Extended sizing was never a range. It was a pilot.

Nobody sends a press release when a size range shrinks. That is the whole tell. The arrival came with a campaign, a casting call, a run of interviews about listening to customers; the departure comes as a filter that returns fewer results than it did last spring, and a category page that has stopped being linked from the navigation.

It would be satisfying to call this hypocrisy. It is duller and more useful than that. Extended sizing, at most large chains, was run as a pilot from the beginning — scoped like a pilot, funded like a pilot, and reviewed against the metrics a pilot is reviewed against. The only people who thought it was a permanent change of direction were the customers, who had no particular reason to know how retail planning works.

What a size actually costs a company

A size is not a scaled photocopy of the next one down. Grading a pattern well means deciding, garment by garment, where the extra centimetres go: how much of it into the body, how much into the armhole, whether the shoulder point moves, what happens to the distance between bust and waist, how a set-in sleeve behaves when the scye grows. Done properly it needs a skilled pattern cutter, several fittings on a body of the relevant size, and time.

Done improperly it costs almost nothing, which is why so much of the market does it improperly. Scale the whole block by a percentage, ship it, and the result is the garment everyone recognises: the sleeve that binds at the top of the arm, the waist that lands too high, the shoulder seam halfway to the elbow, the shirt that gapes for reasons the wearer is invited to take personally.

Then each size becomes a stock-keeping unit, and every unit is a forecast that can be wrong in two directions. Too little and you have lost the sale. Too much and it goes to markdown, which does not merely reduce the margin on those pieces but writes down the whole style. A range with eight sizes has eight forecasts to get wrong. A range with fourteen has fourteen.

Shelf space is the real constraint, and it is zero-sum

In a shop, a rail metre allocated to one thing is not allocated to another. The buyer defending the extended range at the seasonal review is arguing against a colleague with a spreadsheet showing what the same metre returned in the core sizes. The extended range loses that argument most of the time for a reason that has nothing to do with anyone’s politics: it was given less prominent space, less depth, fewer of the trend pieces and more of the safe ones, and it duly sold at the rate that space and that assortment predict.

This is the part that gets misread as evidence. The sell-through figure is treated as a verdict on demand when it is substantially a verdict on the decisions that preceded it. Put the range online only, keep it out of the window, give it the cardigans rather than the coat everyone is photographing, and the numbers will confirm what the placement already assumed.

Why the launch is always loud and the exit is always quiet

A launch buys attention, and attention is measurable in a way the buying team can show upwards. An exit buys nothing and costs goodwill, so it is done by attrition: fewer styles offered in the full range this season, then the range offered only in the continuity lines, then the filter quietly returning basics and outerwear and nothing anyone would call a wardrobe.

By the time a customer notices, there is no announcement to be angry about. There is only the experience of looking for something and finding less of it, repeatedly, which most people file as bad luck rather than as a decision somebody made in a meeting.

The brands that did not do this have something in common

The ones that have held their ranges tend to have built the larger sizes into the block from the start rather than extending outwards from a sample size, which means the grading problem was solved once, at the beginning, rather than deferred to each season’s cheapest option. They also tend to be businesses where that customer is the business, not an adjacent market being tested. Nobody reviews the core range at the seasonal meeting.

That is not a moral distinction either. It is a structural one, and it is the thing worth watching when a chain announces that it is expanding again — which several will, because the cycle runs on a roughly five-year clock. The question to ask is not how many sizes. It is whether the sizes were graded or scaled, whether they are going into shops or into a warehouse, and whether anyone senior has a target attached to them that survives a bad quarter.

Until then, the practical response is the unromantic one. Check what a shop physically holds before you go, buy the thing that fits twice, and treat any range you like as temporary — because on the evidence of the last two years, it is.