Sourcing
20 October 2026
A 21-model catalogue with 293 SKUs invites a first order that covers everything thinly: twenty models, twelve colours, fifty pieces each. That order satisfies nobody. Half the colourways sell out in three weeks, the other half are still on the shelf at month eight, and the reorder decision is made on emotion because there is no data.
Assortment planning is the least glamorous part of this business and it decides the margin. Here is the arithmetic we would do.
The catalogue is the factory's inventory of possibilities. Your order should be built from what your channel can move in a defined period. Two numbers drive everything: expected units per week at full availability, and the lead time between placing a reorder and having it on the shelf.
If the reorder cycle is eight weeks by sea and you sell forty units a week of a fast colour, you need at least 320 units of cover just to survive the gap, plus a safety margin. That single calculation usually tells you that the first order has to be deeper and narrower than instinct suggests.
| First order size | Models | Colours per model | Depth per SKU | Comment |
|---|---|---|---|---|
| 500 pcs | 3–5 | 2–3 | 50 | Test the channel, not the catalogue |
| 1,000 pcs | 6–8 | 2–3 | 50–100 | The usual first container-adjacent order |
| 3,000 pcs | 10–14 | 3–4 | 75–100 | Enough to see a real colour ranking |
| 10,000 pcs | 14–21 | 4–6 | 100–150 | Now breadth is affordable |
The pattern is deliberate: breadth grows roughly with the square root of the order size. Doubling the budget should not double the number of SKUs, because every additional SKU adds a stock-keeping cost, a reorder decision and a markdown risk.
A 21-model catalogue covers Apple and Samsung across several generations. For a first order in Germany, Canada or the US, the sequence that works is:
Across our catalogue, six colourways are photographed because six is what sells. The distribution is consistent: two neutrals — black and a dark navy or grey — take roughly half of all units, two mid-tone fashion colours take a third, and the remaining brights share the tail.
An order that gives equal depth to all six colours will overstock the tail by a factor of three. The colour approval article explains how to lock the neutrals so they match across reorders — which matters more than it sounds, because a navy that shifts between batches makes the whole shelf look inconsistent.
Most price breaks are triggered by total order quantity, not by quantity per colour. That means a ten-model, three-colour order of 1,000 pieces can reach the same tier as a two-model order of 1,000 pieces. Consolidating across colourways is the single most reliable way to improve the landed cost without changing the specification.
Ask explicitly: at what total quantity does the next price step start, and does the total include all colourways. The quotation structure article covers where that line sits in a quote.
Sea freight from South China to a European or North American port is measured in weeks, and port congestion adds variance. The practical rule: place a reorder when stock cover reaches the sum of production lead time plus transit plus a two-week buffer. For most wholesalers that is a reorder point of six to ten weeks of cover, not two.
Which means the first order's job is not just to sell. It is to generate the sell-through data early enough that the second order can be placed before the first one runs out. Narrow and deep does that; broad and thin does not.
1,000 pieces, three markets, one channel. Six models — two current flagships, two previous-generation flagships, two high-volume mid-range — at three colours each: black, navy, and one seasonal accent. Depth: black 70, navy 60, accent 37, per model, which totals 1,002 pieces. Black carries the reorders, the accent creates the fashion signal, and there is enough data after six weeks to know which two models deserve depth on the next order.
That is a boring order. It is also the one that gets reordered.
Roughly three to five models with two or three colours at 500 pieces, scaling to fourteen models at 10,000 pieces. Breadth should grow with the square root of order size, not linearly.
Current flagship plus previous generation, and do not order a model that is weeks from being superseded. The installed base of a year-old flagship is often larger than the new launch.
Usually not — most tiers are triggered by total order quantity across the order. Ask explicitly whether the total includes all colourways before you split an order to reduce risk.