A factory can make an assembly accurately and still leave the future building operator with an awkward problem. The dimensions fit. The connections can be made. The finish looks consistent. Yet a component that will eventually need attention cannot be removed without disturbing something the project expected to remain in place. This is a different failure from poor manufacturing. It begins with an incomplete account of what the assembly must let people do.
My view is that selected service tasks belong in the factory design discussion while the arrangement can still change. That does not mean every imaginable future intervention should be rehearsed or that maintenance preferences should overrule all other requirements. It means the project should identify a defensible set of tasks and test their physical possibility alongside installation. A building method earns its usefulness over more than the period in which it is delivered to site.
This is particularly important when prefabrication concentrates decisions early. Repetition is valuable when the repeated design is suitable. Once materials, fabrication sequences and finishes are committed, a small access problem can become expensive to correct across a production run. The economics of a standard module therefore depend partly on asking the right questions before its apparent efficiency is multiplied. An access opening is not a minor finishing detail when it determines whether a replaceable component can actually be replaced.
Start with an actual task
At Varelim, an October 2024 review concerned ninety-six bathroom modules across four variants. One variant accounted for twenty-four modules. A full-size assembly could be installed, but the specified service component could not be removed through the available access after the finishes were in place. The drawing review had not made that task sufficiently explicit. The problem appeared when people tried to carry out the sequence physically.
The factory team favoured revising the work instruction. That was a reasonable question to examine: perhaps the difficulty arose from the order of operations rather than the geometry. Our operations colleagues wanted the access relationship changed and demonstrated. I supported a six-week redesign at a cost of S$420,000 because a written instruction could not by itself establish that the component had a usable removal path. We moved the opening and adjusted the relationship around it.
The revised arrangement allowed the specified task to be demonstrated. That is the extent of the conclusion. It did not show that every maintenance activity would be easy, that the module would never need destructive access or that every site condition had been represented. The affected release moved four weeks later and we accepted less standardisation. The other three variants continued their approved preparation. The decision had a cost as well as a technical justification.
The useful starting point was a task, rather than a general instruction to make the module maintainable. Who needs to reach what? What must be isolated or removed first? Which finishes will already be present? What space and tools does the procedure assume? These questions can reveal a conflict that the word accessible leaves hidden. They also help the team decide which concerns require a physical model and which can be resolved through an ordinary design review.
Give the physical review a decision to change
A mock-up can become theatre if the organisation has already decided that it cannot alter the product. The visit produces photographs, comments and a list of minor adjustments, while the important interfaces remain commercially fixed. I would rather conduct a smaller review at a point when a consequential finding can still change the arrangement. The value lies in the decision the exercise can influence, not in the size or polish of the demonstration.
That requires some uncomfortable planning. The project must identify who can accept a change, who owns the resulting cost and which commitments have already been made. Factory colleagues should know whether they are demonstrating a proposed design or confirming a design that has passed earlier reviews. Operations should identify the tasks it regards as material before the meeting. Otherwise each side can leave believing that a different question was answered.
The CTO does not replace the project engineer or the specialist responsible for the assembly. My responsibility is to fund the method of investigation, bring the necessary people together and make the consequences of the finding visible. In our module review, the project authorities retained their design and construction decisions. The technology function could recommend further work and explain why an instruction-only response was insufficient, but it could not convert a demonstration into permission for every application.
There is also a commercial argument for precision. A late objection described as general dissatisfaction is hard to price or resolve. A defined removal task, an observed obstruction and a revised arrangement create a much clearer conversation. The team can compare alternatives, identify which modules are affected and separate the cost of correcting this issue from unrelated improvements someone might prefer. That discipline protects the programme from both premature acceptance and uncontrolled redesign.
Keep standardisation useful after handover
Standardisation should reduce avoidable variation while preserving the work the building needs to support. Sometimes that means accepting two arrangements where one originally appeared attractive. Sometimes it means changing the component choice or its position. The answer cannot be settled by counting unique parts alone. A design with fewer parts may transfer effort into servicing, temporary protection, reinstatement or prolonged disruption to users. Those consequences deserve consideration even when the construction team does not carry the later operating budget.
I would now put the selected service-task list into the qualification brief before the first full-size review. The list would be agreed with the relevant operating specialists and tied to the actual arrangement being considered. It would not pretend to predict the building's entire future. Its purpose would be to make the early design choices answerable to a few concrete activities whose omission could materially weaken the proposal.
The record should then travel with the design. A successful demonstration becomes less useful if later substitutions change the component, access opening or surrounding finish without revisiting the task. This is not an argument against change. It is an argument for remembering why a particular relationship was chosen, so that a later team can recognise when its proposed alteration affects the earlier conclusion.
For senior leaders, the broader lesson is about what counts as completion. A method can be ready for manufacture while its operating consequences remain insufficiently examined. The right response is not to demand certainty about every future condition. It is to decide which foreseeable tasks matter, investigate them at a useful stage and retain the limits of that evidence. A factory drawing becomes a stronger business proposition when it describes an assembly people can continue to work with after the installation team has left.