Counting Cookies
Food cravings are serious business, especially when a breakdown in the sales and inventory process might result in consumers not having what they want when they want it. Nabisco wanted to use its coporate intranet to give salespeople real time information that would let them accurately forecast product inventory needs and shift good between locations to meet surges and lags in demand.
The key to Web-to-host success at Nabisco was finding the right tools to give the sales teams information secured within the mainframe.
Product Information UltraQuest and NOMAD Aonix San Diego, Calif. 1-800-972-6649 www.aonix.com Pricing information available from the vendor. |
When you’re at your favorite convenience store looking for that late-night package of Oreos or a can of Planter’s peanuts, the last thing you want is to find that the store is out of stock. To ensure that never happens, Nabisco recently revamped its sales and inventory data access systems, thus empowering its sales force to better track inventory.
Nabisco, with headquarters in Parsippany, N.J., made the changes in order to unlock – and securely share via the Web – the legacy sales and inventory data stored in mainframe databases. (Nabisco is a division of Kraft Foods North America, the largest branded food company in the U.S., which in turn is part of Philip Morris Companies Inc.) With information posted on a corporate intranet, Nabisco sales teams are now armed with crucial information that lets them accurately forecast product inventory requirements for seasonal items and immediately shift goods between locations to meet surges and lags in demand.
"Our sales teams can post product allocation information on the Web," explains Sandra Olshefski, Nabisco’s customer service analyst for product allocation. "Teams can trade allocations between one another, increasing or decreasing inventory where it is needed. The entire system operates sort of like an internal eBay." Sales teams are distributed across the country and are divided either by geography, or grouped as large customers, in cases such as Kmart or Wal*Mart.
By using what Nabisco calls the Customer Allocation System, sales teams distributed across North America can immediately determine the accuracy of sales forecasts, pinpoint exact sales information and see how well their products are selling. Sales teams with too much of a particular product can offer the excess to other teams; sales groups running short of stock can post requests for additional deliveries. Coordinating, validating and expediting all of the activities is Olshefski. She makes sure the trades "are valid," then coordinates disbursement and shipping between locations.
Part of the challenge is ensuring that continuity remains between regularly scheduled product shipments and the "need-it-now, just-in-time" requirements for seasonal merchandise. Because order shipments are virtually instantaneous, Olshefski must stay right on top of the orders and acts as arbiter, flagging shipping discrepancies when the incorrect items are about to be sent to the wrong locations.
By harnessing the power of the OS/390 mainframe with its MVS operating system and leveraging legacy applications for inventory, shipping and distribution, the interactive system helps Nabisco better meet its goal of selling its products. In addition, according to Marilyn Brown, senior manager for logistics systems, the Web-enabled application falls in line with a corporatewide business initiative at Nabisco to preserve and enhance mainframe-based legacy systems using Web extensions and other mechanisms that bring much-needed data closer to the business units. "We’re standardizing the way users access information using the mechanisms that they’re most familiar with," she explains. "A Web browser and Excel spreadsheets are the most common methods."
Once delivered as thousands of ad-hoc screen-based and paper reports, a large amount of corporate data is now delivered from the OS/390 DB2 databases that support the inventory, shipping and distribution applications using both the UltraQuest family of products and NOMAD (both products are from Aonix). The combination of these tools enables Nabisco to extract, format, combine and deliver mainframe legacy data to Web-based applications or Microsoft Excel reports. "NOMAD code is used to pull information from DB2 tables and UltraQuest provides the browser-based viewing functionality," explains Brown. "By directly delivering mainframe data to the Web, we’ve been able to reduce ad-hoc reporting and standardize on our information delivery mechanisms."
Need the Cream Filling Key to the Nabisco project was a product called NOMAD from Aonix. The NOMAD program integrates non-procedural commands with full procedural language and facilitates quick reporting, database maintenance, prototyping and application development. Used in conjunction with the UltraQuest products, NOMAD provides Nabisco with direct Web-to-mainframe access to legacy data and provides the basis for tight integration to DB2. Additionally, NOMAD provides efficient access to IMS, VSAM, IDMS, Teradata, QSAM, Oracle and Sybase data sources. The product contains a variety of application generators and user interface options, including a reporting front end and a graphical user interface design tool, which help speed the development and deployment of NOMAD applications. NOMAD also contains a large number of reporting and analysis functions, including integrated decision-support features, arrays and time series and financial formatting. Features for application development activities include an integrated 4GL and procedural language, programmable windows, database editing, performance monitoring and application generation tools. -P.E.C. |
Forecasting seasonal products comes from an analysis of historical sales data, trends and demographics – along with a review of competitive products. The final seasonal product forecasts generated by all of the sales teams serve as the basis for manufacturing. "We gather all this information to determine how much we need to produce," says Olshefski.
Before allocation information was placed on the Web, sales teams had no way of knowing just how well products were selling – until demand outpaced supply or there was an extreme overstock requiring a liquidation sale. "Our sales teams are distributed throughout the country and never knew whether they had ordered too much or too little," Olshefski explains. "They also had no effective way of letting the other teams know when they had excess product or if they required more items."
Also, without the ability to easily monitor product sales, managers found it difficult to effectively balance forecasts with future manufacturing requirements. "Manufacturing efficiency could have been improved if we knew exactly what our overall product sales numbers were," Olshefski contends. "We knew we could save money and operate more efficiently if we produced only what we needed." While Nabisco did not monitor exact cost savings, Olshefski estimates that, when compared to the former manual method of product allocation, she can now handle four times the number of products in 75 percent of the time.
The key to success, however, was empowering the sales teams with the information secured within the mainframe.
Puttin' on the Ritz As an eight-billion-dollar-plus manufacturer of cookies, snacks and other food products, Nabisco markets its goods in the United States, Canada and more than 85 other countries. Brands include Oreo and Chips Ahoy! cookies, Ritz and Premium crackers, SnackWells, Planters, LifeSavers, A.1 steak sauces, Grey Poupon mustards and Milk-Bone dog snacks. While most Nabisco products are sold year-round and have a long shelf life, the process of forecasting and selling seasonal products is a precise science. Holiday products, such as those sold around Christmas, Easter, Valentine's Day and Halloween, can appear on shelves only for a limited time. And, since they're food products, items can't be saved for next year. In addition, according to Nabisco's Sandra Olshefski forecasting activities must be completed early enough to ensure adequate lead time for production and shipping. For seasonal items, lead time means allowing two to three weeks of shelf time prior to the holiday in question. "All of the seasonal items must be sent within a limited shipping window," Olshefski explains. "There's no opportunity for seasonal product substitution. Not many people will buy a chocolate Easter bunny during Halloween." -P.E.C. |
Data Ahoy!
Before data could be extracted and posted to the Web, Olshefski and Brown did an analysis to determine three things: the exact information needed by the sales teams, the DB2 tables where the data resided and how the integrity and security of the mainframe environment would continue to be protected.
Led by Olshefski, the requirements for the sales teams were created by developing a list of ideas that would help enable them to do their jobs better and help in reaching an accurate decision on whether or not to increase production or reduce liquidation for specific products. Internal research conducted by Brown revealed that the necessary data resided in various locations on the mainframe. Still, while the UltraQuest and NOMAD tools provided direct access from the Web to the mainframe files, Nabisco opted to use an alternative method, with alternative files, for performance and security reasons. That decision was made partly because Nabisco feared that direct access to critical files from the Web might simplify data theft or destruction by outsiders.
"It was important to get the information out to the business units and sales teams," explains Brown. "But, we needed to make sure there was no performance impact on the production machines, and we wanted to make sure that the different DB2 databases remained secure." To address the security and performance issues, it was decided that sales team allocation data would be made available from a database that was separate from the production DB2 tables. A series of NOMAD programs were written to extract the needed data for storage into a NOMAD database. Written by Brown, the 10 extract programs run each evening during the company’s batch processing cycle, formatting the data for UltraQuest.
UltraQuest works by providing a connection to host data and applications through a mainframe-based Web server. The Durable Server Facility (DSF) contained within the product is a configurable resource manager designed to handle high-volume traffic and reduce the cost of mainframe access. This facility manages pools of application servers that run in the Nabisco MVS environment and remains active between transactions, helping to reduce operational overhead. UltraQuest contains tools for testing, tuning and tracking connections against mainframe resources. Additionally, the UltraQuest Library provides a centralized repository for storing, standardizing and reusing reports and menus. Publishing capabilities allow reports created on the desktop to be shared across the enterprise.
In addition to the extract programs, Brown notes that five UltraQuest utilities are used to create the proper report formats for the Web browser. "We merged all of the information from the DB2 tables and created a new sales team table," she explains. "Using their browser, they can see the data in any fashion they desire."
Says Olshefski, "We can view the information by product, by season, by SKU, by groups of products and more."
Brown says the entire project took approximately four months, including requirements creation, prototyping, generating sample applications, and gathering and incorporating user-specified modifications. "The changes requested by the sales teams were minor and required some tweaking," she explains. Working with NOMAD and UltraQuest was "pretty straightforward," she says, once the vendor addressed product implementation issues unique to Nabisco. "We weren’t using the tool the way it was designed," says Brown. "Instead we were going from the mainframe through our corporate intranet instead of directly to the Web. Aonix helped us overcome our issues so that we could use the product the way we required."
Not a Lifesaver, but Close
According to Brown, using NOMAD and UltraQuest to provide Web access to legacy data will enable Nabisco to further exploit the capabilities of existing applications that have proven reliable and successful. "That’s the beauty of this solution," she says. "We already have procedures in place, and, in many cases, just need to make some modifications instead of starting from scratch." It’s now a matter of identifying which additional applications are candidates.
For example, she notes that some existing legacy applications employ ad-hoc reports to deliver information to users. However, those reports proved problematic because they had to be archived and frequently consumed large amounts of Direct Access Storage Devices (DASD) space. "We run out of space all the time because of these ad-hoc reports," she says. "But, to publish the information on the Web, it’s merely a matter of changing a step within the program." Instead of writing to a file, the data is written to either a browser or an Excel worksheet. Excel is the overwhelming preference of users, according to Brown, because it allows users to format the reports more than a browser does.
For Olshefski, bringing legacy data to a Web front end has helped eliminate the mystery behind inventory counts, while enabling Nabisco to remain more productive and more competitive with its seasonal items. "It’s the hundreds of small improvements that have a major improvement on our overall bottom line," she says. "Anything we can to do improve the availability of our products represents better service for our customers."
Philip E. Courtney has published a variety of articles in technical and trade publications, and is president of Quality Product Marketing. E-mail him at [email protected].