Four stages, and each one ends in a document
We narrow the question until a board can answer it. We gather the evidence and check where every piece of it came from. We write down what the evidence cannot settle. Then we hand the file over and sit in the room while the board decides. The same two people carry the file from the first stage to the last, and one of them is always a partner. A full cycle runs eight to twelve weeks.
A decision gets read twice, and the second reader is never in the room
So we write the reasoning while the work is happening. Reasoning reconstructed afterwards is a different thing, and it reads like one.
Regulator wanted the reasoning, not the resultThe first reader is the board on the day of the vote.
The second is a regulator, a court, or a board seated years after the vote, working from whatever survived.
A deck does not survive that reading. It carries the conclusion and very little else: not the alternatives that were dropped, not the date on which the numbers were true.
The order is fixed, and the first stage can end the engagement
The stages run in order, and each one ends with something written. Stages two and three sit inside the evidence review, which is why they share its window. Stage one can also end the engagement. Narrowing the question sometimes shows that the client's own evidence already settles it, and then we stop there.
Narrow the question until the board can answer it
Stage one, two weeks
Most files arrive as a question no vote can settle. We take it apart and hand back one the board can answer, with the alternatives written next to it. We also name the evidence that would change the answer. That gives the next stage a scope instead of an open brief.
What it produces
The framing note. The question in one sentence, the alternatives on the table, the evidence that would move the answer, and the motion the board will vote on.
Gather the evidence and check where it came from
Stage two, in review
Every claim that reaches the board gets a source, a date, and the name of the person who checked it.
Where the evidence is thin, we say thin and show why. Where it does not exist, we say that too, and we write the recommendation so it survives the gap.
What it produces
The evidence record.
One line per claim:
- what it says
- where it came from
- when it was true
- how far it can be pushed
Write the reasoning, including the options we set aside
Stage three, inside the same review
We write up the alternatives that lost as carefully as the one that won, each with the reason we set it aside. This is the part a deck drops, and the part the second reader asks for first. Writing it while the argument is still live is the whole point.
Before it goes to the board
The two people on the file argue it between them before anything goes to the board.
Whatever objection survives that argument goes into the section on what would make the recommendation wrong.
What it produces
The decision document.
01
The recommendation
02
The alternatives with the reason each was dropped
03
The limits of confidence
04
The section on what would make the recommendation wrong
Write the paper for a forty minute board session, and run it
Stage four, two weeks
Most strategy work does not fail in the analysis. It fails in the handover. The board paper is written for a board session that runs forty minutes. The full file sits behind it, in order, for the people who go looking. The session is run by the two people on the file, because the questions that matter are the ones a summary cannot answer.
What it produces
The board paper, with the full file behind it in order, and the session it was written for.
We write down what we do not know, in the same document
Confidence is not one number for the whole file. It sits at the level of the claim. A board reading the paper can see which parts of the argument still hold if one number turns out to be wrong. Nobody has to ask us.
Each claim in the evidence record carries
how we know it
The named source, and the person here who checked it against that source.
when it was true
The date the figure describes, which is rarely the date the paper was written.
what would have to happen for it to stop being true
Written before the vote, so the board knows in advance what would reopen the file.
We mark a figure we checked at source differently from one a client department reported to us, and both differently from an estimate.
Every recommendation carries a section on what would make it wrong
It is a list of conditions, each one written so somebody can check it later.
The numbers that would have to move, and by how much.
The assumption that holds only while a regulator holds its current position.
The date to look again at a thin piece of evidence.
An auditor reading it long after the vote can see we named the risk before anyone knew the outcome.
Tomas Reinholt, Partner, health systems
When the evidence arrives late
Evidence does not arrive on the board calendar. Decide before the vote what would count as a reason to look again, and the late arrival becomes a check instead of an argument.
The material stays with you, in a form you can change
We hand over working files, not a locked PDF. The reason is practical.
A decision document gets amended when circumstances move, and quoted in a submission to a regulator.
The next piece of work extends it, and that work may not be ours.
If a client has to ask us for a copy in order to defend their own decision, the record is not really theirs.
At close, the files transfer with the reasoning intact, and the client can edit any part of it.
“We inherited four hundred slides and no memory of the choice. The next board inherits a document it can argue with.”
We keep the file as it stood on the day
112 files sit in the archive, from 2019 to 2026, across 9 sector regulators. Anders Vik keeps it. It exists for one moment: when somebody reopens a decision and asks what was known at the time. We can show the file as it was on the day of the vote, with the evidence dated and the alternatives still in it. Files from this archive have been cited in proceedings.
The five case files on this site come out of it, written up with the client's own account of the decision.
Case file
The regulator wanted the reasoning, not the result
Sector
Energy and utilities
File status
Reopened by the regulator.
The evidence arrived after the vote
Sector
Health systems
File status
Closed, then checked against the evaluation.
Two committees, one decision
Sector
Financial services
File status
Reopened at the exit review.
A decision nobody could reconstruct
Sector
Public infrastructure
File status
Rebuilt from the archive.
The board that asked for less
Sector
Financial services
File status
Closed at eight pages.
Two of the four on every file, and one of them is always a partner
Four people. Two of them are on your file from the framing note to the board session. The partners cover energy and health systems. The principals cover financial services and public infrastructure.
Mireille Osei
Partner, energy and utilities
Sat on an energy regulator's appeal panel for four years.
Tomas Reinholt
Partner, health systems
Built clinical trial protocols before Plenary.
Nadia Fournier
Principal, financial services
Writes most of the firm's method.
Anders Vik
Principal, public infrastructure
Keeps the decision archive, the one cited in proceedings.
Four terms, the same on every file
The same two people from the framing note to the board session, one of them a partner.
No juniors on the analysis.
A full cycle is eight to twelve weeks, framing to handover.
We do not run a two week version of an eight week question.
Every recommendation names the conditions that would make it wrong, in a form somebody can check later.
Boards use it to set review points.
At close, the files move to you in a form you can edit, with the reasoning left in.
Bring the question you cannot answer yet
The first conversation is about the decision in front of you and the date it has to be made. If framing is all you need, that is two weeks, and we will say so. If the papers you send already answer the question, we will tell you that and send no proposal. All four of us see what arrives. The one who covers your sector replies, and that person would carry the file.