Quoted SMV Versus Production SMV: Investigating the Cost Variance

The efficiency gain that came from changing a number

A production report shows 60% efficiency. An engineer replaces the quoted garment SMV with a newer approved value, and the report rises to 70%. Output has not changed. Neither have the hours worked. The team has changed the measure of work earned by each garment.

This invented example captures a common risk in SMV variance analysis. A different standard can change the reported result before anyone improves the process. The new value may be more accurate, but the increase alone does not prove better performance.

The useful question is what changed between the quote and production. Was the garment specification revised? Did the original breakdown omit work? Was a different method used? Or did the same method take more paid time than planned? Each answer points to a different action.

A sound report preserves the quoted basis and explains the current standard. It then reconciles both with actual time and cost. It gives commercial, engineering, production and finance staff a common account of the order. It should help them resolve a difference without hiding it inside an average efficiency percentage.

Put three different measures beside each other

The quoted SMV is the standard minute value used when the order was costed. It belongs to a particular product version and assumed method. It may have come from a detailed breakdown or an early estimate. Those are different levels of evidence, even when both appear as a precise decimal.

The production SMV should describe the approved method for the product actually being made. Its record needs the relevant operations, equipment and allowance basis. It is not simply the average elapsed time observed during a difficult first day.

Actual labour time is what the defined group of people used or was paid for during the reporting period. It can include delays, extra handling, repairs and learning. Which items belong in the total depends on the stated measurement boundary.

ApparelDiary’s SMV calculation guide covers the creation of a defensible standard. Here, the task is to explain the difference between standards and their cost effects.

Coats Digital’s GSDCost documentation separates method standards from factory efficiency and wage inputs. Each has a different role in time and cost analysis. It also describes comparing actual and standard values at operation level. That separation is useful whether the factory uses specialist software or a controlled spreadsheet.

Make the comparison fair before calculating it

Two totals are comparable only when they cover the same work. A sewing-only quote covers less work than a production figure that includes finishing and packing. Align their scope before comparing them. Moving a task between departments may change their individual SMVs. The factory’s total work can remain unchanged.

Start with the product version and order quantity. Match the size range, material, construction and quality requirements. Then compare the route. An added reinforcement seam is different from a repeated seam caused by a defect. The first can change required work content; the second is repair work that needs its own explanation.

Check units as well. A time entered per dozen must be divided by twelve before it is compared with minutes per piece. A set containing a top and shorts is not the same output unit as one garment. These errors can survive several rounds of approval if the column heading says only “SMV.”

The allowance convention must also match. One value may include an allowance that the other leaves out. The percentages may also use different bases. Reconcile that basis before labelling the remaining gap a method difference. Do not add a further allowance merely because the production result is lower than hoped.

Microsoft’s variance guidance identifies several possible sources. These include route versions, added operations and reported time. The practical lesson is to inspect those records before assigning a cause to the total.

Build an operation-level bridge from quote to approved work

Assume an order will deliver 10,000 accepted garments. Its quote used 18.0 standard minutes per garment. Before production, the team approves a 21.0-minute standard for the final product and method. All numbers in this article are hypothetical teaching examples.

The engineer does not replace 18.0 with 21.0 and discard the earlier file. The report keeps both and explains the three-minute change.

These are differences between completed standard values on the same allowance basis. They are not raw stopwatch readings to which another percentage should be added. The table is an example of reconciliation, not a standard-time library.

The construction change may support a commercial discussion if the agreement permits it. The omitted operation points to a quoting or review error. The handling change needs a closer look. Was the quote based on equipment the factory never had? Or did production choose a less suitable method? The numerical bridge does not settle that question.

For 10,000 garments, the change adds 10,000 × (21.0 − 18.0) = 30,000 standard minutes. That is 16.7% more standard work than the quote contained. It is not yet 30,000 extra paid minutes, because the costing plan also contains an efficiency assumption.

Recalculate efficiency without claiming a process improvement

Suppose the order uses 300,000 paid labour minutes within the agreed reporting boundary. That equals 5,000 labour hours. The plan and actual results cover the same labour group. They also include the same time categories.

The report uses this defined measure:

Efficiency = earned standard minutes ÷ included paid labour minutes × 100.

Using the quoted SMV gives:

10,000 × 18.0 ÷ 300,000 × 100 = 60%.

Using the approved production SMV gives:

10,000 × 21.0 ÷ 300,000 × 100 = 70%.

The ten-percentage-point difference comes entirely from the standard used in the numerator. Both results describe the same output and time record. The 70% figure fits this example’s current method. It does not show a ten-point operational gain.

This distinction matters when an SMV revision is introduced during a reporting cycle. A trend line can jump on the approval date even if the shop floor runs exactly as before. Mark that change in the report and retain a comparison on a consistent basis where it is useful.

The denominator deserves equal care. ILOSTAT distinguishes hours paid for from hours actually worked. Paid hours can include paid absence; the concepts serve different purposes. Name the chosen basis on the factory report. Different time reports may measure different things.

In this example, paid minutes are allocated to the order on a defined basis. A company using attended line minutes would need matching planning assumptions and rates. If the planning basis includes waiting, keep that basis in the actual comparison. Removing it only from actual time would distort the result.

Reconcile the labour cost in three steps

The quote assumed 75% efficiency and a labour cost of 0.100 currency units per included paid minute. The actual average rate is 0.105 on the same cost basis. These are invented rates, not wage benchmarks or recommendations.

The example covers a selected labour cost pool. It excludes fabric, trims, other overhead, selling costs and profit. A complete CM price requires more information, as explained in ApparelDiary’s garment costing guide.

At the quoted standard, planned paid minutes were:

10,000 × 18.0 ÷ 0.75 = 240,000 minutes.

At the approved production standard, while retaining the planned efficiency:

10,000 × 21.0 ÷ 0.75 = 280,000 minutes.

Actual paid minutes were 300,000. The report can now move from the quoted cost to the actual cost without mixing changes together.

The full gap is 4,000 + 2,000 + 1,500 = 7,500 currency units. It also equals 31,500 − 24,000, so the bridge balances.

The first part reflects the changed standard at the planned efficiency and rate. The second reflects 20,000 more paid minutes than the revised plan allowed. The third reflects the higher rate on actual minutes. No residual amount remains unexplained by the arithmetic.

The causes still require evidence. The 2,000 time effect may involve planning, material supply, machine stops, method learning or execution. Calling it an operational variance does not establish that an operator caused it.

Why the calculation order matters

This bridge changes the standard first, then minutes, then rate. This sequence applies the rate change to actual minutes. The rate step therefore includes the combined effect of changed time and rate. Another valid sequence can allocate that interaction differently while reaching the same total.

Use one stated sequence across reporting periods. Otherwise, apparent changes in cause may come from a changed formula. For more complex orders, show the intermediate totals so another reviewer can reproduce the result.

ACCA’s guidance explains planning and operational variances. It compares an original plan, a revised basis and actual results. Its worked labour-variance example also separates time and rate effects. The garment example here uses its own figures and an explicit paid-minute efficiency assumption.

Investigate the time gap without absorbing it into SMV

The revised plan allowed 280,000 paid minutes. Actual time was 300,000, leaving 20,000 minutes to investigate. It would be easy to raise the standard again until the report appeared to meet 75%. That would remove the signal without explaining the work.

Instead, compare operation records, time bookings and events for the period. Look for repeated repair work, missing materials, setup delays or an unplanned method. Separate observed facts from possible causes. A supervisor’s note that “the team was slow” is not the same evidence as a timed sequence with a defined method.

An operator learning a new task may need more time. That does not necessarily change the approved work content. A machine fault may interrupt the method without changing its standard. By contrast, a permanent approved change in handling can justify a new method study. These cases should not all produce the same master-data change.

Event minutes must also reconcile. If twelve people wait for ten minutes, that can represent 120 labour minutes within the affected group. A ten-minute machine stop is not automatically ten labour minutes. It also does not prove that all twelve people were idle throughout.

Use records that identify who was affected and for how long. If two reasons overlap, count each person-minute once in the total. Cause codes can retain both observations. Their total must not create more lost time than the order consumed.

Not every minute will be precisely attributed. An honest unexplained balance is more useful than invented event detail. The next run can collect better evidence at the operations where the gap is largest.

Keep payroll-rate changes out of the work-content argument

The rate step in the example adds 1,500 currency units. The mix of labour grades or overtime premiums may have changed. Another included pay component may also explain it. It does not mean that the garment contains more standard work.

Finance should compare the cost pool behind both rates. A quote may include only direct wages. If the actual rate adds employer contributions and indirect departments, its cost pool is wider. The difference in coverage must be shown before the remaining rate effect is interpreted.

An average rate can also move when work shifts between groups with different rates. Where that matters, calculate the bridge by labour group, then sum it. The total still reconciles. The detail shows rate changes within each group and shifts in the mix of hours between groups.

The example’s rate variance is a cost explanation. It is not a reason to withhold earned pay or reduce required rest. Any action affecting pay or working time must follow applicable rules and employment terms. A variance formula does not establish those obligations.

Handle revisions during an order by their effective dates

A new standard may apply to only part of an order. Suppose a separate example has 4,000 accepted pieces made under an approved 20-minute method. The remaining 6,000 use a revised 21-minute method.

Earned standard minutes are then:

4,000 × 20 + 6,000 × 21 = 206,000.

Applying 21 minutes to all 10,000 pieces would produce 210,000. That adds 4,000 minutes not earned on the stated revision basis. With 300,000 included paid minutes, the revision-based result is about 68.7%, rather than 70%.

Record which units or production dates belong to each method. If the method changes partway through a bundle, record that split. A convenient month-end date may not reflect the work actually done. Retain the approval and the reason for the change.

Historical restatement may be useful when correcting an error, but label it clearly. Keep the originally issued report as well as the corrected comparison. A genuine correction should make the history clearer, not erase the decisions made from the earlier information.

Separate product mix from a change in performance

A factory average SMV can rise because it made more complex products. Consider two styles with unchanged standards of 18 and 30 minutes. Producing 800 of the first and 200 of the second gives 20,400 earned minutes. Reversing the quantities gives 27,600, although total output remains 1,000 pieces.

For a fair order review, apply each style’s relevant standard to its actual accepted quantity. Compare that total with matching labour minutes. A simple average of style percentages can give equal weight to a small run and a large one.

The report can still show a factory total. Build it from the underlying minutes and costs. Where the mix changes sharply, explain it before interpreting a movement in cost per piece.

Reconcile the reporting period and the software baseline

Output and time often cross reporting boundaries. A line may spend time today on work that becomes accepted output tomorrow. A daily comparison can therefore look poor even when the order-level total is reasonable. The opposite can occur when earlier work is completed during a short final period.

Use a consistent cut-off and explain material opening or closing work in progress. A live order may use earned work by stage. That method needs controls to prevent the same work from being counted twice. Do not mix that measure with a finished-garment measure without a reconciliation.

Repairs need the same discipline. A repaired piece that becomes acceptable should not earn the full garment standard twice. A report may credit repair activity for a defined purpose. Keep that credit separate from efficiency based on accepted order output.

The software report may also use a different baseline from the quote. Microsoft’s cost-analysis guidance defines the basis for its completed-order variances. It compares reported activity with the item’s calculated standard cost. The production order’s estimate is a different baseline. A system variance can therefore be correct and still answer a different question from the commercial review.

Document the baseline used by each report. Check the quotation, production estimate and standard-cost record against actual transactions. Their variance totals may differ until those bases are reconciled.

Use the same evidence to forecast the unfinished work

A variance report need not wait until the last garment is packed. Consider an earlier checkpoint in this hypothetical order. The first 4,000 accepted garments have used 120,000 paid minutes. At the actual rate of 0.105, the labour cost to date is 12,600 currency units. The remaining quantity is 6,000, all planned on the approved 21-minute standard.

If the remaining work reaches the quoted 75% efficiency, it requires 6,000 × 21 ÷ 0.75 = 168,000 paid minutes. At the same actual rate, that adds 17,640 currency units. The forecast total becomes 12,600 + 17,640 = 30,240.

If performance stays at the 70% seen so far, the remaining work needs 180,000 paid minutes. Its cost becomes 18,900, giving a total forecast of 31,500. The difference between those forecasts is 1,260 currency units.

The lower forecast needs a reason. Has a measured method change removed the delay? Has the team completed the learning stage, or has a missing resource arrived? A promise to “reach the target” is weaker evidence than a sustained run under the relevant conditions.

Both forecasts assume unchanged scope, rate and quality requirements for the remaining quantity. A new product revision or labour-rate change needs a separate update. Costs already incurred remain in the forecast. Apply the hoped-for rate of progress only to the unfinished work.

This view gives management a range tied to observable conditions. It also shows why timely action matters. An uncertain recovery remains a forecast, rather than a realised saving.

Turn the review into a better next estimate

The final review should identify a change, its evidence and the decision it supports. An omitted operation needs a correction to the estimating process. A changed specification needs a controlled commercial record. A repeated delay may require a production action. A rate difference belongs with the relevant cost owner.

Keep the source records behind the review, not just the final spreadsheet. Link each standard to its method version. Link each time total to the reporting cut-off as well. A disputed entry needs a named reviewer and an unresolved status. Approval of the arithmetic has a limited scope. A revised selling price, new method or payroll adjustment needs its own approval. Those decisions belong to the relevant owners and should remain traceable.

For the main example, the useful conclusion is more precise than “efficiency was low.” The final method contained three more standard minutes per garment than the quote. Actual paid time exceeded the revised plan by 20,000 minutes. The actual rate was 0.005 currency units above the planned rate. Those effects explain the 7,500-unit cost gap on the selected labour basis.

Close each action with evidence from the next relevant run. A revised estimate should use the approved method and a supported performance assumption. It should also retain enough detail to explain any future difference. The value of the report lies in that traceable learning, not in making two percentages agree.

References

1. Coats Digital. GSDCost method standards, cost inputs and operation-level feedback.

2. Microsoft Learn. Common sources of production variances.

3. ILOSTAT. Wages and working-time concepts and definitions.

4. ACCA. Planning and operational variance interpretation.

5. ACCA. Performance Management, December 2012 answers: labour variances.

6. Microsoft Learn. Production-order cost analysis and its comparison basis.