Forecasting branded footwear sell-through by SKU

1. Forecast at the SKU, not the category

Category-level forecasts hide the truth. Two silhouettes in the same brand can sell at very different rates — one Air Max colourway will outsell another 5:1 in the same store. A category average leads you to over-buy the slow one and under-buy the winner.

Break every forecast down to at least style + colour. If your system supports it, go to full SKU (style + colour + size). The extra granularity is the whole point.

2. Size safety stock only on the top 20%

Roughly 20% of your SKUs drive 70–80% of revenue. Those are the ones that must not stock out. The remaining 80% can and should run lean — the cost of a stock-out on a slow line is much lower than the cost of dead stock on 200 of them.

• Rank SKUs weekly by trailing 4-week sell-through, not lifetime units sold.
• Hold explicit safety stock on the top quintile — usually 2–4 weeks of demand at current velocity.
• Let the long tail run at reorder point with no safety stock. If it sells out, it sells out.

3. Calculate a real reorder point

The reorder point is the stock level at which you place the next order. The formula is straightforward:

Reorder point = (weekly velocity × lead time in weeks) + safety stock.

Weekly velocity is the trailing 4-week average of units sold — recent enough to reflect what is happening now, long enough to smooth out a single spike. Lead time is the number of weeks from placing an order to it landing on your shelf, including receiving. Safety stock is only on the top 20%.

4. Trigger reorders on velocity, not stock level

Waiting until a SKU hits a stock threshold means you have already lost sales. Instead, trigger the reorder when 4-week velocity crosses the reorder point — that way you replenish while stock is still on the shelf and the sales signal is still fresh.

This is the single biggest change most branded footwear retailers can make. It turns replenishment from a reactive exercise into a predictive one, and it is the reason distributor-led retailers stay in stock on their best sellers while the rest of the market runs out in mid-November.

5. Cut the long tail without emotion

Every quarter, list SKUs with zero units sold in the last 60 days. Discount them aggressively or return where the distributor allows it. Dead stock ties up cash you need for the top 20% — the ones actually driving your margin.

The emotional bit is that you personally bought some of those lines and you liked them. That is not a reason to keep them. Retail is decided by sell-through data, not taste.

Worked example: a single Nike Air Max SKU

Trailing 4-week sales: 12 pairs per week. Distributor lead time: 2 weeks. Sitting in the top 20% by revenue, so safety stock is 3 weeks of demand.

1. Reorder point = (12 × 2) + (12 × 3) = 60 pairs.
2. When on-hand stock hits 60 pairs, place the next order.
3. Review the 4-week velocity each Monday and recalculate.

If velocity jumps to 18 pairs a week, the reorder point moves to 90 automatically. You order more, sooner, without waiting for a stock-out to tell you.