Demand forecasting for short time series under volatile and uncertain market circumstances

I'm working on demand forecasts for five different market segments, with each their own product type. Each product type has 20 years of monthly data and therefore 240 data points to rely on. My goal is to forecast demand for each product type, looking one and three months ahead.

Currently I'm using the traditional ARIMA approach. Normally this works well, however due to new market circumstances with high uncertainty and volatility the latest predicitons where disappointing. Therefore, I am looking …

data datascience demand demand forecasting five forecast forecasting product series time series type uncertain