Budget against actual: the variances

Here is where the afternoon you spent on the plan gets paid back. Every month, when the statement arrives, you put it next to the column you wrote in advance and look at the differences. The exercise takes half an hour, needs no software, and turns a vague sense that "March was rough" into three specific numbers, two of which have obvious causes and one of which is worth a phone call.

🧾 Three columns and a difference

The sheet is as simple as it sounds: what you planned, what happened, and the gap. March in the example restaurant went like this.

Read the difference column and something does not add up. Sales missed by $2,000, two cost lines are over by $860 between them, one is under — and yet the result is $2,670 worse than plan. Where did the rest of it come from? And why does the food line, which came in under budget, turn out to be the one the kitchen needs to talk about?

Both questions have the same answer, and it is the one idea that makes variance analysis worth doing rather than just worth looking at.

🔀 Sold less, or spent more?

Two completely different things make a cost line miss its plan. The line can move because you sold less than you planned — you bought less food because fewer people ate — or because you spent more per peso of sales than you planned. The first is a sales problem and the kitchen had nothing to do with it. The second is a cost problem and the sales figure has nothing to do with it. A raw comparison mixes them together and produces nonsense: here it says the kitchen saved $190 in a month when its food cost actually got worse.

The fix is one line of arithmetic. Rebuild the budget using the sales that actually happened, and compare against that. Food at 32% of $45,000 is $14,400, not $15,040. Labour at $8,500 plus 17% of $45,000 is $16,150. Rent and everything else do not move at all, because they are fixed. That adjusted plan is what March should have cost given how March actually sold, and against it the picture changes completely.

Now the month makes sense. Slightly under half of the damage was simply selling less, and it costs 51 centavos of result for every peso not sold — that is the contribution figure from topic 3 doing its job. The rest was spending: the kitchen ran a point above plan, the rosters did not come down with the volume, and something worth $650 landed in the fixed block. Four numbers, four owners, and none of them is "March was rough".

🚦 Which differences deserve a conversation

A variance sheet with twelve lines on it will always show twelve differences, and treating all of them as findings is how the habit dies in month three. Set a threshold before you start and only talk about what crosses it. Something like more than $500, or more than 5% of the line works for a restaurant this size; scale it to yours so that three or four items a month survive, never ten.

In March, food at $450 does not cross it and gets watched rather than discussed. Labour at $550 and the $650 in "everything else" do cross it, so the meeting has exactly two subjects. The second one is the easiest to answer and the most commonly ignored: a single number in a fixed block that moved means something specific happened — a repair, an annual bill landing early, a subscription nobody remembers signing — and the answer is usually one line of the bank statement away.

• Look at the trend, not the month. One point of food cost over plan is noise. The same point three months running is a supplier price change, a portion that grew, or something leaving the building.
• Explain the favourable ones too. A line that comes in well under budget is as informative as one over: it can be a genuine saving, or a delivery that arrived after the cut-off, or a repair that was deferred and will be worse later.
• Watch out for inventory. If your food cost is measured on what you bought rather than what you used, a stock-up month looks like a disaster and the next looks brilliant. The financial reporting course explains how to tell.
• Keep a running year-to-date column. A $600 miss in March is small; a $600 miss every month is $7,200 by December, which is more than a month of rent.
• Write one sentence per finding, in the sheet. "Compressor repair, $650" next to the number is what makes next year's budget better, and it takes ten seconds while you still remember.

The last one matters more than it looks. A variance sheet with explanations written on it is the raw material for the next plan: it is how you learn that your fixed block is really $14,000 plus about $2,000 a year of repairs you keep treating as surprises, and how the following December you budget for them instead of being annoyed by them.

🗣️ What actually happens in the half hour

The mechanics are worth being blunt about, because this is the step that everyone intends to do and few keep up. It happens on a fixed date — the day after the statement is available, in the diary, monthly — and it involves whoever runs the floor and the kitchen, not just the owner. Ten minutes to fill in the actual column, ten to adjust for real sales, ten to talk about the three or four things over the threshold. Each one gets a name and a date, or it does not get discussed at all.

What it must not turn into is a search for someone to blame. Most variances are structural — a quiet month, a price rise, a repair — and the two or three that are behavioural come out much faster in a room where the sheet is the subject rather than the person. The chefs and managers who see the adjusted column quickly stop arguing about whether the month was busy, because the adjusted column has already agreed with them about the volume and moved on to what was spent on top of it.

And one honest note: a month where everything lands within the threshold is a good month, not a wasted meeting. It is also fairly common once the plan has been corrected a couple of times, which is what the last topic is about.