What it calculates

A budget-versus-actual analysis compares what you planned to spend and earn against what actually happened, line by line, and flags each variance as favourable or unfavourable. It is the core monthly discipline of financial planning.

How it is calculated

Variance is actual less budget, but the sign convention depends on the line. On revenue, actual above budget is favourable. On costs, actual above budget is unfavourable - the same arithmetic sign means opposite things. Percentage variance puts lines of different sizes on a comparable footing, and rolling the detail up to department and company level shows whether individual variances offset or compound.

How to read the result

Chase the large percentage variances on small lines as well as the large absolute ones, because a cost line running 60% over is usually a broken assumption rather than noise. Distinguish timing from permanent variance: an invoice that landed a month late reverses next period and needs no action, while a rate increase does not. The rollup matters as much as the detail - a department can be on budget in total while hiding a large overspend offset by an unfilled vacancy.

Worked example

Marketing budgets 100,000 and spends 118,000: an 18,000 unfavourable variance, 18% over. Revenue budgets 900,000 and delivers 940,000: 40,000 favourable, 4.4% over. Net variance is 22,000 favourable, but the marketing overspend still needs explaining because the two are unrelated.

Common questions

Should variances be measured against budget or forecast?

Both, for different purposes. Budget variance measures performance against the commitment made at the start of the year and is what accountability rests on. Forecast variance measures the quality of your most recent estimate and is what tells you whether your forecasting process works.

What variance threshold is worth investigating?

Use two tests together: an absolute floor so you ignore trivial amounts, and a percentage so you catch small lines behaving badly. A common pairing is anything above 5% and above a fixed currency amount, tuned so the exception list stays short enough to actually be read.

How do you separate volume from price effects?

Decompose the variance: the volume effect is the quantity difference at budgeted price, and the price effect is the price difference at actual quantity. Selling more units at a lower price can produce a flat total while hiding two large offsetting variances.