The appeal of a capsule is that it is a short range. That is also why most of them fail: the style count comes down, but every piece is still developed on its own — five or more fabrics, unrelated colours, nothing that pairs. The buyer opens the box and finds twelve styles that make twelve outfits. A real capsule is valued for its combinations, not its style count.
1. Count the looks first, then decide the styles
Write the structure as arithmetic and the style count stops being a guess:
- Tops × bottoms: 4 longline tops × 4 bottoms (2 trousers + 2 skirts) = 16 looks;
- Dresses worn alone: 3 = 3 looks (and the knit dress layers over the wide-leg trouser for 4 more);
- Layering: one duster over those 19 = 19 more.
That is roughly 38 looks — the number a buyer works out on the spot at a appointment, and the reason this structure holds up. The reverse also holds: if four of the twelve are singles that pair with nothing, real output drops into the twenties and the capsule has no reason to exist. So the first rule of the ratio is: tops and bottoms in roughly equal number, dresses no more than a quarter of the total, and one outer layer, not three.
| Category | Styles | Internal mix | Job in the capsule | Usual mistake |
|---|---|---|---|---|
| Longline tops (tunic, long shirt) | 4 | 2 knit (jersey, rib) + 2 woven (poplin, crepe) | The base layer of the season; decides whether the bottoms ever get worn | All four in one fabric — the range reads as uniform |
| Bottoms | 4 | 2 wide-leg trousers + 2 skirts (A-line, straight) | Pairs 1:1 with the tops and carries the 16-look base | Four in the same colour family — the look count halves |
| Dresses | 3 | Tiered maxi, shirt dress, knit maxi | The easy purchase: one piece is a whole outfit, usually the best converter | Three versions of one silhouette — same dress in three colours |
| Outer layer | 1 | Duster or long cardigan | Stretches the season, adds a level, turns singles into sets | Built as a heavy coat — the season gets locked |
2. The colour skeleton: three tiers, two colours per style at most
Capsule colour is not "pick a few nice shades" — it is making any two pieces sit together without a fight. Workable proportions are roughly six parts neutral, three parts seasonal colour, one part accent. These are planning proportions, not industry statistics: a conservative customer pushes the first two tiers up, a younger one raises the seasonal share.
| Tier | Share | Placement | Constraint |
|---|---|---|---|
| Neutral base | about 60% | Two or three of black, charcoal, sand, camel, ivory | At least one must appear in both tops and bottoms, or the base does not pair |
| Seasonal colour | about 30% | Dresses + one top + one bottom | Must cross two categories; appearing only on dresses makes it a single-piece colour |
| Accent | about 10% | One style only (usually a dress or a skirt) | More than one breaks the season's coherence and cannot carry a dye minimum |
Two hard rules: two colours per style maximum — a third rarely sells enough to justify another dye lot — and each colour used on at least three styles. Scatter the colours and every shade's volume shrinks, at which point dye minimums and lot-to-lot variation eat the margin. This is exactly where small orders in many colourways go wrong.
3. Fabric reuse: five or six fabrics for twelve styles
Twelve styles in twelve fabrics is waste. The more scattered the fabrics, the smaller each consumption, the harder the minimum, the more dye lots, the slower the reorder. The consolidation that works is one fabric carrying two or three styles: poplin for both the shirt dress and the long shirt, crepe for both the wide-leg trouser and a dress, rib for a top and for body panels. Land on five or six and purchasing, dyeing and inspection each become a single action for the season.
One boundary cannot be crossed while consolidating: knit and woven blocks are not interchangeable. At the same ease, knit recovery makes a garment read closer to the body while a woven holds its shape; carry knit ease straight onto a woven and the finished piece looks empty. That is set out in coverage and ease for EU/US modest; at capsule stage the only requirement is that styles within one fabric share one set of block parameters and do not borrow across fabrics.
4. SKUs and the first order: 132 SKUs is not 132 minimums
Run the numbers and the first order stops being a feeling:
- SKU count: 12 styles × about 2.2 colours × 5 sizes ≈ 132 SKUs;
- Minimums are per style: 12 × about 50 pieces = about 600 pieces, regardless of SKU count;
- How to split it: rotate three colours rather than giving every style two; run five sizes but weight M and L above the extremes, adding tails as needed;
- Leave a reorder margin: hold fabric for one or two styles inside the 600 so a winner does not wait for new cloth.
The mental shift is that SKUs are a selling unit, not a production unit. Split a style into five sizes and two colours and production still runs one marker, one cut, one sewing line. What drives the first order's value is the style count and the pieces per style. Do the opposite — five colours on every style so the SKU sheet looks rich — and the factory faces five small dye lots, with cost and lead time both going out of control.
5. Sampling order: three skeleton styles first
Sampling all twelve together is the classic waste: the block is not settled, so the other nine all get reworked. The order that works is two rounds:
- Round one — three skeleton styles: one knit top, one woven bottom, one dress. Those three cover the season's three basic constructions (knit upper, woven lower, one-piece) and settle drape, hand-feel and the colour relationships in one pass. First samples in about 3–5 days.
- Round two — the other nine: released together once the skeleton is approved, sharing the confirmed fabrics, trims and block parameters so one fitting closes all of them instead of nine separate loops.
The difference is not only time. A problem caught in round one — a neckline that opens, a hem that hangs wrong, a colour that shifts — gets copied nine times if you discover it with nine samples. Which is why a capsule needs more sampling discipline than a loose range, not less.
6. Seven things to confirm before ordering
- The ratio — tops and bottoms near 1:1, dresses within a quarter of the total;
- Pairing output — pull two random pieces: do they make an outfit? If not, the structure is wrong;
- Colour tiers — does a neutral appear in two categories, does the seasonal colour cross tops and dresses;
- Colours per style — two maximum, and each colour on at least three styles;
- Fabric count — consolidated to five or six, each carrying two or three styles;
- SKUs and the first order — total SKUs, total pieces and the reorder margin counted together;
- Sampling order — three skeleton styles first, the other nine released as one batch.
A capsule succeeds or fails on whether the structure is written as a spec: the ratio as numbers, the palette as proportions, the fabrics as a consolidated list, sampling as two rounds. Settle those four and twelve styles are a season. Skip them and twelve styles are just twelve garments.
Frequently asked
- How do I plan a modest capsule collection — where should I start?
- Start with the structure, not the styles. Fix the ratio first: 4 longline tops + 4 bottoms + 3 dresses + 1 duster = 12 styles, tops and bottoms near 1:1 and dresses within a quarter of the total. That gives 16 looks from tops and bottoms alone, plus three dresses worn alone, plus one duster layered over all of them — roughly 38 looks for the season. Then set the palette (about six parts neutral, three seasonal, one accent) and consolidate fabrics to five or six. Draw the styles last. A 12-style first order at about 50 pieces per style lands near 600 pieces.
- Is 12 styles enough for a season, and how should the ratio split?
- Twelve is enough; the ratio matters more than the count. A workable split is 4 tops, 4 bottoms (2 trousers, 2 skirts), 3 dresses and 1 outer layer. Equal tops and bottoms is what guarantees the pairing count; keeping dresses under a quarter stops the season turning into a dress brand; one outer layer is enough because its job is layering, not volume. If four of the twelve are singles that pair with nothing, real output falls from about 38 looks into the twenties and the capsule loses its point.
- How many pieces should a capsule's first order be — do too many SKUs make it unaffordable?
- The first order is counted per style, not per SKU. Twelve styles at about 50 pieces each is roughly 600 pieces; twelve styles × about 2.2 colours × 5 sizes is about 132 SKUs, but SKUs are a selling unit — production still runs one marker, one cut, one line. What drives the order value is style count and pieces per style. To lower it, cut styles or colourways; do not give every style five colours to make the SKU sheet look full, because that becomes five small dye lots and both cost and lead time run away.
- How many colours per style should a capsule carry, and how do I keep the palette coherent?
- Two colours per style maximum, and each colour on at least three styles. Build the palette in three tiers: about 60% neutral (two or three of black, charcoal, sand, camel, ivory), about 30% seasonal colour, about 10% accent. Two hard constraints: at least one neutral must appear in both tops and bottoms, or the base will not pair; and the seasonal colour must cross two categories — a dress plus a top — since appearing only on dresses makes it a single-piece colour. Scatter the shades and each one's volume drops until dye minimums and lot variation eat the margin.
- Can a capsule share fabric with the regular range, and how many fabrics should I hold?
- Yes, and it should. Consolidate to five or six fabrics, each carrying two or three styles — poplin for the shirt dress and the long shirt, crepe for the wide-leg trouser and a dress — so purchasing, dyeing and inspection are each one action and reorders move faster. The one boundary is that knit and woven blocks are not interchangeable: at equal ease, knit recovery reads closer to the body while a woven holds shape, so carrying knit parameters onto a woven leaves the garment looking empty. At Liuyu the capsule runs on the same development line as casual tops, dresses and trousers, sharing base blocks and the fabric library, with first samples in 3–5 days and bulk in 10–20 days.