Yinghang Wu | Founder, ChinaBrandPath

An apparel importer can approve a garment, place a modest first order and still learn very little about the assortment. The first lot may sell, yet the result can be carried by one colour, one size curve, one marketplace promotion or one retailer that ordered more deeply than the rest. Expanding every colourway after that result turns a limited test into a broad inventory commitment before the channel has explained what actually worked.

For a Russian importer evaluating a new apparel supplier, the first lot should answer a narrower question: which sellable combinations have earned a repeat order, which need revision, and which should stop?

Define the pilot at the sellable-unit level

A factory style is not the same object that a marketplace, store or distributor receives. The channel sells a specific colour, size, pack and label version. If the pilot groups all of those combinations under one style code, the importer may see an acceptable overall result while missing weak combinations hidden inside it.

Before ordering, the importer can create a simple assortment map. Each row identifies the sellable combination: colour, size range, pack quantity, channel and the version of the garment and packaging. The purpose is not to create a complex data project. It is to preserve the unit on which the next decision will be made.

This matters when the same garment enters more than one route. A marketplace may sell individual units, an independent retailer may buy a mixed inner pack, and a regional distributor may request a different size curve. Treating these as one pilot prevents the importer from seeing whether the economics are repeatable in each channel.

Separate demand from availability

The strongest-selling combination is not automatically the one with the strongest demand. It may simply have been available for more days, placed in more stores or supported by a promotion.

The pilot record should therefore keep a few context fields beside sales: days available, stock-outs, markdown or voucher periods, cancellations, return reasons and channel-specific pack constraints. None of these fields needs to become a universal performance score. They help the buyer distinguish a demand signal from an availability effect.

Suppose a neutral colour sells through quickly but was also the only option consistently in stock. The decision may be to repeat it, but the importer should not conclude that brighter colours lack demand. Conversely, a size that appears to sell slowly may have been repeatedly unavailable in the channel where it was most relevant.

The first lot should make such uncertainty visible instead of forcing a complete range decision from one aggregate sales number.

Read returns as assortment evidence

Returns are often counted at style level and discussed as a product-quality problem. For an assortment decision, their meaning is more specific.

A return code can reveal that one size curve is poorly described, one colour photograph creates the wrong expectation, one pack is inconvenient for a retailer, or one label version confuses the customer. The importer should connect that signal to the exact sellable combination rather than treating every return as a verdict on the supplier or garment.

This does not mean accepting every customer comment as fact. It means recording enough identity to test whether the same reason repeats for the same combination. A repeated fit comment tied to one size run supports a different decision from scattered preferences across the entire range.

Give each combination one of three outcomes

At the end of the first lot, the importer does not need a single yes-or-no decision for the whole style. Each sellable combination can receive one of three outcomes.

Repeat: the version, channel and basic economics are clear enough for another bounded order.

Revise: the combination may remain viable, but the size curve, image, pack, label or product data needs a defined change before it returns.

Stop: the available evidence does not justify more inventory, or the supplier cannot reproduce the accepted version reliably.

The value of these categories is operational. They turn pilot evidence into an order instruction. “Customers liked the style” is too broad for purchasing. “Repeat black in the accepted size curve for the marketplace; revise the blue photography before re-listing; stop the mixed retail pack until the inner quantity is workable” can guide a real reorder.

Protect the decision from a changed version

An assortment result applies to the version that was tested. If the supplier changes fabric, trim, colour reference, label or pack configuration, the importer should not automatically carry the old result forward.

The change may be acceptable, but it needs an explicit link to the pilot record. Is it the same sellable combination with a documented update, or a new version that requires a smaller repeat test? Without that distinction, the second order may scale an item the channel has never actually seen.

This is why the assortment map should include the accepted sample or version reference. It keeps sales evidence and product identity in the same decision.

Expand only what the channel has explained

The first lot is valuable when it reduces uncertainty. It should tell the importer not merely whether units moved, but which combinations the channel can describe, stock, sell, support and repeat.

If the buyer expands every colour and size after one encouraging aggregate result, the pilot has not controlled risk; it has only delayed the full commitment. A better expansion is uneven. Strong combinations move first. Ambiguous combinations return with a specific revision. Weak or unreproducible combinations stop.

It gives merchandising judgment a stable object: the garment combination, in the exact channel, under the version that produced the evidence. That is how the first lot earns the right to become a range rather than merely becoming more stock.

About the author

Yinghang Wu is the founder of a China-based advisory platform for importers, distributors and retail buyers evaluating consumer brands through product evidence, controlled pilots, channel fit and reorder decisions. Contact: yinghang.wu@chinabrandpath.org | WhatsApp: +8619883990010 | Location: Jinhua, Zhejiang, China.