will quietly cost you. Pull is the discipline of letting genuine need set the pace.
If your operation is drowning in what it doesn't need and short of what it does, XNM's strategic advisory can help you move from push to pull where it counts.