A development company should remember more than the buildings it completes.

In 2022 six project completions involved 112 Tervana employees moving from delivery assignments. The initial personnel plan placed most attention on the next posting. Asset leaders wanted access to the reasoning behind decisions they would inherit, while development leaders needed experienced colleagues on new work. Both requests were legitimate. The unresolved issue was the time a person could actually give to explaining the completed project once a new manager had begun allocating the next assignment.

I worked with development director Adrian Velsaro and the receiving asset teams to define a twelve-week handover responsibility for fourteen employees. We funded SGD 460,000 of overlap expense and identified the subjects each person would help explain. The arrangement concerned knowledge and working relationships; technical acceptance, contractual duties and property certification remained with the appropriate specialists. I did not treat a people handover as evidence that a building was ready for use.

Some project leaders preferred an immediate transfer followed by help when requested. I asked how that informal obligation would appear in the new role’s workload and assessment. The funded period made the trade-off visible. It also required receiving teams to prepare questions rather than assume that retaining access to experienced people would create learning automatically. We recorded which decisions had been examined and which dependencies still required another owner.

The result was a defined arrangement across the six completions, not a claim that every lesson was retained or every later issue avoided. The fourteen colleagues eventually moved fully into their subsequent responsibilities. The cost was accepted because the organisation wanted an explanation that documents alone had not provided. I would use the same question in a different project, but not automatically the same duration or headcount. The relevant decision is what needs to be understood, by whom and through which funded responsibility.

A 2024 review covered nine project-leadership assignments that connected development and operating interests. The role descriptions used broad language about coordination and accountability, but discussions with the appointees produced 27 disputed responsibility points. Several concerned decisions that one sponsor expected the project leader to make while another expected prior agreement. A more impressive title would not settle that difference.

I convened the relevant sponsors with the role holders and appropriate commercial colleagues. Eighteen points were resolved in agreed role records. Nine required decisions by the executive sponsors and remained explicitly open. We spent SGD 175,000 on facilitated reviews, project time and follow-up. The people team documented employment responsibilities; it did not rewrite investment approvals, technical authority or joint-venture agreements through a job description. Matters governed by those arrangements went to the people qualified to decide them.

The initial preference was to publish all nine roles as clarified before the next planning cycle. I opposed that description while substantive questions remained. The cost was an incomplete headline and further executive work, rather than a completed programme report. Role holders received an interim route for escalating the unresolved points so that ambiguity did not simply return to them in private conversations with competing sponsors.

The review produced eighteen settled points and nine visible decisions still required. It did not establish that every future conflict would be easy or that a written responsibility record could replace working relationships. Its value was a more honest account of authority at the time. I wanted performance discussions to reflect the decisions the organisation had actually assigned, not judge an individual against several incompatible interpretations. The example also reinforced a limit in my own function: documenting a conflict can make it actionable, but cannot resolve a matter outside our authority.

In 2023 a gap between major assignments affected 38 experienced employees. The pipeline supported several possible uses for their capability, but the dates were not firm enough to promise every person a particular project. An immediate reduction would lower expense; retaining the group would preserve experience at a known cost. I supported a five-month capability pool only after we described the work to be done during the interval and the point at which the decision would be revisited.

The arrangement cost SGD 1.12 million. It included defined internal assignments, peer support and preparation for likely operating needs, rather than treating employees as available without a meaningful role. Finance challenged the uncertainty in the pipeline. I agreed that the proposal should not present prospective work as committed revenue, and we recorded the alternatives if the expected assignments did not materialise. Employees received the same distinction between a possibility and a confirmed posting.

At the end of the period, 24 colleagues moved into new project roles, nine into asset operations and five left by agreement. Those outcomes reconcile the original 38-person population. Individual arrangements were handled through appropriate employment processes and advice. The account does not claim that every transition matched the employee’s first preference, or that the retention expense was recovered through a measurable project saving.

The decision bought time and preserved options, with a cost management could examine. It was not a standing promise that every future project gap would be funded in the same way. I would consider the capability, pipeline, available work and financial position afresh. The useful discipline was to avoid two misleading extremes: describing all retained capacity as waste, or treating experience as so valuable that its cost no longer required explanation. The organisation made a bounded investment and followed the employment account to a definite outcome.

In 2017 we invited experienced project managers to express interest in helping other teams examine difficult decisions. Twenty-nine colleagues came forward. The initial idea was to designate all of them as mentors, leaving each project to arrange the time. I questioned whether that would create a useful responsibility or merely a list of willing people whose main assignments remained unchanged.

We established twelve defined peer-review assignments with SGD 240,000 of annual time allocation and support. Selection considered the type of experience required and the capacity the colleague’s current manager could release. Seventeen interested employees were not appointed in that round, and we explained the decision without describing their experience as unimportant. The receiving project retained technical and commercial authority. A peer reviewer could test reasoning and offer an account of earlier experience, but did not become an additional approval route.

Some managers worried that formalising the work would make ordinary cooperation bureaucratic. I agreed that colleagues should continue to help one another informally. The defined assignments addressed a more sustained request that required preparation, scheduled participation and follow-up. We wanted that contribution visible in the person’s workload and assessment, rather than praised publicly and treated as a distraction when the home project reviewed performance.

The first review examined whether the receiving teams had used the contribution and whether the assigned time was realistic. It did not attempt to attribute a whole project result to a peer conversation. The outcome was a smaller, more accountable arrangement than the original proposal. I learned that recognition is strongest when it changes the conditions under which a contribution can be made. A title and a thank-you can be appropriate, but they cannot provide the time or decision boundaries that a sustained cross-project responsibility needs.

I helped introduce a project-completion review in 2010 that produced 46 lessons across four projects. The report looked substantial, and I initially treated the number of recommendations as evidence that useful learning had been captured. The next review found that only eleven had a clear receiving owner. Many of the others were broad observations that a later team could agree with without knowing what to do differently.

We returned to the project specialists, rewrote 21 lessons with a more specific decision context and archived fourteen that could not support a useful action. Alongside the eleven with existing owners, those categories account for all 46 original entries. The correction cost SGD 130,000 of facilitated review and project time. I took responsibility for designing an exercise around a visible output before establishing the use the organisation expected to make of it.

Some colleagues wanted to preserve every recommendation in the active register to avoid losing experience. I distinguished an archive from an instruction to future teams. Material could remain available without being presented as current guidance that someone was expected to apply. The rewritten entries identified the circumstances, the question and the receiving responsibility. Technical specialists remained responsible for evaluating the underlying conclusions; the people team improved the learning arrangement rather than certifying the advice.

The result was a smaller and more usable account, but the rewrite itself did not prove that a later project benefited. We still needed to ask whether the receiving team examined the lesson and what it decided. This failure changed the questions I used in subsequent reviews. I became less interested in the size of the document and more interested in the connection between one team’s experience and another team’s judgement. The organisation should be able to remember uncertainty as well as an apparently successful answer.