When the Plant Ran Hot: A Capacity Lesson from a Network Under Strain
By the spring of 2021, a mid-sized contract manufacturer we will call Northriver Components had survived the worst of the early pandemic shock and was, on paper, doing well. Orders had rebounded faster than anyone forecast. The problem was that recovery did not arrive evenly. Demand surged at the same moment that two of its tier-two suppliers were still running short-staffed, freight lead times had roughly doubled, and a single moulding line in its own plant was quietly becoming the constraint that decided everything downstream.
The story below is a composite drawn from several real situations, anonymized. It is worth telling because the mistakes were not exotic. They were the ordinary kind that good capacity planning is meant to catch.
Where the plan broke down
Northriver planned capacity the way many firms still do: by looking at its own four walls. The plant had nominal capacity figures for each line, and as long as forecast volume stayed under the line totals, planners assumed they were safe. What that view missed was that capacity is a property of the whole network, not of any single node. The binding constraint was never the plant average — it was the one moulding cell that fed three product families, plus a resin supplier whose own recovery lagged by a quarter.
Capacity was measured in averages, hiding the bottleneck cell that actually governed throughput.
Supplier capacity was assumed, not confirmed; no one had asked tier-two vendors what they could realistically commit to.
The plan held a single demand number rather than a range, so there was no slack reserved for the upside that actually materialized.
Lead-time growth was treated as temporary noise instead of a structural change to be planned around.
What disciplined network capacity planning looks like
Capacity planning across a network is less about a single grand model and more about a handful of honest questions asked at every link in the chain. Done well, it connects forecast demand to the real constraints that turn that demand into shipped product.
Find the true constraint. Map the flow from raw material to finished good and identify the resource with the least slack relative to demand. That bottleneck — a cell, a supplier, a port — sets the pace for the whole network, so plan to it, not to the plant average.
Plan to a demand range, not a point. Carry a low, expected, and high scenario. Reserving capacity for the high case is cheaper than scrambling for it, and it tells you in advance which suppliers you would need to lean on if the upside arrives.
Confirm supplier capacity in writing. Treat tier-two and tier-three commitments as facts to be verified, not assumed. Ask what volume a supplier can hold, by when, and what their own constraints are.
Build lead-time changes into the plan. When transit and replenishment times shift, re-time your orders and safety stock to match. A plan that still assumes pre-disruption lead times is quietly planning to be late.
Re-plan on a rhythm. Capacity is not set once a year. A monthly sales-and-operations planning cadence keeps demand, supply, and constraints aligned as conditions move.
Had Northriver done these things, the surge would still have been hard — but it would have been a managed problem rather than a surprise. They would have known months earlier that the moulding cell, not the plant, was the ceiling, and they would have had a confirmed second source for resin instead of a hopeful one. The recovery taught a durable lesson: a network is only as fast as its tightest link, and you cannot manage a link you have never measured.
Sorting out where your real capacity constraints sit, and locking down the supplier commitments behind them, is work we do with clients every day through XNM's procurement, sourcing & contract management.