I remember a few years ago on a transformation programme, the data workstream stopped appearing in my week. No escalations, no amber, no requests for time. It had become the part of the programme I thought about least, which I took as a good sign and otherwise did not think about at all.
At the half-year review the board moved a third of its budget and two of its strongest people to the integration workstream, which had been red for a quarter and had everyone’s attention.
Nobody argued. I did not argue either. On the evidence in front of that board it was the obvious call.
One workstream was visibly in trouble.
The other was visibly fine.
Eight months later the data workstream missed a milestone for the first time in two years, and I sat and listened to three senior people explain that it had always been the fragile part of the programme. The record supported them. Two years of governance had produced almost nothing to say otherwise.
Governance runs on exceptions
Every reporting system built for a large programme is an exception machine. RAG ratings, escalation paths, risk registers, exception reports, the weekly call that exists to surface what has moved. The apparatus is designed to find deviation and route attention to it, and it does that well.
None of it is designed to surface prevention. A decision that stopped a problem from forming produces no status line, because there is no status to report. The register never holds the risk that was retired before anyone wrote it down.
So a programme’s information about itself runs systematically warm on what is going wrong and cold on what is being prevented. Nobody chooses this. It falls out of the design.
"The register never holds the risk that was retired before anyone wrote it down."
Three things follow, and only one of them is about people’s feelings.
Money and attention flow toward noise, which is the story above. Budget moves to whoever is loudest about the difficulty they are in, and the quiet workstream funds the rescue.
The wrong lessons get institutionalised. Programmes hold post-mortems on failures and almost never on the parts that worked, so the method that prevented six months of rework leaves with the person who used it, while the failure gets a lessons-learned deck and a process change.
"Budget moves to whoever is loudest about the difficulty they are in, and the quiet workstream funds the rescue."
And the people who prevent problems get read as ordinary, which I have very nearly done to the best operator on a programme, and wrote about on LinkedIn yesterday.
What actually goes missing
The risk retired in week three, before it was ever worth a line in the register.
The dependency renegotiated while it was still a conversation rather than a date.
The scope discussion held privately with a market lead, so it never became a change request.
The integration assumption tested early, which is why the September window needs no contingency.
Every one of those is a decision, made by a named person, with a cost avoided that can be estimated. None of them will appear anywhere in the programme’s record unless somebody puts them there deliberately.
Report decisions, not activities
An activity list dies in the inbox. It reads as effort, it invites no response, and nobody repeats it in another meeting.
A decision travels, because a decision has a shape a listener can hold: something was in the balance, a call was made, and something else did not happen as a result. That is a story, and stories are what get repeated in the rooms you are not in.
The unit of visibility on a transformation is not the update. It is the decision plus its avoided cost, attached to a name.
"Stories are what get repeated in the rooms you are not in."
The fortnightly line
This is the whole artefact. One decision per lead, per fortnight, six fields, no longer than the space below.
The decision. What was decided, in one line, with the date it was taken.
The alternative. What would have happened otherwise, in money, time or risk, with the basis for that estimate stated plainly.
The author. Who made the call, by name. Not the workstream, not the team.
The proof. The fact that shows it worked, or the date by which it will be visible.
The sponsor sentence. The single line your sponsor can say in their own meeting, without you in the room and without checking anything first.
The destination. Which forum needs this, and by when.
The sponsor sentence is the field people skip and the one that does the work. Write it as the sentence you want repeated: “We changed the cutover sequence in March, which is why the August window needs no downtime.” If your sponsor cannot say it from memory after reading it once, it is not finished.
Two practical points about where this lands. It goes into the body of what the board reads, not an appendix, because appendices are read by nobody who allocates budget. And it accumulates in the quiet weeks, which are exactly the weeks that feel safe to go silent. A record built only during crises is a record of crises.
"If your sponsor cannot say it from memory after reading it once, it is not finished."
Where this stops working
Three places:
The first is counterfactual inflation. Every avoided cost is an estimate about something that did not happen, and the temptation to round it upward is constant. A claimed disaster that nobody can verify costs you more than saying nothing, because it teaches the room to discount everything else you report. State the basis, keep the number conservative, and let it be dull.
The second is narration outgrowing delivery. There are people who have learned to report brilliantly and deliver adequately, and everyone in a programme knows who they are within two months. Making real work visible is a correction to a governance defect. Making thin work visible is a different activity, and the same room will eventually price it correctly.
The third is the uncomfortable one. Sometimes quiet is quiet because nothing hard is happening. The test is in the fields above: if a lead cannot name a decision and what it avoided, there may not have been one. That is worth knowing too, and it is the same instrument that finds the opposite case.
The part I keep off the feed
A programme’s record is not a neutral account of what happened. It is an argument about what mattered, and it is written mostly by whoever bothers to write it.
That argument outlives the programme. It decides what the client believes about which parts of the transformation created value, which shapes the next phase and the next budget. It decides which of your people are remembered as the ones who handled the hard year. And it decides, quietly, who gets trusted with the next transformation, because the people making that choice are reading a record rather than watching the work.
"A programme's record is an argument about what mattered, written mostly by whoever bothers to write it."
At programme manager the job is to deliver clean work. The system rewards that, and it should. Then the gate changes, and nobody announces the change. Above that line you are being assessed on whether value moves through an organisation because of you, and value that nobody can see has not moved anywhere. A transformation lead who delivers well and reports thinly is competing against one who does both, and the second one is the safer bet from where the decision is made, every time.
There is a version of this that people find distasteful, and I understand why. The correction is not to talk about yourself more. It is to put the decisions and their consequences into the record with names attached, including names that are not yours. Do that consistently and the record starts telling the truth about the whole programme, which happens to include you.
Problems tell their own story. Competence needs an author.
Reply and tell me about a piece of work on your programme that nobody outside it knows happened. The next issues get built from what comes back.
Roman