SmartForecasts Sparks Company-wide Improvements and Better Supplier Relationships at Siemens Building Technologies
Five years ago, when Bob Perlowski was manager of materials management at Siemens Building Technologies (SBT), he knew there had to be a better way to forecast the material needs for the company's four divisions. Like many companies, SBT was feeling the demands of increased competition and multi-national business growth, and its forecasting process just couldn't keep up. SBT is large-a $1.4 billion subsidiary of the multinational giant, Siemens AG. SBT's divisions manufacture products and technologies used to control heating, ventilation and security control systems in commercial buildings.
Back then SBT's forecasting process was completely manual. As we continued to grow, we ran out of horsepower, remembers Perlowski, now director of production operations. The whole process became a big burden.
The forecasting process involved hours of discussion in a conference room filled with white boards, numbers and post-it presentation charts hanging everywhere. Every number on the charts was filled in by hand. With 14,000 different product items, planners could only forecast the company's high volume items, and were forced to neglect the others.
Planners were so mired in the details that they couldn't detect seasonality, trends or other patterns in the demand data even when they repeated on a regular basis. And, 20 percent of the neglected products had intermittent, slow-moving demand-a type of demand that is particularly difficult to predict because it has few, if any, identifiable patterns in the demand history. Instead, intermittent demand data typically contain a large percentage of zero values with non-zero values mixed in randomly.
The change in SBT's forecasting process has been gradual. In 2000, Perlowski started looking for a solution, and in 2001, SBT installed Smart Software's SmartForecasts Enterprise demand forecasting and inventory optimization software with its intermittent demand forecasting capabilities. There were three reasons the company decided to purchase the software system:
- SmartForecasts is one of the few forecasting and planning tools that can handle the large number of products SBT offers
- The software has a simple, intuitive Windows interface, fashioned after MS Excel, but
- It also has sophisticated capabilities to analyze demand data and create accurate forecast results.
When SBT first started using SmartForecasts, planners used three years of rolling demand history from the company's home-grown order management system and exported it into SmartForecasts as a comma-delimited flat file. The forecast was produced and saved in SmartForecasts and automatically uploaded to the company's ManMan manufacturing system.
The old order management and ManMan systems are now gone. In 2002, the company installed Oracle's database management system company-wide and also upgraded its SmartForecasts installation to a site license. Today, SBT is beginning to see the positive results. It now has a consistent forecasting process across its business units which is based on a common set of planning tools and a common database to store and distribute results.
The new Oracle system is the repository for demand history and forecast data. Sharing the information has been greatly facilitated, and plans are afoot to eliminate the comma-delimited flat file intermediate step. Using its built-in database connectivity wizards, SmartForecasts will link directly to Oracle, greatly simplifying access to data and management of forecast results.
With SmartForecasts up-and-running, better results were noticeable almost immediately. SmartForecasts enabled us to see patterns in our demand data that we had never seen before, said Perlowski. He now understands the seasonality in SBT's business. When we'd get together at our forecasting sessions, he said, we used to talk about numbers; now we talk about trends.
And having a better understanding of the data has helped Perlowski prepare and do his job better. Using SmartForecasts' promotion modeling capabilities, SBT is now able to model the effects of its promotions and manage future demand more efficiently. SBT is also able to forecast many more products than before, helping it to better balance its inventories-reducing stocks of some products and increasing others-while maintaining or increasing service levels. According to Perlowski, there are now fewer change orders and the factory runs smoother.
One of the reasons operations are smoother, cites Perlowski, is that SBT can now provide its suppliers with more accurate information, enabling them to provide better, more timely service. SBT can also do this faster because forecasting time has been reduced by 75% using SmartForecasts.
Since the implementation of its new Oracle system, SBT has rolled out SmartForecasts to its marketing people in all the divisions. The marketing people, who are all resident at SBT's Buffalo Grove, Illinois campus, now do their planning around the forecasting process. Perlowski's group generates a monthly statistical forecast based on historical booking data obtained from the Oracle system. Product managers, using SmartForecasts on their desktop workstations, can examine the results and adjust them based on new or changing market conditions. Everyone then meets to review the adjustments and generate a consensus forecast.
SmartForecasts has greatly increased our confidence in the forecast results, remarked Perlowski. We're now getting good, hard core information, and our communications among business units and our suppliers has improved immensely.
Back to Success Stories