Theory of constraints on the factory floor: what it changes in practice

The focusing sequence and where plants abandon it

The sequence is to identify the limiting resource, decide how to exploit it fully, subordinate everything else to that decision, elevate it only when exploitation is exhausted, and then start again because the constraint will have moved. Plants routinely execute the first step, skip straight to elevation because buying equipment is easier than changing behaviour, and never subordinate at all. Subordination is the hard step: it means other departments accept idle time, accept lower local efficiency figures, and change what they are measured on. Without it the constraint is still starved and still interrupted, and the new equipment simply raises the cost base.

Releasing material to the pace of the constraint

The release rule ties the start of work to the rate the constraint can absorb, rather than to material availability or to keeping the first operation busy. Work is released a defined interval ahead of when the constraint needs it, which caps the amount on the floor and keeps queues short everywhere except the deliberate buffer. The immediate visible effect is that the first operations stop being permanently busy, which alarms supervisors and finance in equal measure. The compensating effect is shorter and much more predictable elapsed time through the plant, which is what customers actually experience.

Buffer management as the daily control routine

Rather than tracking every job, the routine watches the state of a few buffers: the work waiting at the constraint, the work waiting before assembly, and the work waiting to ship. Each buffer is divided into zones, and a job that has penetrated too far towards the front without arriving triggers expediting. This gives the plant a single, unambiguous priority list derived from risk rather than from who complained. It also produces a diagnostic: recording which resource caused each penetration builds an evidence-based list of where the flow actually breaks, which becomes the improvement agenda.

Policy constraints usually bind before physical ones

Frequently the limit on output is not a machine but a rule: a batch size inherited from an old costing method, a shift pattern that leaves the constraint idle at handover, an approval step that holds work overnight, a purchasing rule that orders in quantities the floor cannot use, or an incentive scheme rewarding output per department. These cost nothing to change technically and a great deal politically. Before authorising capital for capacity, it is worth listing the rules that govern how the constraint is run and asking which of them anyone can still justify.

What has to change in reporting for any of this to survive

The approach dies quietly if the monthly pack still rewards absorption, machine utilisation and departmental efficiency, because supervisors will optimise what is reported regardless of what the improvement programme says. Replace local efficiency for non-constraint areas with schedule adherence and buffer status, keep rate measures for the constraint, and report plant-level output and delivery. Finance needs to be part of this conversation early, since the transition period shows lower stock and lower absorbed cost, which looks like deterioration on some conventional reports even while cash and delivery improve.

Frequently asked questions

Does this approach conflict with lean methods?

Less than the debate suggests. Both aim at flow and both attack overproduction; they differ in emphasis, with one concentrating effort at a single limiting resource and the other spreading improvement across the value stream. Plants with a clear dominant constraint and high variety usually get faster results from the constraint focus. Plants with repetitive, balanced flow get more from steady waste reduction. Many operations use constraint logic for scheduling and lean methods for shop-floor improvement without difficulty.

What if we cannot identify a single constraint?

Then either the plant genuinely has spare capacity everywhere and the limit is demand, or the flow is so disrupted that the constraint moves faster than you can observe it. The first case is a commercial problem, not an operations one. The second is usually caused by unstable release and large batches, and it responds to capping the amount of work on the floor first, which stabilises the flow enough for the real constraint to become visible.

How long before the effects show up on delivery performance?

Shorter than most improvement programmes, because the first changes are to release and priority rather than to equipment. Capping released work drains the excess on the floor over a few production cycles, and elapsed time falls as the queues shrink. The lasting result depends on whether subordination holds once the initial attention fades, so build the buffer review into the existing daily meeting rather than running it as a separate project routine.