What is variance analysis in management accounting?
Variance analysis compares a reference value with an actual or new result. The reference may be budget, forecast, last year or an earlier period.
This should not be confused with the statistical analysis of variance known as ANOVA. In controlling and management accounting, the question is normally why revenue, cost, margin or profit changed and which drivers explain it.
1. Choose the right comparison and level
Decide which question the analysis must answer. Budget versus actual tracks the plan. Forecast versus actual measures forecast accuracy. Year over year shows development but is often affected by calendar, acquisitions and portfolio changes.
Choose the level of analysis as well. A company total can hide improving price in one region and falling volume in another. Start with the total and drill into dimensions that can genuinely explain the variance.
- Period and number of selling days.
- Currency and exchange rates.
- Product, customer and channel scope.
- Definition of net revenue, cost and gross profit.
2. Make sure the data is comparable
A technically correct bridge can still mislead if the data is not comparable. Check duplicates, returns, credits, new and discontinued products, and whether quantities use the same unit.
A price-volume-mix analysis normally needs a product or other analysis dimension, quantity and revenue for both periods. A gross-profit bridge also needs comparable cost data.
- The same key identifies the same product or customer in both periods.
- Quarter is compared with quarter and month with month.
- Revenue and cost use consistent signs and currency.
- New, lost and returned items are treated as explicit portfolio changes.
3. Start with the total variance
Calculate the change before trying to explain it. If revenue was 100 in the reference period and 112 in the actual period, the total variance is +12. The same logic applies to cost, gross profit or another selected result.
This total becomes the control for the full bridge. All explanatory effects must sum exactly to the total change, apart from an accepted rounding difference.
4. Separate price, volume and mix effects
The price effect shows the value of a change in unit price on the volume actually sold. The volume effect shows the value of a change in total quantity at a reference value. The mix effect captures a change in the distribution between products, customers or channels.
Several established methods exist, and they allocate interaction effects differently. State the method, sign convention and reference period. What matters is consistent application and a bridge that reconciles.
- Price: units sold × change in price per unit.
- Volume: change in total quantity × the reference average economic value.
- Mix: the value created by shifting volume between components with different price or margin.
- Portfolio: new, lost, returned or credited products separated when this improves understanding.
5. Add cost and currency effects when the question requires them
A revenue analysis may end after price, volume, mix and portfolio. Gross profit also requires changes in unit cost. Improved sales price can otherwise conceal an even larger increase in purchase or production cost.
For businesses operating in several currencies, separate currency from commercial performance. Use a documented constant-currency principle so the organisation can distinguish actual pricing from translation effects.
6. Build a reconciled bridge
Present the starting value, each effect and the ending value in a bridge or waterfall. Add a reconciliation control that shows the difference between the calculated ending value and the actual ending value.
If the bridge does not reconcile, do not use it for decisions. Common causes include missing products, incorrect signs, different aggregation levels or double-counted interaction effects.
7. Move from number to business explanation and action
Variance analysis becomes valuable when connected to causes. A negative volume effect may result from demand, stock shortages, customer loss, capacity or a deliberate profitability choice. The number shows where to ask; it is not the full answer.
Finish each material variance with cause, owner, proposed action and expected timing. Mark what has been verified and what remains a hypothesis.
- What happened and how large was it?
- Which underlying business driver caused the effect?
- Is the effect temporary, structural or a data issue?
- Which action will be taken, by whom and when?
Common variance analysis mistakes
The method must be easier to review than the variance it explains. Avoid these recurring problems.
- Comparing periods with different scope or definitions.
- Calling the full change volume while product mix also changed.
- Mixing currency translation with actual pricing.
- Hiding new or lost products in a residual category.
- Presenting effects that do not reconcile to the ending value.
- Showing a correct table without connecting it to cause and action.