Dinesh Navrayan

In 2021 a customer requested the transfer of a tool set at the end of a programme. The factory knew where the equipment was, but the commercial file did not clearly reconcile the original customer contribution with subsequent modifications. We paused the administrative transfer while the parties established the agreed position. The incident prompted a review of 145 customer-funded tool sets across Selvarune. Twenty-one had incomplete records of ownership, modification approval or return obligations.

Our commercial director argued that a broad review would reopen relationships that were functioning well. He wanted to deal with exceptions only when customers asked. The engineering director was concerned that moving equipment without a clear account of its modifications could create further disputes. I supported a defined reconciliation exercise, using engineering records, commercial correspondence and advice appropriate to each case. We would identify the missing agreement and ask for clarification, without treating incomplete documentation as evidence of wrongdoing by either side.

Over eight months we resolved sixteen of the twenty-one files through documented agreement with the relevant customers. Five remained open with named owners and a description of the unresolved point. The exercise cost S$280,000 in internal capacity and external assistance. In three resolved cases we accepted additional maintenance or preparation work before a future return, rather than assume that the customer would carry the cost. We did not claim that a completed register removed every possible disagreement about a tool’s condition or use.

I would have involved the programme teams earlier in explaining why the exercise mattered. Some initially experienced it as an audit of decisions made before their time. The useful result was a clearer future obligation, not a search for somebody to blame for an old file. That distinction improved the quality of the evidence people were willing to bring forward.

A prospective customer proposed a three-year supply programme in 2023 and asked for a dedicated production cell. Its forecast implied attractive annual volume, but the proposed agreement allowed monthly call-offs to change substantially. Our quotation recovered most of the cell’s fixed commitment through the unit price. If orders arrived as forecast, the economics worked. If they did not, we would have reserved the capacity without a corresponding payment obligation.

The sales lead wanted to accept the structure, arguing that the customer’s launch plan was credible and the relationship could introduce other work. Our plant manager asked which existing programmes would move when the dedicated cell became unavailable to them. I proposed a separate reservation fee and a smaller initial block of dedicated capacity, with the remaining requirement served through shared scheduling. The customer objected to paying for idle time and asked for a lower unit price instead. I explained that the two charges addressed different things and offered a shared-capacity alternative with different scheduling commitments.

We agreed a reserved block representing 60 per cent of the originally requested cell time, with a S$720,000 annual reservation fee and a separately agreed part price. During the first year, actual unit demand was 74 per cent of the original forecast. The reserved portion met the agreed availability conditions, while the unreserved capacity supported two other programmes. We gave up the prospect of being the customer’s exclusive source and incurred S$160,000 of additional scheduling and setup expense across the mixed workload.

The result did not prove that separate fees suit every programme. It showed that this buyer and supplier could agree explicitly on who carried the cost of access to capacity. I would involve the customer’s operations team sooner next time; its procurement team initially interpreted the fee as a price increase rather than a different allocation of a manufacturing obligation.

By 2024 two difficult product families depended heavily on six experienced setters. Managers could name potential successors, but those colleagues had little opportunity to practise the full changeover under supervision. Production demand usually took precedence, and the most capable setters continued to rescue the schedule. The dependence remained largely invisible because the jobs still shipped.

I asked our technical leaders to define a nine-month development programme for eighteen colleagues across Johor and Batam. It would include supervised changeovers during paid production time, with qualification based on demonstrated tasks and clear restrictions on work still requiring support. The finance lead proposed concentrating training into shutdown periods to reduce lost capacity. The technical team explained that some important decisions could only be observed during a real changeover with normal operating conditions. I approved S$540,000 for planned practice, supervision and temporary cover, while asking the plants to state the capacity cost in their schedules.

Thirteen participants completed the defined qualification by the end of the programme. Five continued supervised practice in specific areas; they were not described as qualified merely because they had attended. The number of product-family changeovers requiring one of the original six setters fell from 82 of 96 in the baseline review to 39 of 103 in the subsequent review. Job mix differed, so we treated that comparison as evidence of reduced dependence to investigate further, not a controlled productivity result. The programme also consumed production time that could otherwise have supported saleable work.

I would improve the explanation given to the original setters. Some feared that sharing their expertise would weaken their position. We made the mentoring responsibility explicit and recognised it in their role, but that should have happened before the first training calendar was published. Capability becomes more repeatable when experienced people have a credible reason to help build it.