Increase stock availability through accurate demand planning
Client: A manufacturer with more than 100 years experience in the Australian market
Challenge: Lumpy demand* with previously unpredictable client behaviour
Solution: AutoML demand forecasting methods
A manufacturing client wished to bring advanced demand forecasting and optimisation into their operations. With manufacturing globally becoming more and more competitive there was a need to bring a data driven edge to a traditional company. Having an understanding of future demand for every SKU in their inventory would allow them to plan accordingly, and optimise their operations.
Remi AI was engaged to see what predictability could be found in the client’s sales data. The demand forecasting suite developed by Remi AI includes forecasting methods that can ingest covariate data, and so the data science team leading this pilot project included data on the stock market, GDP growth, debt markets, and weather in the analysis to see whether any data points provided an indication of future demand.
The Remi AI team produced 6 week forecasts across all SKUs in the client’s inventory and was able to achieve accuracy on average across all SKUs of more than 70% which was an impressive result given the sparse nature of the data and lumpy demand. The team also uncovered an interesting insight for the client in that short term debt products were a large driver of demand for their products. The assumed logic here is that cheap debt spurs manufacturing investment which would be music to the ears of reserve bankers everywhere!
*Lumpy demand is a phenomenon encountered in manufacturing or retailing when the items are slow-moving or too expensive, for example fighter plane engines.