Key takeaways
- A variance is only useful when it is explained by a business driver — volume, price, mix, cost rate or efficiency — not just reported as a difference.
- Bridge budget profit to actual profit one driver at a time, and make sure the bridge adds up exactly.
- Always analyse sales mix separately: total volume can rise while contribution falls.
- Good commentary states the variance, the cause, the amount, whether it will recur, and what management should do next.
Every month, finance teams produce a report that compares actual results with the budget. In many companies that report stops at two columns and a difference: revenue was BDT 2 crore below budget, overhead was BDT 15 lakh above budget, profit missed by 18%. Management reads the numbers, asks “why?”, and the finance team goes back to the ledger to find out. The meeting that should have been about decisions becomes a meeting about explanations.
Variance analysis fixes this. Done properly, it breaks the gap between budget and actual into a small number of drivers that operational managers recognise and can act on. This guide explains the method I recommend for management reporting: the profit bridge, the price–volume–mix split, cost variances, materiality thresholds and how to write commentary that leaders actually use. A worked example with simple numbers runs through the whole article so you can follow every calculation.
1. Why a simple “actual minus budget” is not enough
A raw difference mixes several effects together. Suppose revenue is above budget. That could be good news (customers bought more), neutral news (prices rose but so did input costs), or a warning sign (the company sold more low-margin products while high-margin lines declined). The total difference is the same in all three cases, but the management response is completely different.
The purpose of variance analysis is therefore not arithmetic. It is attribution: assigning each part of the gap to the driver that caused it, and to the manager who controls that driver. Sales owns volume and price. Procurement and production own material prices and usage. Operations owns efficiency. Overhead owners own their spending lines. When each variance has an owner, the monthly review becomes a performance conversation instead of an accounting debate.
2. The worked example
We will use a single-product business to keep the arithmetic clear. The budget assumed 10,000 units at a selling price of BDT 50, a variable cost of BDT 30 per unit and fixed overhead of BDT 120,000. Actual results show 11,000 units sold at BDT 48, a variable cost of BDT 31 per unit and fixed overhead of BDT 125,000.
| Line | Budget | Actual | Difference |
|---|---|---|---|
| Units sold | 10,000 | 11,000 | +1,000 |
| Selling price per unit | 50 | 48 | −2 |
| Variable cost per unit | 30 | 31 | +1 |
| Revenue | 500,000 | 528,000 | +28,000 |
| Variable costs | 300,000 | 341,000 | +41,000 |
| Contribution | 200,000 | 187,000 | −13,000 |
| Fixed overhead | 120,000 | 125,000 | +5,000 |
| Operating profit | 80,000 | 62,000 | −18,000 |
All figures in this article are illustrative and rounded for teaching. They are not taken from any employer’s accounts.
On a two-column report, the story looks confusing: revenue is up by BDT 28,000, yet profit is down by BDT 18,000. The profit bridge below explains why.
3. Build the profit bridge driver by driver
The profit bridge — often shown as a waterfall chart — starts with budget profit, adds or subtracts each variance, and ends with actual profit. The rule that keeps the bridge honest is simple: each driver is measured while holding the others constant, and the sum of all variances must equal the total difference exactly. If it does not add up, there is a calculation error or a missing driver.
= (11,000 − 10,000) × 20 = +20,000 favourable
= (48 − 50) × 11,000 = −22,000 adverse
= (30 − 31) × 11,000 = −11,000 adverse
= 120,000 − 125,000 = −5,000 adverse
Check: 80,000 + 20,000 − 22,000 − 11,000 − 5,000 = 62,000. The bridge closes.
Swipe sideways to see the full chart →
Budget-to-actual profit bridge (illustrative, BDT)
View the data behind this chart
| Step | Amount | Effect |
|---|---|---|
| Budget profit | 80k | Starting point |
| Sales volume | +20k | Favourable |
| Selling price | −22k | Adverse |
| Variable cost | −11k | Adverse |
| Fixed overhead | −5k | Adverse |
| Actual profit | 62k | Result |
Notice that the volume variance is valued at budget contribution per unit, not budget selling price. This is deliberate. Selling one extra unit does not add BDT 50 to profit; it adds the BDT 20 margin that unit earns after variable costs. Valuing volume at selling price is one of the most common errors in management reports, and it exaggerates the benefit of growth.
4. Separate sales mix from sales quantity
Most real businesses sell more than one product, customer group or channel, and each earns a different margin. When that happens, the volume variance must be split into two parts: a quantity effect (we sold more or fewer units in total) and a mix effect (the proportion of high- and low-margin products changed).
Take a second illustrative business with two products. Product A earns a contribution of BDT 25 per unit and Product B earns BDT 10. The budget assumed 6,000 units of A and 4,000 units of B — a 60:40 mix and a weighted average contribution of BDT 19 per unit. Actual sales were 5,000 units of A and 6,000 units of B.
Swipe sideways to see the full chart →
Units sold by product: budget vs actual (illustrative)
View the data behind this chart
| Category | Budget units | Actual units |
|---|---|---|
| Product A | 6,000 | 5,000 |
| Product B | 4,000 | 6,000 |
= (11,000 − 10,000) × 19 = +19,000 favourable
= (5,000 × 25 + 6,000 × 10) − (11,000 × 19) = 185,000 − 209,000 = −24,000 adverse
Total volume grew 10%, and a quantity-only view would report a BDT 19,000 gain. But the mix moved toward the cheaper product, and that cost BDT 24,000. The combined volume effect is a loss of BDT 5,000. Without the mix split, management would celebrate growth that actually destroyed margin. This is exactly the kind of insight a CFO expects from finance.
In practice, mix can be analysed by product, customer, region, channel or even order size. Choose the dimension that matches how the business makes decisions. In an export garment business, for example, mix by buyer and product category usually explains far more than mix by individual style.
5. Analyse cost variances: price, usage and efficiency
On the cost side, the same logic applies. A material cost variance splits into a price variance (we paid more or less per kilogram or metre) and a usage variance (we consumed more or less material per unit than the standard). Labour splits into a rate variance and an efficiency variance. Overheads split into spending and, where overhead is absorbed on activity, a volume or capacity variance.
| Cost element | Price / rate variance | Quantity / efficiency variance | Usual owner |
|---|---|---|---|
| Direct material | (Std price − Actual price) × Actual quantity | (Std quantity for actual output − Actual quantity) × Std price | Procurement / Production |
| Direct labour | (Std rate − Actual rate) × Actual hours | (Std hours for actual output − Actual hours) × Std rate | HR / Production |
| Variable overhead | Spending: Budget at actual activity − Actual cost | Efficiency: (Std hours − Actual hours) × Std rate | Operations |
| Fixed overhead | Spending: Budget − Actual | Volume: (Actual output − Budget output) × Std rate | Department heads |
The split matters because the actions are different. A material price variance caused by a currency movement calls for a treasury or sourcing response. A usage variance caused by fabric wastage calls for a production response. Combining them in one line hides both.
6. Set materiality thresholds so the team focuses
A management pack that explains every variance explains nothing. Agree a simple rule with leadership, such as: comment on any line where the variance exceeds both 5% of budget and a fixed amount appropriate to the business (for example BDT 5 lakh). Everything below the threshold is reported but not narrated. The threshold should be documented, applied consistently, and reviewed once a year.
Two exceptions are worth adding. First, any variance that is small but recurring for three months should be commented on, because it signals a structural change rather than noise. Second, any variance that affects a covenant, tax position or regulatory ratio should be commented on regardless of size.
7. Write commentary that leads to action
Commentary is where variance analysis becomes useful or useless. Weak commentary repeats the number (“Revenue was BDT 28,000 above budget”). Strong commentary follows a five-part structure:
- What: the variance and its size.
- Why: the operational driver, in business language.
- How much: the financial effect of that driver.
- Will it recur: one-off, timing, or structural.
- So what: the action, owner and expected impact on the forecast.
8. Common mistakes to avoid
- Valuing volume at selling price instead of contribution, which overstates the benefit of growth.
- Using a static budget to judge cost control when activity changed. Flex the budget to actual volume first, then compare costs.
- Bridges that do not add up — usually caused by mixing actual and budget quantities inconsistently.
- Ignoring mix in multi-product businesses.
- Explaining variances only in accounting terms (“higher accruals”) rather than operational terms (“overtime during the September peak”).
- No link to the forecast. A variance that will continue must change the rolling forecast.
9. Tools: Excel, Power BI and ERP
The method works in any tool. In Excel, keep one tab of drivers (units, prices, unit costs, overheads) for budget and actual, and calculate each variance with explicit formulas so a reviewer can trace them. In Power BI, model budget and actual as separate fact tables sharing product, customer and date dimensions, then build measures for each variance and a waterfall visual for the bridge. In SAP CO, standard cost estimates and variance categories in product costing provide price and quantity variances at order level, which can feed the same management bridge.
Whatever the tool, the discipline is the same: consistent definitions, a bridge that closes, and commentary owned by the people who run the business. For connecting this monthly analysis to the forward view, see my guide on driver-based budgeting and forecasting, and for presenting it to leadership, the Power BI MIS dashboard case study.
Frequently asked questions
What is the difference between a sales volume variance and a sales quantity variance?
The sales volume variance measures the total profit effect of selling more or fewer units than budget. In a multi-product business it splits into a quantity variance (change in total units at the budget average margin) and a mix variance (change in the proportion of products sold).
Should volume variances be valued at selling price or at contribution?
At budget contribution (or standard profit) per unit. One extra unit adds its margin to profit, not its full selling price, so valuing volume at selling price overstates the impact.
What is a reasonable materiality threshold for variance commentary?
Many teams comment on variances above both a percentage (often 5–10% of budget) and an absolute amount suited to the size of the business. The rule should be agreed with management, documented and applied consistently.
Conclusion
Variance analysis turns the monthly report from a scorecard into a decision tool. Start with a profit bridge that closes exactly, value volume at contribution, separate mix from quantity, split cost variances into price and usage, and write commentary that ends with an action and a forecast impact. With that structure in place, the monthly review stops asking “what happened?” and starts asking “what are we going to do about it?”
Written by Liakat Hossain
Finance and accounts professional with experience in FP&A, SAP FICO, Power BI, internal control, reporting, audit, tax, VAT and export-oriented business operations.
This article is original educational content based on professional finance practice. Real-project examples are anonymized, and scenario figures are identified as model outputs rather than guaranteed results.