Can you extend a size curve to cover extended sizes?

No, and this is the most common and most expensive error in the exercise. A size curve describes how demand distributes across a run that shares a fit block; extended sizes generally require a different block, with different grade rules and often different proportions rather than uniformly larger ones. Adding positions to the end of an existing curve assumes the new sizes behave like a continuation of the old distribution, which is an assumption about demand made from a fact about numbering.

Which styles should be extended first?

The ones whose fit is least sensitive to grading and whose full-price sell-through is already strong in the core run. Styles with structured tailoring, complex seaming, or a silhouette that depends on precise proportion carry the highest pattern cost and the highest fit risk, so they are the worst place to learn. Extending a brand's strongest simple silhouettes first produces a cleaner read on demand, because a poor result is more likely to mean the demand was not there rather than that the fit was wrong.

How do you plan depth for extended sizes with no sales history?

Bound it rather than forecast it. The usable inputs are the brand's own returns and customer-service data on requests for larger sizes, the size distribution of any wholesale or marketplace partner already selling to that customer, and the minimum viable depth the size run can support per door. Planning from a published population size distribution is the common shortcut and it is a population swap: how many people wear a size is not how many will buy your brand in it.

Should extended sizes be a separate assortment or integrated into the main range?

Commercially the assortment should be integrated — the same styles, the same season, the same price architecture — because a visibly separate range signals a lesser one. Operationally the planning must be separate: its own curve, its own depth logic, its own sell-through read. The frequent mistake is inverting these, running a distinct capsule for the customer while planning it as an appendix to the core buy.

How long before an extended size launch can be judged?

Longer than a normal style read, because the first season measures availability as much as demand. If the launch shipped a short size run, thin depth, or partial door coverage, a weak sell-through is largely a statement about what was offered. A defensible read needs at least one full season at planned depth with the range in stock through the selling window, and the diagnostic question is always whether the sizes that underperformed were actually available for the whole period.

Extending a Range Into Extended Sizes: What Actually Has to Be Planned

Extended sizes are a second fit block, not extra positions on an existing size curve. This guide covers why the curve cannot be stretched, how to choose which styles extend, how to build depth with no history, the fit and pattern cost that sits upstream of the buy, and how to read the first season honestly.

The curve does not stretch

The single most costly assumption in extending a size range is that the existing size curve can be lengthened. It cannot, and the reason is not statistical — it is physical.

A size curve describes how demand distributes across a run that shares a fit block. Every size in it is a graded transformation of the same base pattern, and the curve's shape reflects a population of customers being fitted by one construction. Extended sizes typically require a different block, because bodies at that end of the range differ from the core run in proportion rather than only in scale. Grading a size 12 pattern up four steps produces a garment that is larger everywhere, which is not the same as a garment that fits.

So the demand distribution does not continue past the end of the existing curve. It restarts. Two curves, planned separately, sitting under one style.

Brands that miss this typically discover it through returns rather than through sell-through — the units sell, and then come back — which is the most expensive way to learn it and the slowest to appear in a plan.

The cost is upstream, and it is per style

The buy is not where extended sizing gets expensive. Pattern development, grading, fit sessions and sampling are, and those costs are per style, not per unit.

That inverts the usual assortment logic. Normally, extending depth is cheap and extending breadth is expensive; here, adding a style to the extended programme carries a fixed development cost regardless of how deep you buy it. The consequence is direct:

Extended sizing rewards a narrow, deep assortment far more strongly than the core range does. A brand that extends thirty styles thinly pays thirty development costs and gets thirty shallow reads. A brand that extends eight styles properly pays eight and gets eight readable ones.

This is a breadth versus depth decision with an unusually steep breadth cost, and treating it like the core assortment is what produces programmes that are expensive and inconclusive at the same time.

Choosing which styles extend

Two criteria, in this order:

Fit sensitivity. How much does this silhouette depend on precise proportion? Structured tailoring, complex seaming, anything cut close through the torso — high risk, high pattern cost. Relaxed silhouettes, knitwear, elastic waistbands, anything with forgiving drape — lower risk, lower cost. Start at the forgiving end, not because it is easier, but because it produces a cleaner signal.

Existing full-price strength in the core run. Extend styles that already work. If a style sells well in the core range and poorly in extended sizes, that is informative. If a style sells poorly in both, you have learned nothing about extended sizes at all.

The temptation is to extend the styles a brand wishes represented it — the statement pieces, the seasonal hero. Those are usually the most fit-sensitive and the least readable, and a bad first season on them can kill a programme that was sound.

Sizing the first buy without history

There is no honest forecast here. There are, however, real inputs, and they are better than the shortcut most brands reach for.

Usable:

Your own returns and customer-service data. Requests for larger sizes, and returns coded as "too small," are first-party evidence about your customer. Most brands have this and never aggregate it.
Partner size distribution. If wholesale accounts or a marketplace already sell your product to this customer, their size mix is a real read on your brand rather than the category.
Minimum viable depth. What the size run needs per door to be shoppable at all. This bounds the buy from below, and it usually binds harder than anything else in the first season.

Not usable, despite being the most commonly cited: published population size distributions. How many people in a market wear a given size is a statement about people, not about how many will buy your brand in that size. Substituting one for the other is a population swap — the same class of error as taking a benchmark measured on one group and applying it to another — and it consistently produces over-buys, because the population is always larger than the customer base.

The workable method: bound the buy with minimum viable depth, centre it on your own first-party demand signals, and deliberately under-buy the first season relative to your own optimism. A sell-out is a good problem in a programme whose next decision is whether to widen. An over-buy in a new fit block is inventory with no second channel to go to.

Reading the first season honestly

The diagnostic that matters is availability, not sell-through.

If the launch shipped a short run, thin depth, or partial door coverage, then a low sell-through number is substantially a statement about what was offered, not about what was wanted. Before concluding demand was absent, the questions are:

Were the sizes in stock through the whole selling window, or did the middle of the run sell out in week three and leave only the ends?
Was the range available in every door and channel that carries the core styles, or only some?
Did the styles chosen have a fit problem, evidenced by return rates well above the core run for the same style?

That last one separates the two failure modes that look identical in a sell-through report. Demand absent shows as slow sell-through with normal returns. Fit wrong shows as acceptable sell-through with elevated returns. They call for opposite responses, and a plan that only reads units sold cannot tell them apart.

Hindsight analysis on an extended launch therefore has to carry returns data alongside sell-through, which is not how most seasonal reviews are built.

What good looks like in year two

By the second season the programme should have produced its own curve — built from its own sell-through, at its own depth, in its own fit block. That curve is the asset. It is what turns extended sizing from a bet into a planned part of the range, and it cannot be borrowed, bought, or inferred from the core run.

Everything in the first season should be organised around generating it cleanly.

See how RetailNorthstar plans a second fit block with its own curve and depth logic under a single style.

Book a Demo →

Related resources

Size & Pack Optimization — Building and applying curves in the core run
Size Curve — Glossary — Why a curve is bound to a fit block
Size Curve Allocation Formula — Turning a curve into door-level units
Breadth vs Depth — Glossary — The trade-off with an unusually steep breadth cost here
What Is Assortment Planning — Where the extension decision sits in the wider plan
Hindsight Analysis — Glossary — Reading a launch with returns alongside sell-through
Option Count — Glossary — Why narrow and deep wins in a per-style-cost programme