A busy factory can still be making the wrong promise.

In 2023, our planning review showed a recurring problem with smaller component orders. They were accepted commercially, but the factory sequence repeatedly moved them behind longer runs. The local utilisation report looked strong. The customer account was less satisfactory. I asked planning and commercial colleagues to identify the requirements that the preferred sequence was systematically postponing, rather than treat each late order as a separate exception requiring another urgent intervention.

We reserved forty-two machine hours per week for that work during a ten-week trial in a defined production area. The arrangement cost JPY 24 million in additional setups, planning support and operating expense. Production managers objected that it would lower the utilisation measure on which the area had been judged. They were right: the reported figure moved from 86% in the comparison period to 79% during the trial. I supported making that reduction explicit rather than redesigning the measure to hide the consequence of my decision.

The selected customer measure improved. Seventy-three of one hundred orders in the earlier comparison cohort had met their original promised dates; eighty-eight of a separate hundred-order cohort did so in the trial account. These were different order populations of equal size, not the same orders measured twice or a controlled experiment. We kept the product-area boundary and the original dates visible. The result did not prove that reserved hours alone caused every improvement, and it did not establish a company-wide service outcome.

The decision gave management a concrete trade-off to examine: a lower local utilisation figure and higher expense alongside a stronger selected delivery result. Some colleagues still preferred the former arrangement because it produced more continuous runs. Commercial colleagues valued the clearer provision for commitments already made to smaller customers. We retained the approach for that area with continuing review, rather than impose the same reservation across every plant. I would now ask how the proposed order mix fits the production measure before the commercial promise is accepted. A factory should not have to discover through repeated escalation that its definition of efficiency excludes part of the business it has agreed to serve.

In 2015, a twelve-person operating group depended on four people who held the required authorisation for a particular setup responsibility. The production plan assumed more flexibility than that distribution supported. When one of the four was unavailable, supervisors had to rearrange work or seek additional support. I initially heard the problem described as a scheduling inconvenience. The more useful account was that the organisation had not funded the learning conditions needed to broaden a specific capability.

We created a programme for the other eight operators, including eighty paid training hours per participant, supervised practice and assessment by the authorised technical and supervisory owners. The total programme expense was JPY 16 million, including teaching, cover and preparation. Some plant managers wanted the programme completed around normal work without taking that time out of the schedule. I disagreed because the factory could not credibly ask people to acquire a new responsibility while pretending that the learning consumed no capacity.

Five of the eight participants subsequently met the relevant requirements and received authorisation. Three continued supervised learning. The group therefore had nine authorised people for that defined responsibility, not twelve. Attendance, paid hours and assessment were reported separately. I did not override a technical judgement to obtain a cleaner completion percentage, and the programme did not promise a particular promotion or employment outcome. Its purpose was a specific operating capability, with the standard and the decision retained by the appropriate owners.

The wider lesson concerned the timing of the operating promise. A manager should not count future flexibility as available merely because training has been booked. The factory still needed a workable plan while the capability developed, and that plan needed to recognise the three participants who required longer. The additional expense was a real choice, not an accounting inconvenience to be recovered through an unsupported productivity claim. I would involve supervisors in the capability account earlier, asking which responsibilities are actually covered and what conditions support the next person’s development. The result was more useful than an attendance total because the receiving manager could understand precisely what work the group was now authorised to perform.

A 2021 proposal sought JPY 280 million for an automated cell intended to improve a troublesome part of the production sequence. Its commercial attraction was clear: fewer interruptions and a more repeatable operating arrangement. The initial discussion nevertheless combined several questions. Some delay came from how fixtures and tools were prepared. Other requirements concerned future volume and the broader process. I asked the engineering team whether a narrower intervention could establish useful evidence about the first problem before we committed to the whole proposal.

We authorised JPY 46 million for fixtures, tool-change preparation and the supporting work required for a defined product family. The technical owners assessed the proposed changes, with independent acceptance and safety responsibilities preserved. The automation sponsor objected that an intermediate programme could create expense that a later cell would partly supersede. That was a legitimate concern. I supported proceeding because the existing proposal treated unresolved preparation problems as though a larger capital commitment had already explained their cause.

Across twelve observed changeovers before the intervention and twelve afterward, median recorded time moved from eighty-three to fifty-one minutes. The observations concerned one product family under the stated working arrangement. They were not proof of a plant-wide capacity increase, a quality improvement or the elimination of every reason to automate. The JPY 234 million difference between the original proposal and the authorised programme remained uncommitted capital, not a realised saving. We also retained the possibility that later demand or process evidence would justify a different decision.

The programme made the next investment conversation more specific. Management could distinguish the preparation issue addressed from the future requirement still under examination. It also made the cost of learning visible, including the risk that some work would not be reusable in a subsequent configuration. I would ask for that distinction earlier in a proposal now. A staged decision earns its place by answering a material question, not simply by costing less. My responsibility was to ensure that the smaller programme produced evidence the larger choice needed, while keeping enough discipline to return to the capital question rather than allowing a promising local result to close it prematurely.

As a plant manager in 2010, I supported reducing twelve supplier relationships in a defined component-support group to six. The proposal promised a simpler purchasing account and fewer coordination points. It looked reasonable from the consolidated spend view. I paid insufficient attention to two specialist finishing requirements whose practical support did not transfer cleanly into the retained arrangements. The smaller list became easier to administer before the factory had established that it was equally able to serve the required work.

Supervisors and engineering colleagues raised the limitation when the new arrangement was used. My first response was to ask whether coordination could improve within the six relationships. That was worth examining, but it also delayed acknowledging the capability gap created by the original decision. We reopened two specialist relationships, bringing the defined group to eight, after the relevant qualification and commercial work. JPY 9 million covered qualification activity, coordination and transition expense. Technical acceptance remained with the authorised owners; restoring a supplier’s name to a list did not establish that its output could automatically be used.

Purchasing colleagues were concerned that the reversal would make the original simplification look unsuccessful. I shared responsibility for that problem because I had endorsed its headline. The useful response was to explain which two requirements had not been adequately covered and why the correction was narrower than a return to all twelve relationships. We retained the six arrangements that could support their intended work and restored two capabilities the operating account still needed. The total of eight described the revised group, not a claim that every supplier decision elsewhere in the business required reversal.

I did not report the original reduction as a continuing efficiency benefit without attaching the correction. The JPY 9 million was additional expense, and the episode changed how I reviewed later consolidation proposals. Administrative simplicity is valuable, but it must be tested against the different purposes hidden inside the population being simplified. I would now ask a receiving production or engineering owner to explain how each specialist requirement will be served before removing its existing arrangement. That would have brought the operating question into the decision while there was still time to address it without reopening a commitment I had already encouraged others to make.

By 2024, a group of twenty-seven component variants had accumulated the same implied promise of routine availability even though their demand and production requirements differed. Some were ordered regularly. Others interrupted the schedule infrequently but required distinct preparation when they returned. Commercial colleagues valued the breadth of the offer, while plant teams increasingly treated the less frequent work as a disruption. Neither description was sufficient. We needed to decide what kind of promise each part of the range should carry.

The joint review retained nine variants as standard, moved twelve to made-to-order terms and removed six from new quotations. The changes were discussed with the relevant commercial, technical and customer owners. Existing contracts, service commitments and appropriate support arrangements were examined separately; retiring a variant from a new offer did not cancel an obligation already accepted. We recorded JPY 37 million in transition expense for stock treatment, tooling arrangements and the work of changing the offer. It was not presented as a free simplification of the factory plan.

Sales colleagues challenged the move to made-to-order terms because an explicit lead-time discussion could make a quotation less attractive. Production colleagues initially preferred a larger reduction in the standard range. I supported the three-part arrangement because the choice was not simply between keeping every variant unchanged and refusing the work. Twelve could remain commercially relevant with different operating terms. Nine justified the routine provision, and six no longer belonged in new quotations after the relevant review. Those categories needed to be understood by the people communicating with customers, not only recorded in a planning file.

The outcome was a more explicit relationship between the range and the factory commitment. It was not the elimination of twenty-seven variants, a claim that no customer was inconvenienced or proof of a permanent cost reduction. The transition required difficult conversations and continued attention to existing obligations. I would begin the discussion earlier when a proposed variant is first considered. An offer can be commercially attractive and still need a distinct production promise. Keeping that distinction visible allows the organisation to make a considered choice about variety instead of gradually creating an operating obligation that nobody remembers deciding to accept.