Decisions with a cost as well as a result

Our 2025 commercial review identified a cohort of customers considering replacement equipment with a combined proposed value of RMB 34 million. Service engineers believed that several existing installations could continue useful work after a defined refurbishment. The new-equipment team regarded those recommendations as a threat to its annual plan. The disagreement was not about whether the customer should receive advice; it was about which commercial outcome the organisation rewarded once the advice was given.

I required the two options to be assessed separately, with their scope, limitations and financial consequences stated. The service team had to explain what refurbishment could and could not address. The equipment team had to justify the additional capability of a replacement, rather than assume that a newer machine was automatically the better choice. I changed the internal review so a supported repair recommendation did not count as a salesperson's failure to close the larger order.

Customers in the reviewed cohort ultimately commissioned RMB 11 million of refurbishment work through us. We did not book the proposed RMB 34 million replacement orders. These were different scopes, and I did not describe the smaller amount as equivalent revenue or claim an unmeasured environmental benefit. We funded the engineering and parts work required for the repairs and followed the resulting service demand. The immediate company forecast reflected the lower sales value rather than hiding it inside an assumed future customer relationship.

The decision made our service proposition more credible internally: colleagues could recommend work consistent with the evidence even when another option generated a larger invoice. It also exposed the need for better refurbishment costing. I would establish that cost record earlier before extending the approach to a wider installed base. A sound customer recommendation still needs an economically workable delivery model, and the company should not confuse a smaller order with a simpler obligation.

After the minority investment, our 2019 review examined engineering and spare-parts decisions that routinely reached my desk. The median elapsed approval time in the defined cohort was nineteen working days. Twelve managers were capable of deciding much of this work, but colleagues still sought my informal opinion when there was disagreement. A delegation document alone would have left the old route intact, with the formal manager responsible for a choice that could be changed elsewhere without notice.

I worked with the executive team and chair to define the managers' remits, the evidence required and the conditions for escalation. Urgent service matters retained their established response route; the review concerned ordinary engineering and parts approvals. We then recorded requests that bypassed the agreed route, including those I had encouraged by answering too quickly. This was uncomfortable because some colleagues interpreted a referral back to the responsible manager as a withdrawal of access to the founder.

Over the next nine months, median approval time fell from nineteen to six working days in the same category of decisions. Two decisions were subsequently reversed after new information emerged. We reviewed the reasoning and the consequences openly rather than use those reversals to reclaim the entire remit. I also stopped holding separate discussions about a manager's decision without ensuring that the accountable person was involved. The chair challenged me when my shareholder perspective began to intrude into an ordinary operating question.

The business became less dependent on my availability, and managers could see that authority included responsibility for learning rather than an obligation to guess my preference. I would spend more time explaining the changed route to long-serving colleagues at the outset. People who had helped build the company reasonably valued direct access; we needed to distinguish continued personal respect from a decision process that could no longer depend on private conversations.

We recruited forty-two apprentices in 2022 for a two-year practical programme after identifying a growing dependence on a small group of experienced technicians. The first operating proposal treated apprentices as additional capacity relatively early and expected teaching to occur around existing assignments. Workshop leaders questioned that assumption. Demonstration, observation and correction required attention from precisely the people whose production and service time was already most heavily committed.

I approved an RMB 8.4 million programme that included protected teaching time and a lower planned output contribution from the participating experts. Finance asked how quickly the expenditure would become productive headcount. I proposed separate measures for recruitment, assessed task capability, programme completion and later retention. We would not count an apprentice as a substitute for an experienced technician merely because a budget period had ended. The relevant supervisors retained responsibility for the work each person could perform and the supervision still required.

Thirty-one participants completed the assessed stage after twenty-four months. Twenty-seven of those remained with the company after another year. The eleven who did not complete were not all unsuccessful for the same reason: some left, some changed roles and some needed a different development route. We reviewed those differences and retained them in the programme account. Experienced technicians also helped document recurring service situations so supervised practice could build on the company's prior knowledge rather than start from an empty file.

The enterprise gained a more explicit method for developing capability, alongside a realistic understanding of the cost and losses involved. It did not eliminate dependence on experienced people. My main correction would be to prepare the mentors more deliberately before the first intake. Being able to perform a difficult task does not automatically make someone comfortable explaining it, assessing another person's attempt or giving feedback that improves the next supervised assignment.

During 2024 and 2025, we piloted a remote-condition service across sixty-three customer installations. Early reporting emphasised the quantity of data collected and the number of notifications produced. Service managers were more interested in whether a notification justified an investigation and whether the customer could act on it appropriately. Several alerts arose from operating conditions the remote team did not understand well enough. More notifications were not necessarily making the service more useful.

I asked engineering and service colleagues to follow each escalated finding through its practical consequence. The customer retained control of equipment operation, and our remote service did not replace required on-site assessment. We changed the review to distinguish an observation, a finding worth pursuing and the resulting action. Commercial colleagues were concerned that this qualified account would be harder to sell than a broad promise of predictive support. I preferred a proposition we could explain through the work it actually helped a technician and customer perform.

During the reviewed pilot period, twenty-five escalations produced seventeen actionable findings and eight false escalations. We analysed the latter for configuration assumptions, incomplete history and interpretation errors, rather than remove them from the presentation. The development and service-readiness programme cost RMB 2.6 million. Annualised service revenue associated with the pilot cohort was RMB 1.9 million; that figure was not presented as profit or as recovery of the full programme investment. We continued the service with narrower claims and clearer customer responsibilities.

The change gave the company a more useful basis for judging the offering. It also reduced the temptation to interpret a growing stream of data as evidence of a better service. I would involve field technicians earlier in selecting the first installations. Their knowledge of older configurations would have helped us identify cases where a remote pattern needed more context before it could support a sensible recommendation.

In 2009, as general manager of Heshanrow's thermal-equipment division, I faced a decline in new orders while customers continued operating the installed machines we supported. A proposed response removed field-service positions alongside assembly capacity. It produced the largest immediate saving on paper, but it treated service workload as if it would decline in the same proportion as new-equipment demand. Our call history and scheduled work suggested otherwise.

I proposed retaining thirty-eight field engineers while reducing excess assembly overtime, consolidating rented space and revising the sequence of discretionary expenditure. Group finance challenged the smaller near-term saving. I accepted the need to reduce cost but argued that weakening service would create a different obligation later, through delayed work and a less capable response to existing customers. We examined actual service demand by region and equipment type rather than relying on the previous year's new-sales allocation to distribute the field team.

The agreed measures reduced expenditure by RMB 22 million over six months relative to the prior operating plan. We maintained the thirty-eight-person field network and changed scheduling to use it more effectively. The reviewed service cohort's overdue planned visits fell from 64 to 29 over the same period, although the result also reflected changes in customer availability. I did not attribute every improvement to the staffing decision or suggest that the retained capacity had no cost. The division accepted a smaller immediate reduction to preserve a defined capability.

This experience later influenced how I founded Wensharo. Service could not be a residual function funded only after new sales had met the plan. I would improve the original downturn response by separating installed-base obligations from new-project workload sooner. Waiting until a general cost proposal was prepared made the discussion adversarial and forced colleagues to defend a number before we had agreed what work the company still had to perform.