When it comes to the business of agriculture, technological innovation doesn’t exactly sprout every year. Arguably, the biggest high-tech development hit the industry in 1892, when John Froelich invented a machine that revolutionized farming, replacing hundreds of human workers with a gasoline-powered mechanical alternative. More than a century later, Froelich’s technology remains the focal point of the farming industry. Its name? The tractor.
Perhaps this is why two recent IT innovations at agribusiness giant Royster-Clark, a CIO 100 honoree, are so impressive. The first, a Web services application designed to expedite customer credit approval, lets the company authorize purchases within seconds, delivering credit decisions in minutes. The second, a distributed computing methodology aiming to eventually run all billing functions overnight, will help the company avoid a million-dollar hardware upgrade that would take months to employ.
At a time when Royster-Clark’s IT budgets were shrinking precipitously, the applications cost just over $30,000 combined. The investments enabled the company to use its existing IT assets more effectively and position itself for growth. The endeavors have also transformed IT at the $900 million company from a reactive mishmash of staid systems to a vibrant and dynamic business partner that tackles problems head-on. “For a company in [the agribusiness] industry, we’re now chewing on some pretty serious stuff,” says Robert L. Paarlberg, managing director of IT at the Collinsville, Ill., facility. “From our perspective, this is the only way to get ahead.”
Better Customer Service
The first of these efforts began in early 2001, when Royster-Clark executives gave Paarlberg the go-ahead to build a low-cost alternative to accelerating the customer credit application process at 350 retail locations across the country. For years, this process had been laboriously slow, taking about two weeks between application and credit-to-company response. Customers applied for credit using paper forms that were mailed to credit managers, faxed to one of two credit bureaus, faxed back to credit managers, mailed to headquarters, entered into a database and finally mailed back to the point of sale.
If a credit manager was out, forms could sit for days. If a credit bureau needed extra time to review an application, the process fell behind even more. Meanwhile, Paarlberg says, customers were forced to wait to see if they qualified to purchase fertilizer and other pricey agriculture products on credit. The system needed an overhaul, and he knew he was the one who had to change it. “The way we used to run this process was downright awful,” he scoffs. “Just about anything would have been an improvement, but we wanted to fix things right.”
To do this, Paarlberg enlisted the help of the then-fledgling Microsoft Professional Services, a subsidiary of Microsoft dedicated to developing the company’s .Net Web services applications. Paarlberg and Director of IT Betty Orr met with a team of Microsoft consultants and outlined the inefficiencies they wanted fixed. To Royster-Clark’s surprise, the consultants agreed to foot nearly half of the $60,000 price tag. Microsoft intended to use the project as a showcase, and Paarlberg wasn’t about to resist. Two months and $30,000 later, Royster-Clark had a Web-based system that eliminates paper altogether; it evaluates applicants automatically against a series of fixed requirements, files the results electronically with a credit bureau and sends a response to the point of sale within seconds.
Today, many customers can apply for credit and minutes later walk out of the store with as many bags of fertilizer as they need. If a customer’s application fails to meet enough of the system’s requirements, the point-of-sale terminal receives a message instructing the customer to contact a central credit manager to discuss other options. Paarlberg and Orr tout this process as the fastest credit review and approval program in the business and say that while they haven’t quantified cost savings associated with the program, simply having it sets Royster-Clark apart.
“This is a huge differentiator in the industry,” says Orr. “The payback is that it’s nice to give customers those credit limits and get them into our fold.”
Laura DiDio, a senior analyst who covers application infrastructure and software platforms for the Yankee Group, agrees: “An application like this speeds the time to market and automates everything securely,” she says. “This is a leading-edge application, and in any industry, that’s a plus.”
The Grid Plan
Around the same time that Orr and Paarlberg first met with representatives from Microsoft .Net, the duo was thinking up an innovative solution to another IT problem—printing billing statements. For years, Royster-Clark had used batch processing, but it was time-consuming and inefficient. Without informing any of his colleagues on the business side, Paarlberg set out to improve this process on his own. Today, though the company isn’t yet employing the solution, his methodology incorporates cutting-edge programming and the company’s existing assets to save time and money without sacrificing quality.
In the spring of 2001, company executives announced a letter of intent to acquire competitor Agro South—a move that would have added 140 retail locations and 40,000 new customers. Paarlberg and Orr understood that this kind of growth would have put monumental strains on the existing bill statement system, an archaic, monthly batch process that ran on a Sun Microsystems server running Solaris and Informix. They realized something had to be done.
The duo spent weeks evaluating options. Finally they determined that the necessary upgrades to compute nearly 150,000 monthly customer statements would require a new server, an expanded database and updated ERP tools. When the final tally for this option surpassed $1 million, Paarlberg and Orr returned to the drawing board and emerged with a different strategy. This second plan revolved around grid computing and outlined a strategy to parse out parallel computations to 405 different Royster-Clark computers while they idled at night.
“If grid computing is a steak, we were set to do a hot dog version of it,” Paarlberg jokes. “It was just a way for us to make use of an untapped resource to maximize efficiency.”
On paper, the plan was almost beautifully simple. Paarlberg, who was educated as a programmer, spent two weekends at home writing a Visual Basic application to generate complex bill statements composed of headers, line-item invoice details, discounts and appropriate taxes. The program would run as an executable behind the company screen saver; when a screen saver went on, the application would kick in and search directories on the Sun server for files designated by customer numbers. Upon finding those files, the program would download them, collect the appropriate billing information, factor in additional fees (late charges, discounts and so on), generate print files, rename them and return them to the same directories from which they came.
Network administrators could keep track of finished print files by referring to the file name itself—files distinguished by customer numbers still needed computation, while files with names did not and were ready for printing. The plan seemed flawless, down to a strategy to print the statements in batches, after hours. There was, however, one hitch—when the acquisition fell through after the terror attacks of Sept. 11, the grid computing plan was put on the shelf, where it has remained ever since.
Today, according to Paarlberg, the methodology is so complete that it could be implemented with only minor tinkering at any time. Because Royster-Clark has no pending acquisitions, however, the grid endeavor is most likely months away from a full-scale rollout. This delay hasn’t stopped analysts from cooing, however. Nick Gall, senior vice president and principal analyst at Meta Group, was “shocked” to hear about the Royster plan, and says that while grid computing is commonplace in the pharmaceutical, bioinformatic and manufacturing industries, it is virtually unheard of in the antiquated agribusiness.
“It is highly sophisticated for a company in this industry to take a large problem and break it up into smaller ones,” notes Gall. “Most agribusiness companies are still wrestling with mainframes, much less thinking about applications like these.”
Of course at this point, Royster-Clark’s plan for grid computing is just that—a plan. While the methodology is inarguably unique, Paarlberg himself is quick to note that until the company acquires a significantly critical mass, there won’t be any substance to its distributed computing approach. Still, Paarlberg takes comfort in what he perceives to be a cost savings of nearly $1 million—he estimates he spent $2,000 in man-hours writing the executable, and figures that having the application on hand will eliminate the need for new hardware down the road.
The Web Future
Ultimately, once the grid computing effort is up and running, Paarlberg says, it will get better with age. First, as customers grow accustomed to receiving such comprehensive invoices, Royster-Clark programmers will incorporate additional information, such as regional advertisements or targeted messages. Next, once the process generates these expanded print files, technologists can write a program to convert individual statements into encrypted webpages, enabling customers to access their accounts online. Ultimately, says Web Systems Manager Julie Schottel, this technology could move the entire payment process onto the Web.
“Once we get our statements online, the next step is allowing customers to pay there as well,” she says. “This methodology has implications that reach far beyond simply utilizing the resources that sit dormant overnight.”
Other innovations lie ahead at Royster-Clark as well (see “Pinpoint Farming,” Page 106). The Microsoft .Net consultants working on another project for Royster-Clark, an enterprise back-office system upgrade that will ultimately link all systems—credit application, cash accounting and invoicing—to one master, Web-based customer database.
Finally, Paarlberg and his colleagues are brainstorming ways to improve inventory management. The fertilizer industry turns on inventories that are closely related to environmental and atmospheric conditions, so ebbs and flows can be unpredictable, making planning difficult. With the help of other business leaders, Paarlberg says he’s currently experimenting with a variety of inventory management systems that incorporate long-range weather forecasting, futures commodity trading and a variety of other factors. These systems might not be as revolutionary as the tractor, but they could help Royster-Clark stay on top of its inventory during a long winter. John Froelich would be proud.