What Does Consistency Cost Compared With Accepting Drift?

The question buyers actually ask is whether consistency is worth the money. It is a reasonable question and it deserves a real answer rather than an assurance. The answer is that some controls pay for themselves several times over and some cost more than they return, and the difference is visible if both columns are written down.

What makes the comparison hard is that the two columns are paid by different people. The cost of the control lands in the unit price and is paid by the buyer. The cost of the drift lands in rework, markdown, and a smaller next order, and is paid by the buyer too, but later and less visibly.

One thing that makes this comparison easier: separate the cost of the control from the cost of the failure it prevents. They are different columns and they are paid at different times, and mixing them is why the same argument is had every season.

The comparison also has a time dimension that catches people out. The cost of a control is paid now and is certain; the cost of drift is paid later and is probabilistic. Certainty feels larger than probability even when the expected figures say the opposite, which is why the arithmetic has to be written down rather than felt.

What Does Consistency Actually Cost?

Four things. Calendar time, because a comparison and a lab dip take days. Material cost, because a dedicated purchase or a reserved dye lot costs more than buying what is available. Sorting cost, because a tighter band rejects more units. And administrative cost, because records take time to keep.

The fourth is the one programmes underestimate and the one that determines whether the others survive. A control that requires an hour per run is a control that gets skipped in the second year, and the cost of skipping it is not visible until a run fails.

There is also a second-order effect on supply. A band the plant cannot hold is a band the plant will argue about, and a supplier who expects to argue builds the argument into the price rather than into the process.

Where Does the Cost of a Tighter Band Land?

Mostly in sorting. A band set at the edge of what the process can hold will reject a percentage of every run, and that percentage is either reworked, sold through another channel, or absorbed into the unit figure by the supplier before quoting.

There is a sixth cost that belongs on this list because it is the one a finance team will recognise: the internal time spent deciding what to do. A rejected reorder consumes days of attention across buying, quality and sales, and that time is real even though no invoice ever arrives for it.

Two of those five are invisible in a claim and visible only in a forecast. The markdown and the returns appear in the accounts; the smaller next order appears only as a booking that did not happen, which is why it is so often left out of the comparison entirely.

What Does Accepting Drift Cost Instead?

Five things, in ascending order of visibility: rework and re-inspection, air freight to rescue a season, markdown on a shade that does not match the rest of the range, returns from customers, and a smaller order next time because the buyer cannot predict what will arrive.

The last one is the largest and the least often counted. A buyer who has been surprised twice reduces the size of the third order, and that reduction is a cost of drift that never appears in any claim.

Two practical cautions about the comparison. Count only the drift events that actually happened rather than the ones that might, and count the internal time at the rate the business pays for it rather than at zero, because zero is what makes every control look expensive and every drift look free.

How Are the Two Compared Honestly?

Over a year rather than per run. A control that costs a small amount per run looks expensive against a single order and modest against four. A drift that costs a great deal once looks enormous against one order and modest against four - unless the fourth order was smaller because of it.

Control cost per run

Calendar, material, sorting and administrative time, per run

Runs per year

The multiplier that makes the comparison honest

Drift events per year

How often a property has actually moved outside a band

Cost per drift event

Rework, freight, markdown, returns and internal time

Lost order effect

The reduction in the next order after a surprise

Decision

Which controls stay, which are dropped, and the review date

The reason the clerical controls do so well is that they prevent decisions rather than defects. A register line does not stop a bag being made badly; it stops someone choosing a different component without saying so, which is the far more common failure on a repeat.

It is worth stating what pays for itself means here: the control costs less than the failures it prevents over a year. It does not mean the control costs nothing, and it does not mean every programme needs every control, because a programme running one style once a year faces different risks from one running six styles every quarter.

Which Controls Pay for Themselves?

The clerical ones, decisively. A retained reference, a first-piece check, a component register and a tolerance table cost hours, not money, and between them they prevent most of what goes wrong on a repeat. No programme should be without them, whatever the unit figure.

A lab dip per lot is the first control that actually costs money, and it earns its place on almost every programme, because shade is the most common cause of a rejected reorder and a dip is far less expensive than a rejection.

A third item belongs on this list: inspecting every unit. Full inspection sounds thorough and is usually slower, more variable and less reliable than a plan, because attention degrades long before the last carton.

Which Controls Cost More Than They Return?

Running a full laboratory test set on every lot. Qualification testing belongs at approval and at the annual review; running it per lot costs a great deal and rarely changes a decision, because the properties it covers do not move run to run.

Also, tightening every band at once. A programme that halves all its tolerances in one step will reject material it could have sold, pay for sorting, and discover that half the bands were never the ones that mattered. Tighten the two that move, on evidence.

Write the trade-off where the next reader will find it, which is on the tolerance table rather than in a separate note. A band with a cost and a review date attached gets reviewed; one in an email from two seasons ago does not.

How Is a Trade-Off Written Down?

As a decision with a number and a date: this band is set here, it costs this, it is reviewed then. Three fields on the tolerance table, and they are what stop a trade-off from being re-litigated every season by whoever is new to the account.

A third option sits between the two and is worth knowing about: reserving a dye lot rather than a whole material purchase. Where the substrate is standard and only the shade is critical, reserving the dyeing gives most of the benefit at a fraction of the cost.

What Does a Dedicated Material Purchase Change?

It removes batch variation as a source of drift, at a higher material cost. For a programme running a single colourway in volume it is often the right decision. For one running six colourways in small quantities it can be the most expensive control on the list and the least effective, because the variation is spread across more lots anyway.

It also affects the inspection arithmetic. A smaller lot takes a smaller sample under ISO 2859-1, so the same acceptance quality limit is judged on less evidence. A buyer who reduces a quantity and cares about a property should say so rather than accept a smaller sample by default.

How Does Quantity Affect the Trade-Off?

Larger quantities spread the fixed cost of a control across more units, which is why the clerical controls look almost free at volume and material at 500 pcs per style and colourway. Smaller quantities raise the per-unit cost of everything, including the records.

That is worth saying plainly: on a repeat of 300 pcs, discussed below the standard minimum, the administrative controls cost proportionally more. The answer is not to drop them but to keep them short - one sheet rather than four.

There is one more item worth adding for programmes that run several colourways: a retained reference per colourway rather than per style. Colour decisions do not transfer between colourways, and a single reference for four of them is how a shade dispute starts.

Two things make the recommendation easier to act on. Do the four clerical controls in one afternoon rather than over a season, and review the whole list once a year against what actually went wrong, because the controls that matter change as a programme matures.

What Is the Honest Recommendation?

Buy the four clerical controls without discussion, add a lab dip per lot where shade matters, tighten the two bands that have actually moved, and run the annual test. That is the whole list for most programmes and it is a fraction of the cost of one rejected shipment.

One caution about the arithmetic: count the cost of drift once, honestly, rather than every time it is convenient. A figure that grows with each telling stops being evidence and starts being a negotiating position, and it will eventually be treated as one.

One note on who should hold the sheet. The tolerance table with costs attached belongs to the buyer rather than to the production side, because it records what the brand is willing to pay for and the production side cannot be the judge of that. It is a single page, it is reviewed once a year, and it is the shortest document in the file.

Frequently Asked Questions

People Also Ask