Home / Journal / Labour minutes per piece

Labour Minutes per Piece: Standard Time in Crochet Production

Updated 19 November 2026 · Standard time, work measurement, capacity arithmetic

A standard minute is not a measurement of how long a piece took. It is a statement of how long the piece should take, made under four conditions that have to be named before the number means anything: a defined method, a defined quality level, a defined operator performance, and a defined allowance for everything that is not work.

Almost every argument about a crochet price eventually arrives at a single line on a cost sheet: labour minutes per piece. It is the line buyers cannot verify, suppliers cannot easily explain, and both sides treat as arithmetic when it is actually a claim. A quote that says 34 minutes is not reporting a fact about the past. It is asserting that a trained artisan, working at a stated pace, using a stated method, with stated allowances, will produce one conforming piece in 34 minutes — and that this will still be true at the thousandth piece, not only at the first.

That assertion can be tested. It can also be built badly, in ways that are invisible for a season and expensive for a programme. This article is about how the number is produced, where each of its inputs comes from, and what a buyer should ask for when the number appears in a quotation without any of them attached.

In this article

1. What one standard minute claims · 2. The clock measures the artisan, not the task · 3. Three ways to obtain a number · 4. Breaking a crochet piece into elements · 5. Allowances: what belongs and what does not · 6. The first hundred are not the thousandth · 7. What predetermined times can and cannot do · 8. From minutes to a schedule · 9. Three ways the line goes wrong · 10. What to ask for in the RFQ · 11. Questions

1. What one standard minute claims

The definition used across industrial engineering is stable enough to quote: standard time is the time required by a qualified, trained operator, working at a defined level of performance, using a specified method and equipment, to complete one unit of work to a specified quality, with allowances for rest, personal needs and unavoidable delay. The arithmetic that follows from it is equally stable, and is the one taught in the ILO work study tradition:

standard time = observed time × rating factor × (1 + allowance)

Four things have to be true before the left-hand side means anything. The method has to be fixed — same yarn, same hook or needle, same assembly sequence, same finishing standard. The quality level has to be fixed — a piece that passes inspection, not a piece that is nearly right. The performance level has to be fixed, which is what the rating factor does. And the allowance has to cover a named list of non-work time rather than a round number added for comfort.

Change any one of them and the number changes, without anything on the shop floor having changed at all. This is why two suppliers can both be honest and still quote eleven minutes apart on the same design: they are not measuring the same object. One has timed a first-off made by the sample room, one has timed a production piece; one has rated the operator at 100 and one at 85; one has added 8 per cent for rest and one has added 25 per cent for rest plus changeover plus rework.

For a buyer, the practical consequence is that a labour line is only auditable if it travels with its conditions. A supplier who can produce the method sheet, the observation sheet, the rating basis and the allowance list is making a claim that can be checked. A supplier who can produce only the total is asking to be believed.

2. The clock measures the artisan, not the task

A stopwatch produces a distribution, not a value. Ten pieces by the same artisan on the same day will not take the same time; ten pieces by ten artisans will spread wider still. The spread is not noise to be averaged away. It is part of the answer, because it tells you what happens on the day the fastest artisan is absent.

The rating factor exists to convert what was observed into what is normal. An operator working noticeably faster or slower than the pace the factory regards as standard produces a time that has to be adjusted before it can be called a standard. The rating system most widely taught, adapted from the Westinghouse scheme and carried in ILO's work study guidance, scores four factors independently: skill, effort, conditions and consistency, each across a set of grades, and the four combine into a single percentage.

Three things about rating are worth knowing even if you never perform one.

First, rating is a judgement, and judgements drift. An analyst who has spent a week watching the same operation begins to treat that speed as normal, which is precisely why a second analyst, or a re-rating after a break, sometimes produces a different answer on unchanged work. The defence is documentation: the grade assigned to each factor and the name of the person who assigned it.

Second, rating cannot rescue a badly chosen subject. If the study was performed on the three quickest hands in the workshop, no rating factor makes the result general. The subject should be a trained operator working at a sustainable pace, and the observation should include enough different operators that the spread is visible.

Third, the honest use of rating is to make times comparable across operations, not to make them look better. A factory that rates every operation at 100 is not measuring; it is asserting that every operation runs at precisely the reference pace, which is never true and which quietly loads all the error into the allowance.

3. Three ways to obtain a number

There are three legitimate routes to a labour minute, and they fail in different ways.

Stopwatch time study

The classic method: break the operation into elements, time each element over a number of cycles, discard obvious outliers and note why they were discarded, rate the performance, add allowances. It is direct and defensible, and it requires the operation to exist. It also requires the analyst to be present without changing what is being observed, which is harder than it sounds in a workshop where the arrival of a clipboard alters the pace for the first twenty minutes.

The number of cycles needed is not fixed. It depends on how much the observations spread: a stable operation may settle after a dozen cycles, an unstable one may need several dozen before the estimate is worth quoting. Reporting "we timed ten pieces" without reporting the spread is a way of hiding the answer to the only question that matters — whether ten was enough.

Video, and frame-by-frame analysis

Video is not a different method so much as a different instrument. It allows the element boundaries to be fixed after the fact, allows a second analyst to check the first, and allows the same footage to be re-examined when a dispute arises months later. For hand work where the elements are short and the hand movements overlap, this is often the only way to get element times that sum correctly, because the eye cannot separate a yarn pick from a stitch completion in real time.

The cost is time: reviewing footage is slower than watching a workshop. The benefit is that the record survives. When a supplier and a buyer disagree about whether the assembly element was included, footage settles it in a way that a spreadsheet cannot.

Back-calculation from production history

If a design has been running for a year, the labour minute is already implicit in the records: pieces shipped, artisan-days booked, hours paid. Divide and you have a number that reflects everything that actually happened, including the rework, the waiting for yarn and the days the pattern had to be re-explained.

This route is cheap and it is usually too pessimistic as a standard, because it captures every inefficiency alongside the work. It is best used as a cross-check: if the stopwatch says 34 minutes and the production records imply 52, one of the two is telling you something important, and it is worth knowing which before the next quotation is issued. A history-based figure also answers a question a stopwatch cannot — what the operation delivers across a season, rather than on a good Tuesday.

4. Breaking a crochet piece into elements

An element is a segment of work with a defined start and end point, chosen so that the boundary can be observed consistently. In crochet the natural boundaries are not always the obvious ones, and getting them wrong is the most common reason a study produces times that do not add up to the piece.

The element groups that usually work for a crocheted toy or accessory are these:

Preparation. Reading the pattern for this piece, gathering the yarn and hook, checking the yarn against the approved lot, setting up the workstation. This is not production time and it is not nothing: on a short run it can exceed the first piece's work, which is one reason small orders carry a different minute per piece than large ones even when the craft is identical.

Body and component work. The crocheting itself, split by component rather than by stitch type — head, limbs, body panel, and so on. Splitting by stitch type produces elements that are easy to time and impossible to schedule against, because no artisan makes all the single crochet for a piece before starting the double.

Assembly. Joining, seaming, attaching limbs, setting eyes and noses, securing any attached part. This is where the safety-critical work sits, and where element boundaries should be drawn tightly enough that the attachment step is separately visible. A standard time that buries the eye-fixing inside "assembly, 9 minutes" cannot be audited against the pull-force requirement, because the number does not say how long the fixing was allowed to take.

Filling and closing. Stuffing to the specified firmness, closing the seam, checking symmetry. Firmness is a quality characteristic and belongs in the element definition, not in the inspector's discretion.

Finishing. Trimming ends, brushing, steaming or blocking if the design requires it, final visual check against the approved sample, tagging.

Handling and packing. Folding, bagging, carton packing, recording. Frequently omitted from the labour minute, which is one of the three errors discussed later.

Two rules make the difference between an element list that works and one that does not. Every element needs a boundary that can be seen — "pick up hook" and "first stitch complete" are boundaries; "start working" is not. And every element needs to be assignable to a workstation: if two elements always happen simultaneously, they are one element for scheduling purposes even if they are two for costing.

5. Allowances: what belongs and what does not

The allowance is where a standard time is most often quietly inflated or quietly starved. Its purpose is narrow: to cover the time a trained operator needs, beyond the work content, to sustain the pace across a shift. It is not a place to hide a badly planned line.

The conventional categories are personal needs, fatigue and unavoidable delay — often abbreviated PFD. ILO's work study guidance puts basic fatigue for light work in good conditions at around 4 per cent, and treats it as a floor to which variable allowances are added for posture, lifting, monotony, noise, heat and the like. Personal needs are typically allowed separately, and delay covers interruptions that are genuinely outside the operator's control: waiting for the next batch, a supervisor's question, a hook that has to be replaced.

What should not be inside the allowance: machine downtime, changeover between designs, rework caused by a specification error, and the learning period on a new design. Each of those is real, each costs money, and each belongs somewhere else in the cost model. Mixing them into a labour allowance makes the figure resistant to improvement — every time the underlying problem is fixed, the allowance stays where it was.

The test is simple. Ask what the allowance covers and get a list. If the answer is a single percentage with no list, the number is a negotiating position rather than an engineering figure, and it will move when the negotiation moves.

There is also a question of direction. A high allowance with a low observed time and a high rating factor is arithmetically identical to a low allowance with a high observed time and a low rating. The total hides the composition, and the composition is what tells you whether the factory is being careful or is covering for something.

6. The first hundred are not the thousandth

Labour time per unit falls as cumulative output rises. This was quantified in aircraft assembly and published by Theodore Wright in 1936 in the Journal of the Aeronautical Sciences, in a paper on the factors affecting the cost of airplanes: each doubling of cumulative production was associated with a roughly constant percentage reduction in direct labour hours, around 20 per cent in the data he examined. The resulting relationship, usually written as a power law with the learning rate as its parameter, became known as the 80 per cent curve.

The formula is worth having in mind because it makes the assumption explicit:

Y = aXb, where Y is the time (or cost) per unit, X is cumulative units, a is the first unit's time, and b = log(learning rate) / log 2. For an 80 per cent rate, b is approximately −0.322.

Two cautions matter more than the formula for hand-made work. The rate is not a law: published estimates across industries run from roughly 75 per cent to 95 per cent, and the flattening at high volumes is well documented — the easy improvements are exhausted first, and a stable process eventually stops learning. And the curve assumes the method is stable. Change the yarn, the hook, the pattern or the finishing standard and the curve restarts, which is why a supplier who re-engineers a design mid-programme and keeps quoting the old minute is quoting something that no longer exists.

For crochet, the effect is real but its shape is different from assembly work. Experienced artisans bring a great deal of transferable skill, so the learning on a new design is shorter than it would be for a novice workforce. What does not transfer is design-specific: where the increases fall, how the piece is held, which seam closes last, how firm the filling should be for this particular shape. That knowledge accrues over the first few hundred pieces and then stabilises.

The practical consequence for a buyer is that a quotation should say which point on the curve it refers to. A first-order price built on first-off times is legitimate, provided the ramp is described. A re-order price built on first-off times is not a conservative price; it is an unexamined one, and it will either be renegotiated or quietly recovered somewhere else.

7. What predetermined times can and cannot do

Predetermined motion time systems assign times to basic motions from a published table rather than from a stopwatch, so a method can be costed before anything is built. The best known, Methods-Time Measurement, was released in 1948 and exists in several variants; its unit, the TMU, is defined as 0.00001 hours, which is 0.0006 minutes or 0.036 seconds. A second is 27.8 TMU.

The appeal is consistency: no rating judgement is required, because the table times already represent a defined performance level, and two analysts applying the same table to the same method should reach the same answer. The limitation is granularity. These systems work by decomposing work into reach, grasp, move, position, release and similar motions — a vocabulary that fits repetitive, short-cycle assembly well and fits hand crochet poorly, because the "motion" in crochet is a compound action whose duration depends on stitch geometry, yarn friction and tension rather than on the distance the hand travels.

Where predetermined times do earn their place in a crochet operation is at the edges of the craft: trimming, tagging, bagging, cartoning, label application. These are short, repetitive, motion-dominated tasks, they are often the least studied part of the piece, and they are frequently the reason a quoted labour minute turns out to be short.

The failure mode to avoid is using a predetermined system as a substitute for observing the craft, then presenting the result with the authority of a table. A table-derived number for hand crochet is an estimate of a method, not a measurement of a practice, and it should be labelled as one.

8. From minutes to a schedule

A standard minute becomes useful only when it is divided into available minutes, and the denominator is where most capacity plans fail. The arithmetic is not complicated:

pieces per period = available minutes × efficiency ÷ standard minutes per piece

Available minutes are not headcount multiplied by hours. They are headcount multiplied by paid hours, reduced by attendance, by the share of the day that is not production at all, and by the fact that not every artisan works on the same order at the same time. Efficiency is not a moral quality; it is the ratio of standard minutes produced to minutes paid, and in a distributed handwork network it varies by site, by season and by design.

Two consequences follow, and both are frequently missed in negotiations.

The first is that capacity is a statement about a period, not a quantity. A supplier who says "we can do 40,000 pieces" has said nothing until the period is named, and a supplier who says "we can do 40,000 a month" has said nothing until the share of that month's capacity already committed is named. The number that matters is the free capacity in the window you need, and it is a smaller number than the headline.

The second is that a change in standard minutes moves the delivery date more than a change in headcount usually can. Adding artisans to a distributed network takes recruitment, training and a period of lower output per head. Removing minutes from a piece — by simplifying an element, by fixing a jig, by moving trimming into a separate station — takes effect on the next batch. This is the practical argument for measuring: the measurement is what tells you which of the two levers is actually available.

It is also the reason capacity conversations are more productive when they are about elements rather than totals. "Your 34 minutes is too high" is an argument. "Your assembly element is 9 minutes and includes eye-fixing; we think that element should be split and the fixing should carry its own standard because it is the safety-critical step" is a conversation that ends with a better number.

9. Three ways the line goes wrong

Across quotations, the same three errors recur.

Timing the wrong artifact. The study is performed on a sample-room first-off, made slowly and carefully by the person who developed the pattern, in a room with better light and no interruptions. The result is a number that describes a favourable condition and is then applied to a production run. The direction of the error is not always the same — a first-off can be slower because the method is still being worked out, or faster because it was made by the most skilled person available — which is why the error is hard to spot from the total alone.

Rating drift. The rating factor is set once, informally, and never revisited. Over a year, the real pace of the workshop changes — new artisans, different mix of designs, a different supervisor — while the factor stays. The standard minute then describes a workforce that no longer exists, and the variance shows up as an unexplained gap between planned and actual output.

The unmeasured tail. The study covers the crocheting and stops. Trimming, checking, tagging, bagging and cartoning are excluded because they are "not really production", and then a schedule built on the standard minute misses by the amount of the tail. In hand-made work the tail is proportionally larger than in machine work, because the finishing of a hand-made piece is itself hand work.

All three have the same signature: the number is not wrong by a random amount, it is wrong in a direction, and the direction is stable. A buyer who sees planned and actual output diverging by a consistent percentage for three months is not looking at variability. They are looking at a standard that was built with one of these three errors in it.

10. What to ask for in the RFQ

The point of all this is not to turn a buyer into a time study analyst. It is to make the labour line answerable with four short questions, each of which a competent supplier can answer without new work.

Which method was timed? The answer should name the yarn, the hook or needle, the assembly sequence and the finishing standard, and should identify the approved sample the method corresponds to.

What was observed, and on what? The number of cycles, the number of operators, and whether the subject was a first-off or a production piece. A supplier who cannot say whether the study was done on a first-off has not done a study that can be relied on.

What does the allowance cover? A list, not a percentage. Personal needs, fatigue, delay, and whether changeover or rework is inside or outside.

Which point on the learning curve does this price assume? First order at first-off times and re-order at steady-state times are both legitimate; quoting one and applying the other is not.

None of these questions requires the supplier to reveal a margin, and none of them is adversarial. They are the questions that make the number a shared object rather than a negotiating position, and they are the difference between a price that can be defended at the first reconciliation and a price that has to be reopened.

Related reading: costing after the quotation, which covers where the labour line sits among the others; yarn count and consumption, for the material line that sits beside it; and production progress reporting, for what to measure once the order is running.

What exactly is a standard minute?

The time a qualified, trained operator needs to complete one unit of work to a specified quality, using a specified method, at a defined performance level, with allowances for rest, personal needs and unavoidable delay. It is calculated as observed time multiplied by a rating factor and then by one plus the allowance — which means it is only as meaningful as the four conditions attached to it.

How many pieces can one artisan make in a day?

There is no general answer, and any supplier who offers one without naming the design, the yarn, the stitch density and the finishing standard is guessing. What can be answered is the shape of the arithmetic: available minutes multiplied by efficiency, divided by the standard minutes for that specific piece. Ask for the standard minute and the assumptions behind it rather than for a daily output figure.

Should I accept a labour figure without seeing the study?

You can accept it as a price and still ask four questions about it: which method was timed, what was observed and on what, what the allowance covers, and which point on the learning curve the price assumes. A supplier who can answer all four is working from a study. A supplier who cannot is working from a target, which is a different thing and behaves differently when the order changes.

What allowance is normal for hand crochet?

ILO's work study guidance puts basic fatigue for light work in good conditions at around 4 per cent, with separate allowances for personal needs and for variable factors such as posture, monotony and environment. The total that results depends on the workstation and the shift pattern, which is why the number should arrive as a list rather than as a single percentage. A total with no list cannot be checked and will not move when the underlying conditions improve.

Does a lower labour minute always mean a lower price?

No. A lower minute with a larger allowance, a longer learning ramp or a bigger minimum order can cost more overall than a higher minute with a short ramp. Labour minutes are one input to a price, and they are the input most easily moved without changing the price — which is precisely why they should be examined alongside the order quantity and the re-order assumptions.

How often should standard times be re-measured?

Whenever the method changes, and periodically even when it does not. A change of yarn, hook, pattern or finishing standard restarts the learning curve and invalidates the old minute. Absent a method change, drift still accumulates through workforce turnover and design mix, so a re-check at each annual re-quote is a reasonable minimum and is cheaper than discovering the drift through a missed delivery.

Why does my quote not match minutes multiplied by an hourly rate?

Because the hourly rate in a quotation is rarely the wage rate. It carries the overhead that has to be recovered per productive hour, the cost of supervision and quality control, and the share of non-productive time that the allowance does not cover. The multiplication is a useful sanity check, not a reconciliation — if the two figures are far apart, the useful next question is which components sit in the rate.

Want the timing sheet we use? Send the design, the yarn and the finishing standard you are working to, and we will return the element breakdown, the observation record and the allowance schedule we would apply to your order.

OEM/ODM process · WhatsApp