The sound of a working factory can be reassuring. People are occupied, machines are running and the production report shows activity. I learned to be more curious about what that activity serves. Which orders are moving, and which are waiting again? A sequence can use equipment efficiently while repeatedly postponing requirements that do not fit its preferred batches. If those requirements have already been accepted commercially, the organisation has a problem that a high utilisation figure cannot resolve on its own.
That observation does not make utilisation unimportant. Idle equipment can represent expense, lost opportunity or poor preparation. The question is whether the measure is being asked to explain more than it can. It describes one aspect of how a resource is used. It does not establish that the factory is producing the right combination of work for its commitments. I want the operating discussion to connect those two accounts instead of allowing either commercial urgency or production efficiency to silence the other.
The preferred batch is not the whole order book
Longer runs often have practical advantages. They can reduce repeated setup activity and make the working sequence easier to organise. A production manager who values those advantages is not necessarily resisting the customer. They may be protecting a real operating condition that the commercial promise has not acknowledged. The difficulty begins when the organisation accepts smaller requirements without deciding how they will enter the sequence, then treats their repeated delay as an unexpected failure of execution.
In that situation, the small order becomes an exception that never stops being exceptional. It moves to another date, returns through escalation and is eventually fitted in at an additional cost nobody included when the order was accepted. The customer may see only another unexplained postponement. Production may see an interruption to a plan it was expected to keep stable. Both experiences arise from a promise whose operating provision was never made explicit enough for the people who had to honour it.
Our 2023 trial began with this kind of pattern in one defined area. We did not start by declaring every longer run inefficient or every smaller order commercially indispensable. Planning and commercial colleagues identified a group of requirements that the preferred sequence was repeatedly postponing. The question became whether a deliberate reservation could serve those commitments more credibly than repeated intervention. That was a specific proposal about an identified part of the business, not a universal instruction for all four plants.
We reserved forty-two machine hours per week during a ten-week trial. The arrangement required JPY 24 million in additional setups, planning support and operating expense. Those costs mattered. A leadership account that mentioned only better delivery would have made the decision appear easier than it was. Production managers also anticipated a lower utilisation figure under the existing measure. I supported stating that expectation before the trial, so the people implementing the choice were not later blamed for a consequence management had knowingly accepted.
A different promise needs an operating provision
The reservation was a way to make an accepted commercial requirement visible in the plan. It did not make every requested date feasible, remove technical constraints or authorise work outside established conditions. The relevant production, engineering and quality responsibilities remained in place. An operating provision is useful only when the people using it understand the boundaries. Otherwise the organisation risks replacing one unsupported promise with another, this time under the name of responsiveness.
The commercial discussion also had to become more precise. If smaller requirements are valuable, the institution should understand why and what serving them requires. Their value may concern the customer relationship, the product offer or another legitimate business purpose. It should not be assumed merely because the order has been accepted once. Equally, the factory should not reject the requirement solely because its current measure favours a different sequence. The useful decision brings the purpose and operating consequence together.
During the trial, reported local utilisation moved from 86% in the comparison period to 79%. The selected delivery account moved from seventy-three orders meeting their original promised dates in an earlier hundred-order cohort to eighty-eight in a separate hundred-order cohort. I keep the word separate prominent. These were different populations of equal size, not the same orders observed twice. Their composition and the wider operating conditions could influence the result. The comparison was informative without becoming a controlled causal claim.
The original promised date was also important. A delivery measure can look stronger if the organisation quietly changes the commitment against which performance is assessed. That would not answer the customer question we were examining. We wanted to know whether the selected orders met the date accepted in the defined account, with any changes explained rather than absorbed into a more convenient denominator. A useful measure needs its definition to survive the pressure to present an encouraging result.
Keep the less flattering measure in the result
The final discussion included three facts together: lower local utilisation, additional expense and a stronger selected on-time result. Different colleagues placed different weight on them. Some preferred the old arrangement because it produced more continuous runs. Others valued the clearer provision for smaller customers. My job was not to discover a single measure that made the disagreement disappear. It was to make the trade-off sufficiently clear that management could decide whether this arrangement served the business it wanted to operate.
We retained the approach in the defined area with continuing review. That did not establish the same reservation as appropriate everywhere. Another plant might have a different order mix, equipment constraint or commercial proposition. A trial should improve the quality of the next question, not become a slogan detached from the conditions that made it useful. I would expect the decision to be reconsidered if its cost, demand pattern or technical context changed materially.
The experience also changed where I would begin next time. The useful moment is before the commercial promise becomes routine. Which kinds of order does the proposed operating model serve well? Where will a less convenient requirement enter? What expense or lead-time condition should be understood before it is offered? Asking these questions earlier can avoid years of treating a structural mismatch as a series of exceptional scheduling failures.
I still value a well-run machine and a disciplined production sequence. I simply want them connected to the institution’s actual commitments. A factory earns confidence when it can explain the mix it is designed to serve and the consequences of changing that mix. Sometimes the right choice will preserve the long run; sometimes it will make room for a smaller requirement at a visible cost. The leadership responsibility is to make that choice before the customer becomes the person who reveals that a busy factory has been keeping a different promise.