Which forecasting method accounts for sporadic demand?

Croston's method is specifically designed to forecast demand that is sporadic or intermittent. Traditional forecasting methods like moving averages, linear forecasts, and Holt-Winters are generally effective for stable and continuous demand patterns but often struggle with data that exhibits irregular spikes or lulls, which is characteristic of sporadic demand.

What sets Croston's method apart is its unique approach—rather than simply averaging demand over time or fitting a linear trend, it separately estimates the demand size and the intervals between non-zero demands. This means it can provide more accurate forecasts for items that do not sell consistently or have random purchase patterns. Consequently, it effectively captures the essential characteristics of sporadic demand, making it a suitable choice for businesses dealing with such forecasting challenges.