Behavioural Change Toolkit · Advising Senior Leaders

The One-Page Board Narrative

The paper that gets a decision, not a nod.

The core principle: a board paper’s job is not to inform. It’s to get a decision made. If your paper has no decision in it, you have written a status update, and it will be noted and filed.


The template

[Change name] — Board Paper

From: ___ Date: ___ For: ☐ Decision ☐ Discussion ☐ Noting


The ask

(First. Always first. Never make a board hunt for it.)

We need a decision on _______________ by [date].


In one paragraph

We are [doing what] so that [outcome]. [N] people in [which roles] need to [do what differently]. We are [N] weeks in. The system will be ready; the question is whether people will use it.


Where we are

| | Status | Trend | |—|—|—| | System / delivery | | | | Behaviour adoption | __% (baseline __%) | ▲▼ | | Business outcome | | | | Team health | | |


The one thing that will determine whether this works

_______________ (One thing. Not five. Boards remember one.)


The decision required

| Option | What it costs | What it gets us | Risk | |—|—|—|—| | A | | | | | B | | | | | C — do nothing | | | |

Recommendation: ___, because ___.


What I’m worried about

_______________


What I need from this group specifically

  • ___________ — from [name], by [date]
  • ___________ — from [name], by [date]

Writing rules

Lead with the ask. Executives read the first two lines properly and skim the rest. Burying the decision on page two means it doesn’t get made.

Always include “do nothing.” It makes the trade real and forces the group to own the choice. It’s also occasionally the right answer, and offering it makes you look like an adviser rather than an advocate.

One worry, stated plainly. A paper with no worry reads as either naive or managed. The worry is what makes the rest of it credible.

Name people in the ask. “Leadership support is required” gets nothing. “I need [name] to confirm the legacy switch-off date by the 14th” gets a decision or a visible refusal — both useful.

Numbers with baselines and trends. “34%” means nothing. “34%, up from 11%, target 70% by week 12” means something.

Never spin. Boards see a lot of papers and are good at detecting managed ones. The credibility you get from reporting your own bad news is worth more than any individual number.


The verbal version — 90 seconds

You’ll often get less time than you expected. Have this ready:

“Three things.

One — the system will be ready on time. That was never the risk.

Two — adoption is at 34%, up from 11%. It needs to be 70% for the benefits case to hold. The gap isn’t understanding — 84% can explain why we’re doing it. It’s that it takes six minutes and nothing has come off anyone’s plate.

Three — I need one decision today: [the decision]. If it doesn’t happen by [date], the consequence is [specific].

My recommendation is [X].”

Practise it out loud. The version you’ve said aloud three times is materially better than the version you’ve only written, and boards can tell the difference.


Framing for different audiences

Audience Frame it as Lead with
Board / ExCo Risk to the benefits case The number and the decision
Finance Cost of partial adoption — you pay for the system either way Money
Operations Disruption management and the dip The plan for the dip
Risk / Compliance Two sources of truth during parallel running The control gap
HR / People Capacity, saturation, and manager load The people cost

Same substance, different entry point. This isn’t spin — it’s meeting people where their accountability sits.


The sentence that works on most boards

“We’ll deliver the system on time. What I can’t yet promise is that people will use it — and the benefits case depends entirely on the second thing, not the first.”

It’s true of almost every technology-enabled change, it’s uncomfortable in a productive way, and it reframes the discussion from delivery to adoption in a single line.


Change on a Page · Measurement Plan — Leading and Lagging · Challenging Upward — Scripts

← Back to the Behavioural Change Toolkit

Last updated: 11 August 2026