|
Businesses often resort to stockpiling inventory to ensure they are ready to respond quickly to fluctuations in demand. But in a fast-paced, time-sensitive marketplace, that’s an inefficient and costly way to manage supply chains. Reducing inventory while maintaining, or even enhancing, service levels is one of the first items that needs to be addressed when trying to optimize business processes.
While many Enterprise Resource Planning (ERP) systems allow inventory policies and replenishment strategies, they don’t automatically set these policies. Furthermore, most of the inventory optimization tools used to determine appropriate safety stocks (buffers against running out of items while inventories are being replenished) assume a normal distribution of demand in their calculations. But common sense and experience with numerous IBM clients demonstrate that "normal" just doesn't apply for significant portions of stock in a typical warehouse. There are simply too many variables – from business expansions to weather to sales promotions – to think a single set of rules will apply to forecasting. This is especially true for spare parts operations in such industries as automotive, electronics, and aerospace and defense.
These disparities have prompted IBM Research and IBM Global Business Services to take a fresh look at calculating supply chain inventory.
As part of an overall supply chain management suite called Inventory Management, IBM Research is developing an advanced tool to more effectively handle a wide range of demand patterns. By analyzing inventory, the new tool classifies the products and calculates buffer stock levels, reorder points and batch sizes for every product in a warehouse.
The underlying principle uses advanced clustering, which serves to reduce the amount of data and to induce categorization. It groups items with similar historical demand distributions and then applies cluster-specific safety stock calculations to individual items. The method has been used successfully for making forecasts, but until now, not to improve safety stock calculations.
The parameters that result from such calculations lead to far better correlations between actual and planned service levels. They also contribute to greatly reduced safety stock requirements – which can range from as much as 15 percent to 40 percent.
Such outcomes have been verified in numerous instances. In fact, the dynamic inventory optimization system tool has been successfully used on more than 30 client engagements. In practice, one retailer is now using the system daily to gather data on 70,000 items from 80 stores, totaling more than 5.6 million datasets. The system then analyzes the data, considers factors such as promotions, creates order proposals for each supplier in each location and generates reports the company’s management can use immediately. For example, the system can determine the way a promotion affected a particular item and forecast the inventory requirement when a similar promotion runs again.
This sophisticated automation allowed the company to quicken the pace of stocking inventory and filling outstanding orders and to better forecast inventory shortfalls and overages, which helped it to save money. In the simplest terms, the company is better able to manage inventory and, at the same time, boost service levels.
IBM Research continues to work on this unique tool and includes multilevel inventory optimization, as well as calculation programs for computing buffer stocks in extreme cases. One example might be when companies are dealing with a number of events that occur in specific time periods, such as holiday seasons in the retail industry or distributions that are not close to normal.
The latter situation often occurs in the spare parts business, where IBM researchers are currently focused on developing solutions that will be applicable across a number of industries. For example, IBM is using the tool in a pilot project designed to help a major European automobile manufacturer better manage its spare parts inventory, with a possible worldwide rollout planned through 2009.
The Dynamic Inventory Optimization System is now delivered through the Center for Business Optimization (CBO) as part of the Dynamic Inventory Optimization Solution. For more information on putting this powerful inventory control tool to work for you, contact the CBO or On Demand Innovation Services.
Read the DIOS brochure.
Related Case Study Max Bahr
| 
