The Complete Guide to Planning a Small Business Year

Small business planning fails in one of two ways. An annual plan written in January that nobody opens after March, or no plan at all, where whatever arrived in the inbox that week sets the agenda.

The fix isn't more planning. It's a shorter cycle, fewer commitments, and deciding in advance what evidence would change your mind.

The rhythm

| Cadence | What happens | Horizon | |---|---|---| | Weekly | Are we on track? Blockers cleared. | This week | | Monthly | Budget vs actual. Forecast vs actual. | This quarter | | Quarterly | Choose three priorities. Re-check the constraint. | Ninety days | | Annually | Budget, and only the direction that genuinely spans a year | The year |

Most of the value is in the quarterly and monthly rows. The annual layer is thinner than people expect — a budget and a direction, not a detailed plan.

Plan in ninety days

Three priorities. Not twelve. Twelve is the same as none, because everyone quietly picks their own three and you find out at the end of the quarter which three each person chose.

Good priorities are structural rather than routine, finishable in ninety days, and owned by one named person. And each needs a finish condition written at planning time — "onboarding documented and run by Sam without me, for three clients" can be checked in December; "improve onboarding" can be argued about indefinitely.

Then the step that makes the rest real: decide what you'll stop. Your quarter is already full. New work displaces something, and the only question is whether you choose what or whether reality chooses by dropping whatever seems least urgent.

And review before you plan. An hour on what you said you'd do, what actually got finished, and — the most useful question — what consumed the time that wasn't on the list.

Expect to finish about two of three. Three every quarter means you're planning too conservatively; one means you're over-committing.

Full cycle and the half-day agenda: quarterly planning.

Build the budget from your own history

Skip benchmark percentages. Export twelve months, categorise it, and split fixed from variable.

Fixed costs set your survival floor — the number you must produce every month whatever happens, and probably the most useful figure in the whole document. Variable costs scale with revenue and are far less dangerous.

Budget fixed costs against conservative revenue, not your best month. Then treat anything above that as a separate decision rather than assumed spending — that's what stops a good quarter turning into permanently higher fixed costs a bad quarter can't support.

Include the four things people leave out: your own pay at a real number, tax reserve as it accrues, annual costs divided by twelve, and replacement for equipment that will eventually fail.

And be clear that budget is not cash. A budget says what you'll spend; the 13-week forecast says whether you can pay for it in March. You can be exactly on budget and unable to make payroll.

Full method: how to build a budget you'll actually use.

Forecast in three layers

Layered, in decreasing certainty — this is what makes forecasting possible with limited history.

1. Committed. Contracted revenue, mapped to when you'll be paid rather than when you deliver. If this covers fixed costs, the month is safe regardless of what else happens.
2. Weighted pipeline. Open opportunities × your own close rates by stage. Template probabilities are a guess with a decimal point; before twenty closed deals, don't weight at all.
3. Not yet sourced. Derived from planned activity and your conversion ratios — and never presented with the confidence of layer 1. Most forecasting failures are an aspiration treated as a projection.

You can forecast about one sales cycle ahead and no further. If the quarter looks thin two months out, that's a prospecting problem today, not a forecasting one.

Produce three cases — low, expected, high — and commit fixed costs against the low one.

Then record forecast against actual every month. After three months the useful thing appears: your own bias, which is usually stable and correctable.

Full method: forecasting revenue with limited history.

Stress-test before reality does

Four scenarios cover almost everything that actually happens:

• Your largest client leaves. Concentration is the risk small businesses most reliably underprice.
• Revenue drops a third for two quarters.
• A large invoice pays 60 days late. This breaks more profitable businesses than lost revenue does.
• You're unavailable for a month. The most revealing, and usually the least modelled.

Run each against the 13-week forecast so the answer is a number of weeks, not an adjective. Then write the response down in advance — what gets cut, in what order, at which week.

Decisions made under financial pressure are reliably worse than the same decisions made calmly three months earlier, and the delay while you work out what to do is frequently more expensive than the choice itself.

Do it even when the answer is reassuring. Knowing your downside is what lets you take reasonable risks — decline bad work, commit to a hire, price with confidence.

Full models: stress-testing a small business.

Know when the plan is wrong

Persistence and stubbornness look identical from the inside. The only defence is deciding in advance what would change your mind.

Write the falsifier when you commit, not afterwards:

By 30 November I expect 20 conversations, 8 proposals, 3 signed. Fewer than 4 proposals by 31 October means the offer isn't landing, and I rework it rather than pushing harder.

Watch leading indicators, not revenue — revenue lags a full sales cycle, so waiting for it puts the decision six weeks late. Conversations, proposal rate, close rate: each points at a different fix.

Separate the three failure modes, because they look the same and need opposite responses: the idea is wrong, the execution is wrong, or it needs more time. The honest test is whether the trend is flat or slow — flat after a fair test means stop; slow means continue.

And give things a fair test. Changing course every six weeks produces activity and no compounding.

Full diagnostic: a bad quarter or a broken strategy.

What this connects to

• The startup version — if the business is new, the one-page plan comes first. This guide is what replaces it once there's history to plan from.
• Execution — the weekly review is where the quarter stays honest. Three priorities at the top, one line each, every week.
• Growth decisions — whether to scale is a planning decision, and it needs the budget and forecast to answer.

The mistakes, collected

1. An annual plan and nothing else. Dead by March.
2. Twelve priorities. Everyone picks their own three.
3. No finish condition. It drifts and gets carried forward.
4. No stop-doing list. The plan has nowhere to fit.
5. Budgeting against optimistic revenue. Costs commit; revenue doesn't.
6. Presenting unsourced revenue as pipeline. The main way forecasts mislead.
7. No falsifier written in advance. Every later judgement is a rationalisation.

Where to start this week

Two numbers, both about an hour of work.

Your survival floor — twelve months of expenses, split fixed from variable. The fixed total is the minimum the business must produce every month.

Your committed revenue for the next three months. If it covers the floor, you know the business survives those months regardless of what else closes.

Most owners have never calculated either, and together they change how the next quarter gets planned.