# The Project I Almost Killed, And What Five Years of Hindsight Taught Me About Momentum
I still remember sitting across from my programme manager in month three, watching her make a genuinely reasonable case for why we should stop.
She was not wrong on the facts: the project had no visible wins, the team was tired, and the internal stakeholders who had commissioned the work had already moved on to three other priorities.
We were rebuilding a regulatory reporting framework for a business unit that had been quietly non-compliant for years, not dramatically, not dangerously, but consistently enough that someone senior had eventually noticed and handed me the problem, the kind of problem that arrives without fanfare and leaves without applause.
She laid out the argument for pausing clearly, professionally, with a slide deck that made stopping look almost responsible.
I nearly said yes, not because I believed it was the right call, but because I was tired too, and tired people find well-reasoned arguments more persuasive than they should.
The Situation
I was working on this project in 2019.
The business unit had been operating on a patchwork of manual processes and institutional memory for the better part of a decade.
Nobody had deliberately built a broken system: it had simply grown in the way most broken systems grow, one pragmatic workaround at a time, until the workarounds became the system.
My job was to replace it with something that would actually hold under regulatory scrutiny.
The challenge was that progress, in this kind of work, does not look like progress for a long time.
I was not building features, I was excavating.
Every week we found another process that existed only inside someone’s head, another data feed that connected to a spreadsheet no one had updated since 2016, another exception that had been handled manually so long it had stopped being seen as an exception at all.
To anyone watching from outside the room, we appeared to be standing still.
My programme manager’s case for pausing was rational: a pause, she argued, would give us time to regroup, re-engage stakeholders, and come back with a cleaner plan.
The logic was sound, and that is precisely why it was dangerous.
In complex rebuilds, “coming back with a cleaner plan” is usually a polite way of describing the moment a team loses its nerve and never quite recovers it.
I had seen it before, the pause that becomes a pivot that becomes a quiet cancellation eighteen months later.
What I said instead was this: we are not pausing, but we are going to do one thing differently.
Every week, I would identify one thing that is measurably better than it was seven days ago, not a milestone, not a formal deliverable, just one thing, documented, visible, shared with the team.
It sounds almost embarrassingly simple, that is, I suppose, the point.
What the Next Five Years Taught Me
The first insight is that momentum is not a feeling, it is a record.
What changed from month three onwards was not the pace of work, it was the existence of evidence.
Every Friday, there was something concrete to point to: a data feed validated, a manual step eliminated, a process documented for the first time.
Individually, each item was unremarkable, collectively, they became the proof that the project was alive.
And that proof did something I had not fully anticipated, it made the team stop measuring progress against the original plan and start measuring it against last week, that is a much more honest comparison, and a far more sustainable one.
I have since watched organisations spend enormous energy on formal programme governance, traffic-light reports, steering committees, milestone reviews, while neglecting the simpler practice of recording what actually improved.
The bureaucracy of progress is not the same as progress itself.
The second insight is that stopping teaches something you cannot easily unteach.
Had we paused in month three, the framework would probably have resumed eventually, projects like this rarely die entirely, they get restarted, rebranded, handed to someone new.
But the team would have carried a piece of learning that I think is genuinely corrosive in professional environments: that when progress is hard to see, stopping is the appropriate response.
That lesson travels, it shows up in the next project, and the one after that, as a lowered threshold for retreat.
The damage of an unnecessary pause is rarely visible in the project itself, it is visible in the people.
The third insight is the one that took longest to articulate clearly: consistency is not the slow path.
Most senior people I know, and I include myself in this, with some embarrassment, are instinctively drawn to the bold intervention, the restructure, the strategic pivot, the announcement that signals decisive leadership.
These things have their place, but I have watched more value created by sustained, unglamorous consistency than by any single bold move, and I have watched more value destroyed by the instinct to reach for drama when patience was actually what the situation required.
The regulatory framework that is now running across three regions was not born from a brilliant insight in month three, it was built one documented improvement at a time, over two years, by a team that had decided to stay in the room.
What This Means in Practice
If I am leading a complex programme right now, a technology rebuild, a regulatory change, a cultural shift, and progress is invisible, the question worth asking is not whether to pause, it is whether I have created the conditions for small progress to be seen at all.
Most organisations are reasonably good at celebrating the launch and catastrophically bad at recognising the incremental work that makes launches possible.
The team that ships quietly, week after week, without a milestone in sight, is doing the most important work in the building, they are also the most likely to be told to pause.
I build the record, I show the week-on-week line, I give the consistency somewhere to land.
The projects that become standards are rarely the ones that started boldly, they are the ones that refused to stop when stopping looked reasonable.