Variance analysis: turning a difference into a reason

Variance analysis is the practice of decomposing the difference between what you expected and what happened into its causes. A month that came in fifteen per cent under plan tells you only that something went wrong. The same month split into price, volume and mix tells you which lever moved, and therefore what to do — and those are three different conversations with three different people.

The habit it replaces is the one where somebody explains the month narratively — it was quiet, the big customer went slow — and everybody nods. Sometimes that story is right. Variance analysis is how you find out.

The three that matter for most businesses

• Price variance: you sold at a different price than planned. Discounting, a promotion that ran longer than intended, or a currency movement on imported goods.
• Volume variance: you sold a different quantity. Demand, capacity, or a sales process problem.
• Mix variance: you sold a different combination. Revenue can hit plan while margin misses badly because customers bought the cheaper line — this is the one that hides most often behind an on-target revenue figure.
• On the cost side the same split applies: you paid a different price for inputs, or used a different quantity of them.

Working through one

1. Start with the total variance and its sign. Under or over, by how much, in money rather than percentage.
2. Split revenue into units sold and average price achieved. Compare each against plan separately — this alone resolves most variances.
3. Check mix by looking at the split between product lines against the planned split. If units and price both look fine and margin still missed, mix is the answer.
4. Do the same on the cost side: input prices against planned, quantities used against planned.
5. Rank what you find by money, not by percentage. A ninety per cent variance on a small line matters less than three per cent on the main one.
6. Write down the cause and what changes because of it. A variance identified and not acted on is arithmetic for its own sake.

Favourable variances deserve the same examination as adverse ones, and almost never get it. A month that beat plan because a customer ordered early has not beaten plan — it has borrowed from next month, and next month will look like a problem that needs solving. Unexamined good news is how businesses respond to a phantom decline by cutting something that was working.

Where it goes wrong

• Comparing against a plan nobody believed. If the budget was aspirational, every month is adverse and the analysis stops being read.
• Percentage thinking. Rank by absolute money or the small lines dominate the discussion.
• Explaining rather than analysing. It was quiet is not a variance category; fewer transactions at the same average value is.
• Analysing everything. Set a threshold — in money — below which a variance is noted and not investigated.
• Doing it too late. A variance explained six weeks after the month is history; the same variance in week one is still actionable.

Where it lives

Variance analysis is the analytical half of the monthly pack described in management accounts, and it needs a plan to compare against — the forecast side is covered in financial projections template. Ettex Books provides the actuals, and the decomposition itself belongs in Ettex Sheets, where the split into price, volume and mix stays visible rather than being asserted.

The simpler monthly comparison that most small businesses should start with, before decomposing anything, is covered in budget vs actual.