How to Build a Budget You'll Actually Use

Most small business budgets are built once, in a spreadsheet, using percentages from an article, and never compared to anything afterwards. That's a document, not a control.

A budget is a set of decisions made in advance — what you'll spend on what, decided while you're thinking clearly rather than in the moment when something feels urgent.

Build it from your own history

Skip the benchmark percentages. "Marketing should be 8% of revenue" is an average across businesses nothing like yours.

Export the last twelve months from your accounting package and categorise every expense. It takes an afternoon and it's the entire foundation.

Two things fall out immediately:

• What you actually spend, which is usually different from what you think
• Subscriptions and commitments you'd forgotten about, which for most small businesses is a genuine and immediate saving

Split fixed from variable

The most useful structural decision in the whole exercise.

Fixed costs happen regardless of revenue: rent, salaries, insurance, core software, your own pay. They set your survival floor — the number you must produce every month whatever happens.

Variable costs scale with delivery: subcontractors, materials, payment processing, per-unit fees. Higher revenue means higher variable costs, which is fine and not dangerous.

FIXED (monthly)                    VARIABLE (% of revenue)
  Salaries + your pay    £____       Subcontractors      __%
  Rent / space           £____       Materials           __%
  Insurance              £____       Payment fees        __%
  Core software          £____       Delivery costs      __%
  Accounting / legal     £____
  ─────────────────────────────
  MONTHLY FIXED          £____   ← this is your survival floor

Your survival floor is the number that matters most in this document. It tells you the minimum revenue that keeps the business alive, and it's the number to check before any decision that adds fixed cost.

Budget costs against conservative revenue

The common failure: budgeting against optimistic revenue, then discovering the costs were committed and the revenue wasn't.

Take your realistic low case — not your best month, not your average of a good year. Budget fixed costs so they're covered at that level.

Then treat revenue above the conservative case as a separate decision rather than assumed spending. It funds discretionary things: a marketing experiment, a tool, a bonus, reserves. Deciding that in advance is what stops a good quarter turning into permanently higher fixed costs that a bad quarter can't support.

Include the things people leave out

Four categories that reliably get missed and reliably cause problems:

Your own pay, at a real number, as a fixed cost. If the budget only works because you take nothing, the business isn't profitable — it's subsidised by you, and that should be visible rather than hidden.

Tax reserve. Set aside as it accrues, not when the bill arrives. Ask your accountant for the right percentage for your structure rather than guessing low.

Annual and irregular costs, divided by twelve. Insurance renewals, software billed annually, professional fees, equipment replacement. These feel like surprises and are entirely predictable.

Replacement and repair. Laptops fail, equipment wears out. A small monthly allocation turns a crisis into a purchase.

Budget is not cash

A distinction worth being explicit about, because conflating them causes real problems.

| | Budget | 13-week forecast | |---|---|---| | Question | Can we afford this? | Can we pay for it in March? | | Horizon | A year | A quarter | | Unit | Amounts | Timing | | Built from | Categories of spend | Bank movements |

You can be exactly on budget and still unable to make payroll, because a large client paid at 60 days instead of 30. The budget won't show that; the 13-week cash flow forecast will.

Most small businesses need both. They're not alternatives.

Review monthly — the variance is the point

The budget itself does nothing. The monthly comparison is the whole value.

Fifteen minutes, once a month, alongside your other numbers:

                   BUDGET    ACTUAL    VARIANCE
Revenue            ______    ______    ______
Fixed costs        ______    ______    ______
Variable costs     ______    ______    ______
Contribution       ______    ______    ______

Three questions on the variances:

• Which were one-offs and which are the new normal? A one-off is noise; a pattern is a budget that needs changing.
• Did anything come in under for a bad reason — deferred maintenance, marketing you skipped because you were busy?
• What does this change about the rest of the year?

Then update the budget when reality has clearly moved. A budget defended against evidence for eleven months stops being useful in about March.

When to cut, and what

If revenue drops below the survival floor, the order matters:

1. Discretionary spend first — experiments, non-essential tools, anything with no contracted commitment
2. Variable costs — these fall with revenue anyway, but check they actually are
3. Renegotiate fixed commitments — many suppliers prefer a lower price to losing you
4. Your own pay, temporarily and deliberately, with a date to revisit
5. Headcount, last, and with professional advice — employment obligations vary by jurisdiction and improvising here is expensive

Cutting marketing first is the common instinct and frequently the wrong one, because it lengthens the downturn you're cutting to survive.

The mistakes

1. Benchmark percentages instead of your own history. Yours is specific.
2. Not splitting fixed from variable. You never learn your survival floor.
3. Budgeting against optimistic revenue. Costs commit, revenue doesn't.
4. Leaving out your own pay. Hides whether the business actually works.
5. Confusing budget with cash. On budget and out of money is a real state.
6. Never comparing to actuals. The variance is the entire value.

What to do next

Export twelve months of expenses and split them into fixed and variable. The fixed total is your survival floor, and it's probably the single most useful number you don't currently know.

Then check it against your worst month of the last year. If the floor is higher, that gap is your most urgent finding.