Quick Answer: Before approving a 2027 budget, a CEO should be able to answer eight questions: what revenue growth the plan assumes and why, which costs are reversible, what gets cut first if revenue misses, what the budget funds that is genuinely new, how headcount maps to work, what it assumes about AI, what the board will challenge, and who owns each number.

Table of Contents

Key Takeaways

Every number in a budget is derived from the revenue assumption, so the revenue assumption is the only line worth interrogating first.
Splitting costs into fixed and reversible tells a CEO how much of the plan can actually be changed mid-year if revenue misses.
Pre-agreed spending triggers, set before January, remove the emotion and delay from mid-year cost cuts.
A budget that funds nothing new is last year’s plan with inflation applied, and it should be named as such rather than presented as strategy.
Every line in an approved budget needs a named owner, or the plan has no mechanism for accountability when it drifts.

Budget season produces a document that looks like arithmetic and behaves like a forecast. The spreadsheet is precise to the dollar; the assumptions underneath it are guesses, and the guesses are where the risk lives.

The questions below are ordered the way a CEO should work through a plan — from the assumption that drives everything, down to the accountability that makes the plan enforceable. None of them require a finance background to ask. All of them are hard to answer well, which is the point.

1. What revenue growth does this plan assume, and where does it come from?

Start here, because every other number is downstream. Headcount, marketing spend, infrastructure, and the cash position at year end are all sized against a revenue figure that someone chose.

The useful follow-up is decomposition. Growth from existing customers, growth from new customers, growth from price, and growth from new products are four different bets with four different risk profiles. A plan that says “22 percent growth” without saying which of those four it comes from has not been examined.

Ask what the same figure looked like in last year’s plan and what actually happened. A team that missed its growth assumption two years running and has submitted a third identical number is telling you something.

2. Which of these costs can we actually reverse?

Split every cost line into reversible and fixed. A reversible cost can be stopped inside one quarter without breaking a contract or damaging the business — month-to-month software, contractor hours, discretionary marketing, travel, events. A fixed cost cannot: leases, salaried headcount, multi-year licenses, committed capital projects.

Reversible cost: spending that can be halted within a single quarter without breaching an agreement, losing a capability the business depends on, or incurring a termination penalty that exceeds the savings.

The ratio between the two is the number that matters. A plan that is 85 percent fixed is a plan with almost no steering wheel. That may be the right structure for a stable, capital-intensive business and the wrong one for a company entering an uncertain year — but the CEO should know the figure before signing, not discover it in May.

3. What gets cut first if revenue comes in 15 percent light?

Name the list now, in writing, while nobody is under pressure. Mid-year cuts made in a hurry are made politically — the loudest department keeps its budget and the quietest one loses a headcount it needed.

Attach the list to a trigger rather than a date. “If bookings run more than 15 percent below plan for two consecutive months, tier one of the reduction list executes automatically” is a decision made once, in calm conditions. It converts a future argument into a prior agreement.

This is also the cheapest stress test available. If a leadership team cannot produce a credible cut list, the budget has no slack in it, and that is worth knowing in December rather than June.

4. What does this budget fund that we were not already doing?

Most annual budgets are the previous year’s budget with adjustments. That is not a criticism — continuity is how operating companies work — but it should be visible rather than disguised as strategy.

Ask for the plan to be split into three buckets: run the business, improve the business, and change the business. Then ask what percentage sits in the third bucket. The number is often close to zero, and leadership teams are frequently surprised by it.

A CEO who cannot point to the line items that fund something genuinely new is approving maintenance. That may be correct for the year ahead. It should still be a choice.

5. Does headcount match the work, or last year’s org chart?

Headcount plans tend to be built by asking each function what it needs, then trimming. That process reliably produces an org chart shaped like last year’s, because each function argues from its current structure.

The alternative question is what work the company has committed to in 2027 and what roles that work requires. Where those two lists diverge — roles funded for work that is ending, work committed with no role attached — is where the plan will break.

Pay particular attention to roles being backfilled automatically. A departure is the cheapest opportunity a company gets to re-examine whether a position still matches the work, and it is routinely spent refilling the old job description. The discipline of declining the default option applies to headcount as much as to opportunities.

6. What does this budget assume about AI, explicitly?

Nearly every 2027 plan contains an AI assumption. Very few state it. The assumption usually hides inside a headcount line