In-Depth

Making Chicken Nuggets More Efficient: AS/400-Based Data Warehouse Lets Zartic Have Data Its Way

• By Jon William Toigo
• 04/01/2001

The benefits of a combined Enterprise Resource Planning and Business Intelligence solution have never been so succinctly stated. At Zartic Inc., the combination amounts to a more efficient chicken nugget.

Zartic, headquartered in Rome, Ga., is a full-service beef, poultry, veal and pork processor that operates four USDA-inspected production facilities and delivers product, nationwide, via Zartran, the company’s refrigerated transportation subsidiary. The 1,100-person company produces millions of chicken nuggets annually – and more than 1,000 other products ranging from frozen hamburger patties to chicken wings – for sale to food service operators, restaurants and distributors.

Supporting Zartic’s operations is a small MIS organization comprised of nine staff members and helmed by Director Jeff Kirkpatrick. According to Kirkpatrick, the MIS crew oversees the operations of key applications hosted on the company’s AS/400 midrange system, as well as an assortment of Compaq, IBM NetFinity and "clone" servers, running the Microsoft Windows NT operating system. A total of about 200 users are connected to a modest NT network, he notes, but the tools provided by the MIS department enable them to make decisions that keep Zartic operations streamlined and efficient.

Kirkpatrick notes that the company is currently experimenting with handheld radio frequency temperature units that will automate the generation of government-mandated temperature monitoring data, both on the production line and when products are moved into inventory. He looks forward to integrating the data directly into the AS/400-based inventory control system, perhaps, within the year, further reducing manual data entry and increasing efficiency. This effort, he notes, is quite a departure from the type of tasks that MIS handled only two years ago.

Until 1999, Zartic employed a set of business systems that leveraged basic database-centric functionality and offered little in the way of customized query and reporting capabilities. Kirkpatrick says that the idea of a data warehouse and data mining applications for end users had considerable appeal, "We were spending a lot of time to design and generate reports and queries and to download data sets to end users who loaded them into Excel spreadsheets to do analysis."

Replacing this labor-intensive activity with an automated data warehouse, however, seemed like a distant dream. Most products available at the time supported only UNIX or NT hosts. Most required a complete redesign of the database, "a native OS400 database with some relational capabilities," that had provided the core of inventory control and management for many years.

"We researched the possibility of data warehousing on and off for about two years, and looked at five products, most of which were NT-, UNIX- or PC-based," recalls Kirkpatrick, "We decided that we needed something simple. We wanted a solution that would run on the AS/400 platform, which we have found to be more reliable than NT. Plus, an AS/400-based solution would capitalize on our expertise."

Based on Kirkpatrick’s evaluation criteria, the options were quickly distilled to one: DataTracker, a business performance management solution from Silvon Software. DataTracker provided Online Analytical Processing (OLAP) functionality that enabled companies to construct data warehouses or data marts using input from transaction-based application systems.

The product allows Zartic to decide what information (such as sales, financial or manufacturing data) they want to capture in a data mart. DataTracker then extracts, loads and summarizes the requested data and makes it available for "data mining" using a set of multi-dimensional analysis tools. Unlike many competing OLAP products, which limit the availability of detailed data, Silvon’s DataTracker integrated summary and detailed data into one seamless database structure.

Silvon CEO and President Mike Hennel credits the successful deployment of DataTracker at Zartic to Silvon’s 13 years of experience in software development for the AS/400 platform. Since 1987, the company has developed an extensive collection of software products in the business performance management and software development management space.

In 1997, Silvon sold off the software development business to focus on its business intelligence product family. Companies, like Zartic, were quick to seize on the initial DataTracker product offering to address their business intelligence requirements.

Hennel notes that DataTracker has become a core technology for a family of products called Stratum. Stratum has established Silvon as a "pure play" in the business performance management software market, says the president.

The Stratum product family includes four enterprise resource planning-focused modules (sales, marketing, procurement and manufacturing management) and three business performance analysis modules (e-business analytics, customer relationship analytics and activity cost management). Originally developed for AS/400, Stratum has been ported to Windows NT and RS/6000 platforms, as well, according to Hennel. It can be used with a wide range of database engines, including DB2, UDB and Oracle.

All of the Stratum modules take advantage of the four key components of the DataTracker engine, which Hennel summarizes as a database administrator (DBAdmin), a data loader (DBLoad), a schema generator (ADM) and an OLAP Optimizer. These are, essentially, the mechanisms by which the data warehouse is defined, populated with data and made available for analysis.

Says Hennel, "Stratum takes the best-of-breed analytical models and queries developed by users of our DataTracker engine and offers them as templates that can be customized by new users. The use of DataTracker by Zartic, and others, provided us with many valuable insights that we considered carefully when we created the Stratum products. We are the first to offer pre-built analytical applications for the AS/400 space, though the market will probably get more crowded within 12 months."

Key to the success for DataTracker at Zartic, according to Kirkpatrick, was the speed and non-disruptive nature of its deployment. Says the MIS Director, an initial meeting with an analyst from Silvon in February 2000 was all that it took to identify warehouse requirements and design the warehouse itself.

"We were delighted that implementing the solution did not require a redesign of our database. Silvon helped us to program the interface to our payroll, manufacturing, sales, inventory and accounts payable/accounts receivable data. We bought an AS/400 Model 170 to host the warehouse, then used the interface programs to populate the data warehouse," Kirkpatrick recalls.

In the ensuing three weeks, while Kirkpatrick’s staff was busily populating the warehouse database, users were trained in the tools for modeling and analyzing the data. Both users and MIS staff were pleased with the outcome.

According to Senior Programmer/Analyst Mark Davidson, "We used to have a report that would show sales in 2000 and in 1999. Different users wanted slightly different views of the data – emphasizing, say, the quantity of beef patties produced and sold, and to whom. We used to have to create custom reports for each and every user. Now, we help the user to set up a view. When they get the hang of the DataTracker analysis options, they can alter their own queries, get their own reports, view them online or print them. They can also drill down into the data, develop new queries and break down the data further in whatever ways are meaningful for them."

Kirkpatrick adds that the centralization of data in a warehouse eliminates the possibility of error, "With downloads and extracts, there is always the possibility for errors that result from different people using different information and being out of sync with each other. Now, the users can use the tools against a common database. They have more control over what they want and that has started them thinking about what exactly they do want to see. I think it makes them more productive."

Since the solution officially went live in June 2000, DataTracker has provided Zartic’s production teams with new and more efficient ways to look at data on purchasing, raw materials, inventory, order history and line downtime. Where users once had to shuffle through hundreds of pages of production information, the DataTracker solution now enables them to quickly access key information from the data.

In most cases, what users are looking for is more productivity – which translates to more uptime on production lines. In a food safety-driven industry, such as meat processing, effective product scheduling is a key determinant of line efficiency. Switching between a raw product and a partially-cooked or cooked product requires that the entire line be broken down for washing – which means significant downtime. The analytical solution enables production managers to carefully analyze downtime trends and schedule upcoming runs for maximum efficiency. In fact, Zartic has identified nine different kinds of downtime on its production lines and can use Silvon to better monitor and avoid them all. Silvon, also, makes it simpler for users to explore what-if scenarios and look at past production and yield to help boost line productivity.

In addition to improving productivity, another benefit that Zartic expects to realize from its business performance management solution is improved management of what it calls "special commodities" contracts established between the vendor and schools and government entities. Under these contracts, the government supplies the meat for processing and contracts for a particular yield – for example, 70 percent of original poundage, accounting for bone removal, seasonings and breading, etc. Zartic must pay a penalty if it fails to meet specified yields. Through this system, school food service operations across the country are supplied with quality commodity processed products that are equivalent to commercially-available products and that meet established specifications for yield and portion control. While benefiting both the vendor and the consumer, the arrangement puts added pressure on Zartic to precisely track and control problems or trends that may affect yields.

Silvon helps Zartic walk the fine line between profit and loss on these special contracts by maintaining tight control over production variation. In fact, Zartic has set up 15 customized views and specialized categories tailored to commodity customers. By making the views flexible and user-customizable, Zartic has built in the ability to address the constantly evolving requirements and changing terms related to these contracts.

Zartic’s business intelligence system is currently in operation at three Georgia locations linked by a T1-based WAN, and its sales force is next in line for training. "One thing we’re very interested in tracking is why sales drop off," says Kirkpatrick. "If a customer was buying a certain amount of product six months ago, and all of a sudden we’re not seeing those sales, we want to know why."

With Silvon, Zartic sales teams can easily link sales trends to production or other events. They can identify any correlations between a change in sales and a new product introduction, a new salesperson or production facility, or a new competitor. Such information is expected to help the company to identify opportunities to strengthen customer relationships by determining the real issues affecting purchases – and to be proactive about protecting sales.

In the future, according to Kirkpatrick, access to the Zartic data warehouse and DataTracker analytical tools will be extended to salespeople in the field. "Together with the DataTracker tools, we, also, licensed Silvon’s remote access package for on-the-road data retrieval via laptop and a virtual private network (VPN) connection across the Internet." The solution will help to realize the vision of business intelligence at Zartic: to put actionable data as close as possible to the decision maker who needs it.

In the final analysis, Silvon’s business performance management solution is, not only, helping Zartic automate and streamline how it accesses and uses data, it, also, provides a flexible platform for addressing issues unique to the meat processing business. As more of its users get up and running, Zartic expects to find a growing number of ways to make its operations, and its chicken nuggets, more efficient.

Jon William Toigo is an independent consultant and author of The Holy Grail of Data Storage Management. He can be reached at jtoigo@intnet.net