The material bill went up. Was it the market, or did the shop floor scrap more than usual? Without a standard, nobody can say. With one, the month's variance splits into 10 named pieces, each with somebody who can act on it, and they sum to the last rupee.
A manufacturer that costs its output at whatever it happened to pay this month has no way of telling a good month from a bad one. The material bill went up: was that the market, or did the shop floor scrap more than usual? The labour bill went up: was it overtime, a rate rise, or slower work? Actual cost on its own is a number with no reference point.
A standard cost gives it one. It says, in advance, what one unit of this product should consume and what each of those inputs should cost. Every month the actual is held against that standard, and the difference is broken into named pieces, each with a person who can do something about it. That is the whole purpose. A standard that nobody is measured against is a spreadsheet nobody opens.
A standard cost is a promise about quantity and price per unit. Variance analysis is the monthly audit of that promise, split so that the buyer answers for price, the shop floor answers for quantity, and nobody can hide behind the other.
The most common failure is not in the analysis. It is in the standard itself. A standard set from last year's average, rounded, with a contingency added "to be safe", produces variances that are always favourable and never believed. A standard set from the engineering drawing with zero allowance for scrap produces variances that are always adverse and equally ignored.
A usable standard has 3 parts for each material, and 2 for each hour of work.
How much of the input one good unit consumes, including the normal loss. If the sheet yields 96 brackets in a good week and the drawing says 100, the standard is 100 divided by 96 of a sheet per bracket, not 1 over 100. Normal loss is in the standard. Abnormal loss is the variance.
What the input should cost per unit of measure for the coming period, at the supplier and terms the business intends to use. Not last year's average and not the cheapest quote ever received. If purchases are in a foreign currency, the standard carries a stated exchange rate, so the currency movement can be shown on its own later.
For labour and for overhead, the hours one unit should take at a normal pace, and the cost of one such hour. The overhead rate is the budgeted overhead for the period divided by the budgeted hours, which is why the budget has to be agreed before the standard can be.
Standards are set once a year in most businesses, and that is usually right: a standard that moves every month is an actual with a different name. When a material's market price shifts permanently, change the standard at a stated date, and record the old and the new, so the variance before and after can still be read.
One product, a steel bracket, so every figure can be checked by hand. The numbers are illustrative and do not describe any real business.
| Standard per bracket | Quantity | Price | Cost |
|---|---|---|---|
| Steel | 2 kg | 50 per kg | 100 |
| Paint | 0.5 litre | 40 per litre | 20 |
| Labour | 0.5 hour | 200 per hour | 100 |
| Variable overhead | 0.5 hour | 60 per hour | 30 |
| Fixed overhead | 0.5 hour | 100 per hour | 50 |
| Standard cost | 300 |
The fixed overhead rate of 100 per hour comes from a fixed overhead budget of 500,000 for the month over a budgeted 5,000 hours, which is 10,000 brackets at half an hour each.
The month is not what the budget expected. The factory made 9,000 brackets, not 10,000. Here is what it actually consumed and paid.
| Actual for the month | Quantity | Price | Cost |
|---|---|---|---|
| Steel bought | 19,000 kg | 52 per kg | 988,000 |
| Steel used | 18,500 kg | ||
| Paint bought and used | 4,600 litres | 39 per litre | 179,400 |
| Labour | 4,700 hours | 205 per hour | 963,500 |
| Variable overhead | 290,000 | ||
| Fixed overhead | 510,000 |
At standard, 9,000 brackets should have cost 9,000 multiplied by 300, which is 2,700,000. What they actually cost is discussed at the end, because the honest total depends on a decision about where the steel price variance is taken. First, the pieces.
Taken at the moment of purchase, on the whole quantity bought, not only on what was used. The buyer bought 19,000 kg of steel; the buyer answers for all 19,000.
Steel: 19,000 multiplied by (50 minus 52) = 38,000 adverse.
Paint: 4,600 multiplied by (40 minus 39) = 4,600 favourable.
Valued at standard price, so that the shop floor's number does not move when the buyer's does.
Steel: 9,000 brackets should use 18,000 kg. 18,500 kg were used. (18,000 minus 18,500) multiplied by 50 = 25,000 adverse.
Paint: 9,000 brackets should use 4,500 litres. 4,600 were used. (4,500 minus 4,600) multiplied by 40 = 4,000 adverse.
On the hours actually paid. 4,700 multiplied by (200 minus 205) = 23,500 adverse.
At the standard rate. 9,000 brackets should take 4,500 hours. 4,700 were worked. (4,500 minus 4,700) multiplied by 200 = 40,000 adverse.
Variable overhead is assumed to move with hours, so it splits the same way labour does.
Expenditure: the 4,700 hours worked should have carried 4,700 multiplied by 60 = 282,000 of variable overhead. Actual was 290,000. 8,000 adverse.
Efficiency: the 200 extra hours each carried 60 of variable overhead. 12,000 adverse.
This is the one that confuses people, because fixed overhead does not move with output but the standard cost per unit pretends that it does. The pretence is called absorption, and the variance is the price of it.
Expenditure: budget 500,000, actual 510,000. 10,000 adverse. This is the only fixed overhead variance that describes money spent.
Volume: the standard absorbs 50 per bracket. 9,000 brackets absorbed 450,000. The budget expected 500,000 to be absorbed. 50,000 adverse. No money was spent here. The factory simply made fewer units than the rate assumed, so the fixed cost was spread over fewer of them, and 50,000 of it was left unabsorbed.
A fixed overhead volume variance is not a cost overrun. It is a statement that the business is carrying capacity it did not use this month. The remedy is sales or a smaller factory, never a cheaper electricity contract.
| Standard to actual | Amount | Owner |
|---|---|---|
| Standard cost of 9,000 brackets | 2,700,000 | |
| Steel price | (38,000) | Purchasing |
| Paint price | 4,600 | Purchasing |
| Steel usage | (25,000) | Production |
| Paint usage | (4,000) | Production |
| Labour rate | (23,500) | HR and production |
| Labour efficiency | (40,000) | Production |
| Variable overhead expenditure | (8,000) | Production |
| Variable overhead efficiency | (12,000) | Production |
| Fixed overhead expenditure | (10,000) | Plant management |
| Fixed overhead volume | (50,000) | Sales and planning |
| Actual cost incurred | 2,905,900 |
The 10 variances sum to 205,900 adverse, and 2,700,000 plus 205,900 is 2,905,900. Check it against the actual table: steel used at standard 925,000, plus the steel price variance 38,000, plus paint 179,400, labour 963,500, variable overhead 290,000 and fixed overhead 510,000. The same 2,905,900. No plug.
Now read it. The month is 205,900 over standard, and the first instinct in the meeting will be the steel price, because steel is what everybody talks about. Steel price is 38,000 of it, less than a fifth. The larger story is on the shop floor: 200 hours more than the output justified, which shows up 3 times, as labour efficiency, as variable overhead efficiency, and, because those hours were paid at a higher rate, inside the rate variance too. And the single largest line, the 50,000 volume variance, is not a factory problem at all. It is 1,000 brackets that were not ordered.
Notice that the steel price variance of 38,000 was calculated on 19,000 kg, but only 18,500 kg went into brackets. The other 500 kg is on the shelf. There are 2 ways to handle this and they give different stock values.
Neither is wrong. What is wrong is doing one in the costing system and the other in the accounts, and discovering the gap at the year end audit. Decide once, write it down, and make sure the stock valuation and the variance report come from the same ledger.
Four things reliably turn a standard costing system from evidence into noise.
The example above is 1 product with 5 inputs. A real bill of materials has 40 lines, some of them subassemblies with their own bills, and a factory makes 300 products. The variance calculation is the same, but 3 things happen at scale.
Niyantrak Consulting sets up standard costing for manufacturing businesses as a framework first, a system second. The framework is the standard setting process, the loss allowances, the decision on where price variance is taken, and the ownership of each variance line. Without that, the software reports numbers nobody has agreed to answer for.
Niyantrak ERP carries the system side: a standard cost per item with its history, a purchase price variance captured on the receipt against the order rate, material issued to production orders so usage is measured per job, work centres with their own rates, and overhead absorbed and settled through the ledger so the stock valuation in the balance sheet is the same figure the costing report starts from. It runs on your own computer, and nothing leaves the building.
A short conversation is usually enough to tell whether the standards, the ownership, or the system underneath them is the problem. Niyantrak Consulting sets up standard costing for manufacturing businesses on exactly this basis.