Building next year's operating plan (start now, not in December)

Most small businesses build next year's budget somewhere between Christmas and New Year's Eve, in about two hours, by taking this year's numbers and adding a growth percentage. Then it goes in a folder and nobody opens it again until someone asks in April why the numbers look nothing like it.

The problem is not the effort. It is the timing. A plan built in late December is a description of what you hope happens, because every decision that could have changed the outcome — hiring, pricing, a lease, a marketing commitment — needed to be made months earlier. Built in September, the same plan is a set of choices you still have time to make.

A plan is not a forecast

These get used interchangeably and they are different documents with different jobs. A forecast is your best guess at what will happen. A plan is what you are committing to make happen, and what you will spend to do it.

The distinction matters because it changes how you respond when reality diverges. A forecast that turns out wrong gets updated. A plan that turns out wrong triggers a decision — you either change what you are doing or you consciously accept the new trajectory. If your annual document cannot produce that moment, it is a forecast wearing a plan's clothes.

Build it in five layers, in this order

Order matters, because each layer constrains the next. Doing it backwards — deciding what you want to earn and reverse-engineering the revenue — is how you end up with a plan nobody believes.

• 1. Revenue, by driver.Not last year plus 12%. Build it from the terms of your revenue engine — customers and churn, or capacity and utilization, or spend and conversion. If you cannot name the drivers, that is the first thing the exercise has told you.
• 2. Cost of sales.Usually a margin percentage applied to the revenue build, adjusted for anything you know is changing — a supplier increase, a new product with different economics.
• 3. People. Every role, its fully loaded cost, and the month it starts. This is the single biggest lever in the plan and the one most often entered as a lump sum, which hides exactly the timing question that matters.
• 4. Everything else. Rent, software, insurance, marketing, professional fees. Most of it is last year with known changes; do not spend three hours refining a line worth $4,000.
• 5. Cash and capital. Working capital as revenue grows, capital purchases, loan payments. This is where a plan that looks profitable reveals itself to be unfundable.

That last row is the reason to do this in September. The plan is profitable and still consumes $27,000 of cash, because of a truck and a loan that never appear on the income statement. Discovered now, that is a problem with four months of options — delay the truck, finance it, stage the second hire, arrange a facility. Discovered in March, it is an emergency.

Two scenarios, not seven

Build the plan you actually intend to execute, then build one downside case: revenue 15% below plan, everything else unchanged. The only question that case needs to answer is whether you stay solvent and which specific commitments you would pull.

Naming those in advance is most of the value. "If we are 15% light by the end of Q2, the second hire waits and the trade show is cancelled" is a decision made calmly, in advance, rather than under pressure with incomplete information. More than two scenarios is almost always procrastination.

Then actually use it

A plan earns its keep only if reality gets compared to it every month. Set this year's actuals beside the plan, look at the variances that exceed some threshold you set in advance, and ask which are timing and which are trend.

Timing variances resolve themselves — an expense landed a month early. Trend variances do not, and they are the ones that compound. Catch one in February and you have eleven months to respond. That is the entire return on the exercise, and none of it is available if the document was built in the last week of December.