The Pack That Gets a Budget Approved in One Meeting
A budget meeting that runs long is usually a meeting where the pack did not say what was being decided.
Most approval packs present next year’s numbers in the same format as this year’s reporting and hope the reader arrives at the conclusion the preparer already reached. Sometimes that works. When it does not, the meeting turns into a line by line interrogation of a document that was never built to be interrogated line by line, and it finishes without a decision, because no decision was ever actually put.
So build the pack backwards. Write the decision list first, before a single page of the pack exists, and let the pack be whatever those decisions need in order to be taken. If a page does not help somebody take one of the decisions on that list, it is reference material and it goes behind the tab at the back where reference material lives.
The second position, and the one people resist: the page that does the work is the bridge from the current year, not the plan itself. Nobody can evaluate an absolute number for a year that has not happened. Everybody can evaluate a movement, because a movement has a cause, and a cause can be agreed with or argued with. Put the bridge on page two and watch how the conversation changes.
Who is approving, and why the pack exists anyway
Some readers have a board with independent directors and a formal resolution. More have two shareholders who are also the two people who run the company. Some have one owner, and a lender who will be sent the same document afterwards because the facility requires an annual plan. Write for the role rather than the institution: somebody has to say yes, and after they have said it the company has to know what they said yes to.
That second half is why the pack exists even where the approver built half the plan themselves. It is the same instinct behind everything on the management reporting side of an engagement: the number is worth less than the record of who agreed to it. In a company where the owner set the revenue direction and made the headcount call, the pack is not persuasion. It is the record that fixes what was decided, so that the variance conversation nine months later is about performance rather than about memory.
Page one: the decisions
One page, at the front, before any financial statement. Every row is something the approver has to resolve in this meeting or explicitly defer.
| Column | What goes in it |
|---|---|
| The decision | Stated as a question with a yes or no answer, or a choice between named options. Not “consider the hiring plan.” |
| What it changes | The lines of the plan that move if the answer goes each way, named by line |
| What it costs, and when | Cash effect and the month it lands. Where the cost is a commitment rather than a payment, say which. |
| What happens if it is deferred | Usually the honest answer is that it gets decided by default in six weeks by whoever is closest to it |
| Recommendation | Yours, written down, with the reasoning in one sentence |
| Decision taken | Blank on issue. Filled in the meeting, in the room, before the next row is opened. |
That last column is the whole point of the page. A pack that comes back with page one still blank did not get approved, whatever anybody said as they were leaving.
Keep the list short and be ruthless about what qualifies. A first budget usually has between four and eight real decisions in it. The hiring plan is always one. Anything requiring a commitment the company cannot easily reverse is one. A pricing move is one. Everything else is arithmetic that follows from those, and putting arithmetic on the decisions page teaches the approver that the page is padding.
Page two: the bridge
The bridge starts at what the current year will actually land at and finishes at the plan, with every step in between attributed to a cause. Not to an account. To a cause.
The distinction matters because an approver cannot do anything about an account. They can do something about a cause. “Salaries and wages up by a stated amount” invites no useful response. The same movement split into a full year effect of people hired part way through this year, a wage movement applied to everyone, three planned roles, and one role deliberately not replaced, invites four separate responses, and three of those four are decisions on page one.
Sequence the steps so that the ones the company controls come last:
- The starting point. What the current year finishes at, using the latest reforecast rather than the original budget, and say which one you used.
- One-off items dropping out. Everything in the current year that will not repeat, listed individually. Anything you cannot name individually is not a one-off, it is a habit.
- Full year effect of decisions already taken. Hires who started in the second half, a price change made mid year, a contract that started or ended part way through. This block is not a proposal. It is arithmetic on things that already happened, and marking it clearly stops the approver arguing with the parts of the plan that are no longer open.
- Known contract movement. Wins already signed, and losses already notified. Signed and notified only. A renewal you expect belongs in the next block.
- Assumed movement. Volume, price, mix, renewal rates, cost inflation on inputs. Each one a separate step, because each one is an assumption somebody can dispute independently.
- Proposed investment. The hires, the capital, the programme spend. Every step here maps to a row on page one.
- The plan.
Read that sequence as an approver and you can see what it does. By the time you reach step six you have already accepted steps one through five or asked for them to be changed, and step six is the only part left to decide. A bridge that mixes the seven together is a bridge that gets argued from the top every time somebody objects to anything.
The rest of the pack
| Section | What it is | Why it is in this order |
|---|---|---|
| The plan, at summary level | Revenue, gross margin, operating cost, result, by quarter | Read after the bridge, so it is read as a destination rather than as a proposal |
| The plan, by month | Same lines, twelve columns | The seasonality is visible here and nowhere else |
| Cash | Monthly closing cash and headroom, the thinnest month marked | The single page a lender-facing approver turns to first |
| Capital and commitments | Every item, with the month the commitment is made and the month the cash moves | Kept apart from operating cost because it is a different kind of decision |
| Covenant projection, where there is a facility | Each covenant projected to each test date in the plan year | Built the way the lender reporting pack and covenant certificate is built, from the agreement’s own definitions |
| The assumption register | Every assumption, its owner, its source, and what it drives | Behind the plan, not in front of it. It is the audit trail, not the argument. |
| Scenarios | Two, not five | Below |
| Detail by department or account | Whatever the approver might ask for | At the back, unread in a good meeting |
Two scenarios, and neither of them is “optimistic”. Run a downside that is specific enough to act on, built by moving the two or three assumptions that actually carry the plan, and state what the company would do in it and at what point it would have to decide. Run one upside if there is a real decision inside it, meaning a case where doing better than plan requires spending something in advance. Three cases of increasing cheerfulness is not analysis. It is a range presented as work.
The assumption register is not the scenario section. It records what was assumed and who owns it, so that in the second quarter a variance can be traced to an assumption with a name on it rather than to a general sense that things turned out differently.
What to leave out
A prior year comparative for every account down to the last expense line. The approver has been reading those monthly all year.
The methodology. How revenue was built is a decision that was taken weeks earlier by the people who build revenue, and re-litigating the method during an approval meeting is the single most reliable way to lose the meeting. One sentence naming the method, in the assumption register, is the correct treatment.
Anything the finance function is still arguing about internally. An unresolved question presented to an approver as a choice they did not know they had is how a meeting acquires a ninth decision it was not prepared for.
Circulation, and the rule that saves the meeting
Send it far enough ahead that the approver can read it, which in practice means at least three working days, and never the night before. Then send a note, separately, naming the three things you expect to be argued about. This feels like handing over your weak points. It is the opposite: it moves the argument off the clock, so that the objection arrives by email on the second day and can be answered with a schedule rather than from memory in the room.
Do not present the pack in the meeting. A preparer who walks the approver through page after page has converted a decision meeting into a reading session, and the decisions get taken in the last eight minutes when everyone is tired. Open on page one, row one, and go.
When it is not approved
Approval is not the only good outcome, and a pack that produces a clean refusal has still done its job.
What must not happen is the third outcome, where the meeting ends warmly, nothing on page one is filled in, and the plan is treated as approved by everyone who was in the room. That plan will be loaded into the ledger, reported against for two quarters, and then disowned in the first uncomfortable month. If page one comes back blank, the budget is not locked, the reforecast dates are not yet published, and the correct next step is a second meeting with a shorter decisions page rather than a load.
What we do not do
We do not present a plan we would not sign, and where the owner has directed a number we do not believe, we write the direction and the disagreement into the assumption register rather than absorbing it quietly into the model.
We do not build a scenario section to demonstrate rigour when the two or three assumptions that carry the plan can be moved on a single page.
And we do not put a number on the decisions page without the month it lands in, because a commitment with no date attached is the one an approver says yes to and then does not recognise when the cash goes out.