Address small problems while they can still be reversed

Migrations leave behind unvalidated states. Reorganisations transfer properties to new entities and forget the small things. Regulatory changes get implemented at speed and never revisited. Mergers consolidate organisations on paper while leaving two parallel object hierarchies in the system. None of these qualify as defects. What they leave is sediment. Tiny gaps in the data that grow without alarm until a reconciliation breaks a property turns up under two identifiers or an expired obligation continues to generate postings.

When the posting run completes I check first for volume not for type. A small number of contracts will never post correctly and I gave up on the root cause years back. Now I track them on a separate list. Finance reaches out when the lease liability for a particular building has moved between periods. Sometimes I find the explanation. Sometimes I close the ticket with a note to keep watching.

The portfolio I am responsible for covers several hundred properties spread across multiple legal entities. Budget commitments are recorded against it. Audit reports cite it. Political stakeholders raise questions about it. The system was designed to deliver a single source of truth. In practice it delivers that view plus a spreadsheet on the side where I align what the system itself cannot. I know this arrangement is not sustainable. Whoever takes over this role after me will take over the spreadsheet as well..

There are screens I avoid and reports I never run. Not because they are irrelevant. The last time someone touched them something broke that took us three weeks to trace. I know which transactions are safe. Everything outside that perimeter feels expensive.

When someone proposes a change consolidating cost centres moving properties into a new entity activating a new disclosure report my first reaction is defensive. How many interfaces does that touch. Which mappings are hard-coded somewhere I cannot see. What will it break in LUM in CAFM in the consolidation. I usually answer with “let us park that for now.” I cannot frame what I suspect clearly enough to escalate without sounding like I do not know my own portfolio. So I swallow it.

If someone asked me to walk through the system end-to-end I could give the high-level tour. If they started drilling why this property sits under that entity why this lease uses that condition type why this number ties to LUM but not to Finance I would run out of answers fast.

There are screens I never open. Reports I never run. Not because they are irrelevant, but because the last time someone touched them, something broke that took us three weeks to trace. I know which transactions are safe. Everything outside that perimeter feels expensive.

When someone proposes a change — consolidating cost centres, moving properties into a new entity, activating a new disclosure report — my first reaction is defensive. How many interfaces does that touch? Which mappings are hard-coded somewhere I cannot see? What will it break in LUM, in CAFM, in the consolidation? I usually answer with “let’s park that for now.” I cannot frame what I suspect clearly enough to escalate without sounding like I do not know my own portfolio. So I swallow it.

If someone asked me to walk through the system end-to-end, I could give the high-level tour. If they started drilling — why this property sits under that entity, why this lease uses that condition type, why this number ties to LUM but not to Finance — I would run out of answers fast.

I stopped asking whether the system is genuinely correct and shifted to making it through each budget cycle. Acknowledging that something at the foundation could be wrong would mean opening a door I cannot close again. Internal audits oversight committees political questions the restatement of public figures expectations I do not have the capacity to carry. As long as the numbers do not appear absurd they pass as good enough.

Over time RE-FX and I have adjusted to one another. My routines my spreadsheets my unwritten rules adapt to its quirks. The unusual things become “just how we do it here.” I have become the missing manual. In a public entity where staff turnover retirements and reorganisations are part of normal operation that is its own kind of risk.

The cost of real change quietly grows. Repairing the system would mean unravelling years of tacit knowledge and undoing the personal logic I have built on top of it. The system has become personal. It runs on my memory my judgement my presence. And one day it will run on my succession.