Running out of a bestseller during a growth push, or sitting on dead stock from a slow season, both usually trace back to the same root cause: forecasting by feel instead of by data.
Sales velocity is the starting point
Average units sold per week, per SKU, over the last 90 days gives a baseline. From there, adjust for known seasonality and any planned promotions, rather than guessing at a round number.
Reorder points, not reorder panic
A reorder point, the inventory level that triggers a new order, should account for supplier lead time plus a buffer. Without one, reordering becomes reactive instead of planned.
Slow movers deserve a decision, not neglect
SKUs that consistently underperform should get an active decision: discount them out, bundle them, or discontinue them. Letting them sit ties up cash that could fund a bestseller's next order.