Hundreds of Thousands of Items Distributed Accurately Every Day
In an industry where it is dangerous to stand still, the retail business is synonymous with change. However, if you are constantly on the move and repeatedly changing your processes you need a flexible IT structure. "We want our software to adapt to our processes, not the other way around," states Bernd Nagel, IT services system manager at Müller Ltd. & Co KG, explaining why his company does not use a standard ERP system, but instead supports retail operations with proprietary applications.
From its small beginnings in Ulm, Germany, Müller has become one of the top 20 companies in the industry in Germany with more than 400 branches, 18,500 employees and revenues of USD $2.23 billion. And it continues to expand with dozens of new branches opening every year, with stores now in Switzerland, Austria, Spain and Slovenia.
Müller's key differentiator is its large range of low-priced products, which is constantly adapted to suit the changing demands of consumers. Retail stores include a variety of departments and specialist areas – from medicines and perfume to stationery, household, laundry and handicraft products, books and multimedia. Nagel explains the company's philosophy, "The products that a customer wishes to buy have to be available locally at all times. We also have to respond to new demands as quickly as possible." The advertising tagline sums up the company's view: "Müller brings happiness."
Logistical Challenge: Store Requirements Planning
Behind this simple philosophy lies a complex logistics operation, ensuring that the philosophy becomes a reality. Müller sells around 300,000 items and its operations encompass over 479,000 square feet of storage and a logistics center in Ulm. The fleet of more than 50 trucks and 240 cars, as well as 40 large-haul vehicles, has to be precisely controlled for optimal efficiency during the 8 million miles traveled each year.
Müller's IT team is tasked with the management of the product range - from accurate demand forecasts to timely deliveries. Due to the large volume and number of items, and the difference in customer preferences from location to location, store requirements planning is extraordinarily complex. The aim is to ensure supply based on the expected sales profile of each individual branch. At the same time, the company must determine the optimal cost-to-benefit ratio with third-party goods, overall stock levels and individual branch inventory.
Expansion into international markets has made this logistical challenge even more complex. Some of the goods are shipped directly from suppliers in the operating countries and the some are shipped from the central storage facility in Ulm. Additionally, each country has specific regulations governing operations, particularly concerning the sale of medicines.
Three Central Applications Control the Flow of Goods
The process is controlled by three central applications in addition to local branch systems: a retail operations system, a warehouse management system and a forecasting system. The retail operations system is the central application, which controls workflow for all the processes related to each item. It contains key specifications, supplier data and inventory details. A 20-member team of internal and external specialists has continually developed the system since 1998. The warehouse management system stores details of goods in and goods out as well as current inventory levels. Both the warehouse and retail operations systems update each other and control the flow of goods.
The forecasting system uses two year's worth of historical data, to forecast which goods should be shipped to each location and in what quantities. In addition, it constantly ‘learns' by comparing its estimates with actual sales. The forecast system primarily identifies relationships, such as which items sell best under which circumstances. Daily, the system calculates expected branch sales for the next twelve months, maintaining shelf stock levels to reduce gaps in stock and to increase shelf turnover.
The interaction between these applications provides inventory planning for all stores. Based on forecasts, sales and inventory, the system compiles the daily branch deliveries. The warehouse management system uses a route map to calculate branch delivery schedules. At the same time, the retail operations system automatically generates orders to suppliers based on forecast, sales and stock data. Upon confirmation by a Müller buyer, the order is transferred to the supplier via EDIFACT.
This time-critical processing based on extremely large data volumes demands a highly efficient IT infrastructure. "The system requires approximately two hours for a large branch, compared to just 15 minutes for a small branch," reports Nagel. The company's success relies heavily on these core processes running smoothly. Nagel continues, "We closely monitor the processing of planning data to ensure that everything is running as it should."
Millions of Delivery Data Records Demand Maximum Performance
The applications generate a large volume of delivery data. Up to 800,000 transactions are processed each day. In addition, over a hundred million historical records are retained for forecasting. The applications require an extremely efficient database. Since the introduction of its server systems, Müller has relied on Sybase ASE. Explaining the choice of Sybase ASE, Nagel says, "Benchmarks showed that Sybase ASE is significantly faster than other databases on the market when processing high volumes of data." Since the initial implementation, the requirements have escalated dramatically. Tables in the retail operations system now include up to 60 million data records, while the forecasting system saves up to 120 million data records per table. The retail operations system database has reached a size of 250 GB, and 300 GB for the forecast database. ASE has kept pace with this data growth, without increasing administrative efforts.
"Sybase ASE is extremely straightforward and is simpler to manage than other databases," states Gernot Wolf, systems administrator at Müller. "For example, the optimum parameters for operative performance can be configured. They are meaningful and easy to understand. When I do something, I know exactly what will change. Since these are displayed in ASCII files, we can also manage the parameters even when the database is offline. And thanks to the functionality of ASE, complex processes are simple to perform. For example, a temporary database is used to display jobs, instead of traditional tables."
The Future: Improving Process Automation
Nagel is striving to achieve even greater process optimization for the flow of goods in the future. His aim is to achieve transparency into branch inventory and sales information. "We want to be able to recognize instantly what stock a given branch has at a specific time." At present the data still has to be transferred separately, because inventory management in the branch system does not run on the same basis as in the central system. Plans for new Java-based point-of-sale software will also ensure a seamless flow of data.
In order to quickly and flexibly adapt processes to meet changing requirements, Müller developed a proprietary system for process analysis. Discussions with specialist departments revealed how the processes appear today, what requirements exist for the future and how processes can be defined. According to Nagel, "This process analysis generates application scenarios, which serve as a basis for our program developers."
Serving as the basis for the core systems, Sybase ASE will continue to support expansion, efficiently manage data growth and fuel process improvements, ensuring Muller never stands still.