Standing up a function that outlived me

Context

When I arrived, change management wasn’t a function. It was a set of habits. Design and scope modifications moved through the project by whoever happened to be handling them, and the approval trail depended on who remembered to ask. On a lump-sum job that is not a paperwork problem — it is the mechanism by which a contract quietly stops meaning what it says.

I was the sole cost, estimating and forecasting authority on the portfolio. The path of least resistance was obvious: absorb change control into my own work. I already had the context, and doing it myself would have been faster than teaching anyone.

The mandate I gave myself

I decided against that, for a reason that is the whole point of the story: you can do far more by educating and delegating than by holding the work. Handling every change myself would have made me indispensable and capped the outcome at one person’s throughput. Building the function meant the capability outlived my involvement — and it meant somebody else got the responsibility and the room to grow into it.

That is a trade. It is slower at the start, and it costs you the thing consultants are quietly incentivised to protect.

The decisions

Making the downside legible before asking for money

There is always pushback — nobody funds a control function on principle, because the return is invisible until it isn’t. So I didn’t argue in principle. I showed what unmanaged change actually does to a lump-sum job: modelled impact against the real portfolio, set beside AACE’s published research on change-order cost growth. The argument stopped being “we should have governance” and became “here is the number we are exposed to, and here is what it costs to close it.”

Designing the approval logic before designing the process

The workflow, the states and the decision rights came first. Procedure followed. Governance that starts as paperwork gets routed around; governance that starts as a decision model gets used.

Setting the structure, then getting out of the way

I laid out the plan, the document structure, what needed controlling, the expectations and standards, and the lessons I had already paid for. Then I let him run it.

He shadowed me for the first few weeks. I also had to teach him construction itself — how to read a drawing, what a specification is for, which standards govern what. He didn’t arrive knowing the domain; he arrived wanting to.

What I deliberately did not do was supervise the operating of it. Control doesn’t produce anything good. You set the structure, you are honest about the standard, and then you have to have faith in people — and what you find is that people do great things. Had I stayed in the loop to keep it correct, it would have stayed correct and stayed small.

Outcomes

The department exists. It grew from one person to a team of nearly ten. The practitioner I trained runs it independently today — I have no role in it.

Alongside it, rebuilding the change lifecycle to close approval gaps and prevent scope leakage contributed to a 15% reduction in total project cost through renegotiated change orders.

What I carry forward

Real growth comes from sharing what you know. The instinct to hold the work is the instinct that caps it — at your own hours, your own attention, your own tenure.

Structure is the thing worth controlling. People are not. Set the standard clearly enough that someone can meet it without you in the room, then leave the room. The measure of a function you built is whether it still runs after you have stopped touching it.