On this page
- Quick answer: the four sources of drift
- What grading decides, briefly
- Why the same size stops being the same size
- How to notice drift early
- Manufacturing perspective: what retention actually costs
- Practical example: three reorders, two outcomes
- Common mistakes
- How we hold a standard
- Frequently asked questions
- Conclusion
This article is about the second problem. How grading itself works — blocks, grade rules and why a constant scale fails — is covered here. What follows is why a size that measured correctly in March measures differently in September, and what prevents it.
The reason it matters commercially is that fit drift does not announce itself. It appears as a slightly higher return rate, a few more size complaints, and a slow loss of confidence in your own size chart — none of which is traceable to a single order, and all of which are expensive.
Quick answer: the four sources of drift
| Source | How it shows up | What prevents it |
|---|---|---|
| Pattern re-created instead of retained | A gradual change nobody can point to | Keeping the fit-approved pattern with a version date |
| Fabric substituted or changed | A garment that feels tighter or looser at the same measurement | Re-fitting when the cloth changes |
| Grade rules re-derived | Some sizes drift more than others, usually at the ends of the range | Retaining the grade rules alongside the base pattern |
| Different pattern maker | Consistent but different — the shoulder sits slightly somewhere else | Working from the approved pattern rather than from the garment |
None of the four is carelessness. Each is what happens naturally when a style is re-created rather than repeated, which is why the defence is retention rather than vigilance.
What grading decides, briefly
Grading takes an approved base size and produces the other sizes by applying an increment at each point of measure. The increments differ by point, because bodies do not scale uniformly — a shoulder grows far more slowly than a hip across a real size range. That is what allows a garment to fit at both ends of the range rather than only in the middle.
Two consequences follow, and both matter here. First, a wide size range is a larger piece of work than a narrow one at the same total quantity, because there are more increments to get right. Second, an approved fit in one size does not guarantee a good fit elsewhere, so the ends of the range are where drift shows up first.
Why the same size stops being the same size
Re-creation instead of retention
The most common cause, and the most invisible. A factory that no longer holds the approved pattern will draft a new one from the last garment produced or from the specification. Both are plausible sources and neither carries the fit history, so the new pattern is correct against a document rather than against a body. It will be close, and it will not be identical.
Repeated across three or four reorders, close becomes noticeably different. The garment still measures correctly against the chart it was produced to — a chart that has itself been subtly rewritten each time.
A fabric change without a re-fit
Cloth is part of the fit. A heavier or more fluid fabric on the same pattern behaves differently on the body: it may hang lower, hold a line where the previous cloth collapsed, or recover differently after wearing. The measurements can be identical and the garment can still fit differently. This is why a fabric substitution should be treated as a light development rather than a purchasing decision.
Grade rules re-derived rather than retained
Grade rules are values, and values get re-estimated. If they are not retained with the pattern, each re-derivation is a fresh set of small judgements — and the errors accumulate at the extremes of the range rather than at the base size. This is the reason a range can fit well in the middle and badly at the edges.
A change of hands
Pattern makers move roles. When a style is picked up by someone else, the only way to reproduce the fit is to work from the approved pattern and the approved garment rather than from the specification alone. A specification describes what the garment is; the approved pattern records how it was made to be that.
How to notice drift early
Drift is detectable long before it becomes a returns problem, provided somebody is measuring something other than the finished garment's compliance with its own chart.
- Keep a physical reference. One approved garment per style, sealed and dated. Measuring a current production unit against it takes minutes and catches drift immediately.
- Track return reasons by size. A rising share of fit-related returns concentrated in one size is the clearest early signal, and it usually appears before the complaints do.
- Track size curve shift. If customers start buying one size up from their usual, something has moved, even if nothing has been reported.
- Audit on reorder, not on complaint. Measure the second production run against the first rather than waiting for a problem to be escalated.
Manufacturing perspective: what retention actually costs
Brands often treat retention as an administrative nicety. Inside a factory it is operational, and its absence is visible.
- A retained pattern removes the fitting work. A reorder built from the retained pattern needs no new fit round. One built from a re-draft does, whether or not anyone calls it that.
- A retained swatch removes a conversation. Shade approval becomes a comparison against a physical standard rather than a discussion about a photograph.
- A retained pre-production sample defines inspection. Without it, inspection is against an interpretation made by whoever is on the line that week.
- Retention makes production portable. A brand holding its own approved standard can change supplier without rebuilding a garment from a description.
The cost of all four is a shelf, a drawer and a labelling habit. The cost of losing them is a season of quietly widening tolerance, which nobody notices until the returns data is read.
Practical example: three reorders, two outcomes
| Brand A | Brand B | |
|---|---|---|
| Pattern | Retained, dated, held by the factory and copied to the brand | Re-drafted on each run from the specification |
| Grade rules | Retained with the pattern | Re-derived each time |
| Colour | Approved swatch kept with a date | Approved from a photograph on a phone |
| Fabric on run three | Substituted, with a re-fit run first | Substituted, no re-fit |
| Inspection standard | The pre-production sample from run one | The current chart, re-read each time |
| After three seasons | The size chart still means what it meant | A chart that has been gradually rewritten |
Brand B's third run is not obviously wrong. It measures correctly against the chart it was produced to — a chart that no longer describes the garment customers bought in season one. That is what makes drift so expensive: it is invisible until somebody measures the two garments side by side.
Common mistakes
- Treating the pattern as the factory's property by default. Establish ownership and retention before you need them.
- Changing fabric without expecting a re-fit. Cloth is part of the fit, not a surface on top of it.
- Not dating anything. An undated approved standard cannot demonstrate that it is the same standard.
- Widening the size range late. New sizes need grade rules that may not exist yet, and the ends of a range are where drift appears first.
- Reading only the compliance report. A garment passing inspection against its own chart proves nothing about consistency over time.
- Assuming the second run will match the first. It will, if the standard was retained. That is the whole condition.
How we hold a standard
Luxudress has manufactured womenswear in Guangzhou since 1995. Pattern, cutting, sewing, finishing and packing are under one roof, which is what makes retention practical rather than aspirational — the pattern maker who approved your fit is in the same building as the line that repeats it.
- We retain the fit-approved pattern, dated, against the style.
- We retain the grade rules alongside it, rather than re-deriving them.
- We keep the approved swatch so a reorder is matched rather than chosen.
- We keep the pre-production sample as the inspection standard for every run that follows.
- We tell you when something has moved. A fabric substitution or a change of pattern maker is named, not left to be noticed.
- We re-fit when the cloth changes, because a substitution is a light development rather than a swap.
If you are reordering a style that has drifted, or moving a style to a new supplier, send us the approved garment and the size chart and we will tell you what is needed to re-establish it. OEM manufacturing covers repeat production, and quality control covers how the standard is applied.
Frequently asked questions
What is grading, in one sentence?
Grading is applying a set of per-point increments to an approved base pattern so that the same garment is produced in other sizes. The increments differ at each point because bodies do not scale uniformly, which is why a constant scale factor produces a garment that measures correctly and fits nobody. The mechanics are in how pattern making works.
How do I know if my fit has drifted?
Measure a current production unit against a retained approved garment from the first run. It takes minutes and it is the only check that actually answers the question. Returns data will tell you eventually, but a rising fit-return rate in one size usually appears a season after the drift began.
Does a fabric substitution really change the fit?
It can, at identical measurements. Weight, drape, stretch and recovery all change how a garment behaves on a body — how it hangs, where it sits, how it recovers after wearing. The chart can be met exactly and the garment can still feel different, which is why a substitution should trigger a re-fit rather than a material approval.
Should I own the pattern?
It is worth agreeing explicitly either way. Owning it makes your production portable and makes consistency your factory's obligation rather than its choice. Not owning it is a legitimate arrangement, provided everyone knows that a change of supplier means re-establishing the fit from a garment rather than from a pattern.
Is a wider size range harder to keep consistent?
Yes, because there are more increments to preserve and the extremes of a range amplify small errors. If you are adding sizes, the new grade rules are a development decision rather than an arithmetic extension, and they should be confirmed on a fit rather than assumed.
How often should the standard be re-checked?
At every reorder, cheaply, and once a season properly. The cheap version is a measurement against the retained sample; the thorough version also reviews the returns data by size. Both are worth doing, and neither is worth doing instead of the other.
Conclusion
Grading decides whether a size range fits. Consistency decides whether a size stays the same size across seasons, and it depends almost entirely on retention: the approved pattern, the grade rules, the approved swatch and the pre-production sample, all dated and all kept.
Drift is not usually a mistake. It is what happens when a style is re-created rather than repeated, and it stays invisible until somebody measures two garments side by side. Measuring them is cheap. Not measuring them is what shows up in the returns rate.