Every Department Needs a Historian

Every programme I have worked on has had the same ghost in it.

A meeting ends. Everybody nods. Three weeks later nobody can agree on what was said, and two of the people who were in the room are completely certain in opposite directions.

Here is the part that took me longest to accept: almost nobody is lying. People remember the version of the meeting in which they come out well, and they remember it sincerely. Memory is not a recording. It is a story we tell ourselves afterwards, and the story is always flattering.

So the question stops being who is telling the truth, and becomes a much duller one. Who wrote it down.

What each of us remembers

"We agreed to change it."

"I thought that was only a suggestion."

"That was never mine to own."

What the record settles

Said

the words, and who said them

Agreed

what was actually decided

Dated

when it was confirmed

The person best placed to take notes is the person who cannot

Delivery teams do not skip documentation because they are careless. They skip it because the person best placed to capture what was said is the same person holding six to nine workstreams together, and that person is in the meeting to solve something, not to transcribe it.

You know how it actually goes. You are half listening because you are mentally unblocking the dependency that is going to ruin your Thursday. Someone says something that is quietly a brand new requirement. You clock it, you think “I will note that after this”, and then the next call starts four minutes later and it is gone. Not forgotten exactly. Filed somewhere behind the things that were on fire.

Multiply that by every meeting, every week, for the length of a programme. That is not a documentation gap. That is an entire parallel project nobody is tracking.

What actually goes missing

It is never the big decisions. Those get a deck. What evaporates is everything around them:

• The thing said in passing that was, on inspection, a new requirement
• The small accommodation you agreed to because it was easier than arguing, and then genuinely meant to come back to
• The decision that was quietly reversed, without anybody announcing a reversal
• The knowledge base that has now been restructured for the fourth time, each time at someone’s verbal request
• The user journey that changed shape between two Tuesdays
• The dependency that somebody else owned, confirmed they owned, and has not mentioned since

None of these are dramatic on the day. They are dramatic in month five, when somebody asks why the scope looks nothing like the proposal, and the honest answer is that it was changed forty times in conversation by people who do not remember changing it.

Memory is partisan, and the loudest memory wins

When a programme slips, the meeting about the slip is never really about what happened. It is about whose fault it is. And that is not settled by evidence. It is settled by whoever sounds most certain and can produce something in writing.

If your side cannot produce anything, you lose that meeting. Not because you were wrong. Because you were busy.

This is the bit that used to genuinely enrage me, and it is the reason I now think about this as a structural problem rather than a personal one. Being right and being able to demonstrate you were right are completely different capabilities, and only one of them survives an escalation.

Meetings are not tidy, and pretending otherwise helps nobody

Calls are messy in ways no methodology accounts for. Somebody forgets to mute and the entire room behind them joins the meeting. Someone who is three weeks late on their own deliverable finds remarkable energy to be sharp about yours. Tempers arrive, because deadlines make people frightened and frightened people are not graceful.

None of this is unusual. None of it is worth taking personally. But all of it eats the fifty minutes in which a decision was supposed to be made, and none of it ends up in anybody’s notes.

A historian does not fix any of that. What a historian does is make it survivable, because whatever did get decided in minute fifty two is written down before the call ends, by someone whose attention was never anywhere else.

So give somebody that job. Only that job.

Not a note taker. Not the most junior person, handed a notebook because they are cheapest. Not someone who also has three deliverables, because the moment delivery pressure arrives the notes are the first thing to go.

One person, whose entire role is the record:

• They are in every meeting that matters, and they are doing nothing else in it
• They capture what was said, who said it, what was agreed, and what was only implied
• They separate decisions from requests, and requests from wishes
• They flag the moment something crosses out of scope, on the day, while it is still cheap to say so
• They maintain one version of the truth that everybody can read, so there is never a question of whose notes are correct
• They are never pulled onto delivery work, because the week they are is the week the record stops

The last point is the one organisations get wrong, because an engaged, capable person sitting in meetings and typing looks like spare capacity. It is not spare capacity. It is the thing that makes everyone else’s capacity defensible.

A shield and a sword

I think of the output as two things at once.

It is a shield, because when the blame conversation comes, and it always comes, you are not reconstructing six months from memory against somebody who has already decided the answer. You have dates. You have the request, the date it arrived, the date you flagged the impact, and the date it was approved anyway.

It is a sword, because the same record is what you take upward. Escalation without evidence is just complaining with a bigger audience, and everybody in the room knows it. Escalation with a timeline is a different conversation entirely, and it tends to be a short one.

As a shield

You are not reconstructing six months from memory against somebody who has already decided the answer. You have dates.

As a sword

Escalation without evidence is complaining with a bigger audience. Escalation with a timeline is a short conversation.

I want to be careful here, because it would be easy to read that as adversarial. It is not, or at least it should not be. The record protects the client exactly as much as it protects you. It is the thing that lets both sides stop arguing about the past and get back to the part where work gets done. Most of the time its real value is that the fight never starts, because nobody can build a grievance on a version of events that is written down and agreed.

What not having one costs

You already know, because you have paid it.

It costs scope you delivered and never billed. It costs commitments you honoured and got no credit for, because nobody remembers you making them. It costs reputation in rooms you were not in. It costs the specific, grinding morale damage of being blamed for a delay you flagged in writing, except you did not flag it in writing, because you were busy fixing the delay.

1. Unrecorded changesaid in passing, never written down
2. Conflicting expectationstwo sides, both sincerely certain
3. Rework and delaythe same work, paid for twice
4. Blame without evidencesettled by whoever sounds most certain

A historian is cheaper than any one of those. It is one of the few roles I can think of that pays for itself the first time it is needed, and quietly earns its keep on every single day it is not.

The lesson, learned the expensive way: somebody is going to write the history of your project. If it is not you, it will be whoever has the most to gain from how it reads.