DEMAND PREDICTION FOR A TEXTILE MANUFACTURER
A client in textile industry with 7,000+ products and delivering to six markets approached us to create a demand prediction model to enable them to optimize their inventory management and increase their inventory turnover. The prediction was aimed at short-term adjustments / orders as well as long-term planning of up to 1.5 years.
In order to achieve desirable precision of demand prediction for each product, our solution combined varying granularities and several dimensions of data. This meant that product specific attributes were combined with product category attributes, market specific attributes and company specific attributes. The resulting factors were then given an estimated weight and combined into a final forecast.
We achieved c.20% reduction in inventory while still being able to serve all orders within the required time frame. In addition, the daily forecast recalculation enabled a swift reaction to any unexpected sales / orders, thus mitigating the risk of unforseen events influencing the availability of products.