A garment labelled the same size fits differently depending on who made it, and the variation is not a manufacturing error. Sizing is a commercial decision rather than a measurement.
There is no binding standard
Voluntary size standards exist, but no rule compels a brand to follow one, and the labels themselves are not defined in law the way weights and volumes are.
Each retailer therefore sets its own size chart, and two garments from different brands can differ substantially in every dimension while carrying identical labels.
The number on the label functions as a brand's internal reference rather than as a description of the body it is meant to fit.
Every brand fits to its own body model
Patterns are developed on a fit model, a person whose measurements represent the customer the brand believes it is serving.
A retailer aimed at younger buyers builds around different proportions from one aimed at an older market, and the resulting garments differ well beyond overall size.
Changing the fit model changes how every garment in a range fits, which is why a customer can suddenly find a familiar brand no longer works for them.
Grading multiplies the differences
A pattern is cut once and then scaled up and down to create the other sizes, using rules that decide how much each measurement grows at each step.
Those grading rules are proprietary, and the assumption that bodies enlarge proportionally is only roughly true, which is why fit tends to degrade at the ends of a range.
The most extreme sizes in a range are often the furthest from the fit model and consequently the least reliable.
Labels drift over time
Sizes have been relabelled downwards over decades, so that a garment carrying a given number today is more generous than one carrying the same number in the past.
The incentive is straightforward: customers respond positively to the number they expect, and a brand that resists the drift appears to run small.
Because each brand drifts at its own pace, garments of different ages and origins accumulate inconsistency that no chart can reconcile.
Manufacturing adds the final variance
A single range is often produced in several factories, and cutting, sewing and fabric behaviour differ enough to produce measurable variation between batches.
Fabrics with stretch conceal some of this, which is part of why they have become so common, since a forgiving material reduces the consequences of imprecise sizing.
The practical result is that measurements published by the seller are more useful than the label, and that returns are treated as an ordinary cost of selling clothing online.