Using Risk Management to Optimize Inventory
PNG is a suite of tools for inventory control of items with highly variable demand. The tool combines the capabilities of two of LMI's advanced inventory optimization models—Peak Policy and Next Generation (NextGen). Peak Policy optimizes inventory levels for items with infrequent, intermittent demand, while NextGen handles items with more frequent demand. Assignment logic, based on demand frequency, automatically routes items to the appropriate model. PNG improves forecasting accuracy while balancing the risks associated with the various dimensions of inventory performance, such as inventory value, customer service, and procurement workload.
Avoid Investing in the Wrong Items
The unstable demand patterns for many items makes them inherently unforecastable and results in large forecasting errors. Much like the flip of a coin, the processes that are generating the demand defy prediction. Buyers using traditional methods to forecast items with intermittent, unstable demand patterns frequently end up with the wrong items on the shelf. Due to such forecasting errors, enterprises experience overinvestment in some items, shortages for others, and excessive and unstable buyer workload. Safety stock and order quantity computations lose key information in the data, compounding these problems.
Balance Investment and Risk
LMI’s advice: stop forecasting. Instead PNG employs a risk-based hedging strategy that uses demand history and item characteristics to set procurement levels for unforecastable inventory. Input data includes five years’ (or the available) transactional demands, prices, current asset data, replenishment lead-times, minimum buy quantities, and other item characteristics. The tools generate tradeoff curves that display tradeoffs between wait time, inventory value, and annual buys. A supply chain manager picks a point on a curve that’s best for desired business outcomes, and PNG calculates minimum and maximum levels for each part. PNG doesn’t require accepting a forecast with unknown financial implications—there’s no question of how much of the forecast to honor. It also avoids separate decisions for safety stock and order quantities—selecting a single point on a curve is both simpler and more comprehensive.
Reduce Inventory, Improve Customer Service, Lower Workload
Setting inventory levels directly based on demand transactions rather than on artificial statistical models improves forecasting accuracy and results in significant benefits. PNG helps planners make integrated decisions and tradeoffs that for items with volatile, intermittent demand:
- Reduce inventory investment without degrading customer service;
- Improve customer service without increasing inventory investment;
- Lower the number of procurement actions.
Rigorous testing of PNG against the best-known inventory approaches shows that for items with volatile demand the tools can:
- Reduce inventory investment up to 15 percent;
- Reduce procurement workload up to 50 percent;
- Reduce customer wait time 20–30 percent.