Through the course of a decade, Stout has provided a broad range of valuation advisory services to a leading global developer, manufacturer, and supplier of products in the transportation industry.
We were recently engaged by this client on a different topic: how to manage inventory and reduce carrying and obsolescence costs. Our client held an estimated $250 million in inventory across 16,000 SKUs and was incurring an estimated $20 million in carrying and obsolescence costs annually. Our goal was to help our client become more efficient and accurate in forecasting demand, which would impact its inventory levels and, ultimately, its costs.
We conducted an initial pilot in which the client provided our team with historical data over a 36-month period for 1,600 SKUs. The data was input into the proprietary artificial intelligence (AI) platform to generate a model that would develop forecasts for the initial batch of 1,600 SKUs. To test the pilot model, we compared actual demand with the model results. The model results showed substantial reductions in both forecast errors and volatility, with errors declining from the current 38% to 7% with a corresponding decrease in forecast volatility. The pilot took two weeks once the data became available to Stout.
During that timeframe, we were able to gain insights, test multiple hypothesis, and identify previously unknown correlations. This work enabled us to generate early-learning results, develop a complete analysis, and gain additional insights – all without having to build or implement new technology, processes, or systems, use significant resources from our clients, or interfere with existing processes. In fact, one of the primary benefits to the client was that we could easily integrate our models into the existing process.
We are currently expanding the engagement to complete the process for all 16,000 SKUs and expect to deploy a production model that will substantially improve the forecasting process and reduce inventory levels in the entire supply chain. In production, the AI system will automatically gather data from the enterprise resource planning (ERP) system and will create, maintain, and improve forecasting models based on dynamic data. The model will constantly measure, monitor, learn, predict, and act in an automated way, improving efficiencies within the forecasting process.