billion dollar baby

For some of us, an impulse buy at the hardware store means a light-up keychain. For others, apparently, it's a riding mower or utility tractor outfitted with a 6.5-bushel rear bagger, 48-inch front blade or 12-volt oscillating fan.

You read that right. According to someone who should know—Loren Troyer, director of order fulfillment for Deere & Co.'s Commercial and Consumer Equipment Division—most of the company's riding mowers, garden tractors and ATVs are bought on impulse by customers who drop by a hardware store for a hammer or set of hinges. Once that shiny green riding mower or tractor catches their eye, however, hammers and hinges are quickly forgotten.

Those impulse purchases also tend to be highly seasonal (two-thirds of Deere's annual retail sales occur between April and July). Taken together, those two factors pretty much explain why Deere's dealers are eager to stock as many tractors and mowers as they can store. And in the past, that's exactly what they did—with the company's blessing. To encourage dealers to stock (and by extension, sell) as many vehicles as possible, Deere offered them free financing.

As much as the dealers may have liked that arrangement, not everybody was happy. The finance people in particular had begun to question the wisdom of tying up so much money in inventory. "[W]e basically encouraged our 2,500 dealers in North America to stock as much as they could," says Troyer. "[A]s a result, two-thirds of our entire assets as a division consisted of finished goods either at the warehouse or at the dealers."

Translated into dollars, those inventories represented a whopping $1.4 billion in 2001. And all indications were that inventories would continue to swell. Deere's own growth projections showed that if it continued on its current course, the division would be carrying as much as $2 billion in inventory in four years' time.

Crunch time
Faced with those projections, the company realized it was time to move forward with an inventory optimization project. In 2001, it began a search for software powerful enough to optimize inventories at various stages of the supply chain and on through to the showrooms of its 2,500 dealers. Specifically, what Deere needed was software that would help it determine optimal stocking levels for its plants, warehouses and dealers, balancing the desire to keep inventories to a minimum with the need to maintain sufficient stocks to avoid hurting sales.

It didn't take long for word to reach Deere of an emerging company, SmartOps. And what it heard captured its attention: The new company reportedly specialized in sophisticated supply chain optimization, and its program appeared to be particularly well suited to a multistage supply chain like Deere's.

If the company was new, its basic premise was not. The launch of the Pittsburgh-based SmartOps actually represented the culmination of a decade's worth of research by its founder, Dr. Sridhar Tayur, a professor of operations management and research at Carnegie Mellon University (CMU), and his colleagues at CMU.

What they were bringing to market was no less than a revolutionary approach to inventory optimization. Unlike the typical inventory software of the day, which essentially assumed that nothing would change once the plan was created, the SmartOps model was designed to take uncertainty into account—reflecting the real-life potential for floods, port congestion, labor disruptions, hurricanes, and so forth.

To do that, Tayur and his colleagues created algorithms that took traditional supply chain information (lead times, historical demand, growth projections) and combined it with data on unexpected occurrences, including how likely they were and how they would affect supply and demand. As difficult as that may sound, the researchers felt they could do nothing less. "The modeling has to represent the complexity of the real supply chain and the software must be robust enough to reflect real world conditions," says Martin Barkman, SmartOps' senior vice president for commercial operations.

Getting dealer buy-in
In 2001, Deere embarked on a pilot program with SmartOps to model different ways to reduce its inventory by 50 percent. Some of the information fed into the model was standard stuff: historical data on dealers' prior sales, dealers' projections for future sales, desired customer service levels, the company's growth projections, lead times, and shipping frequency, for example. Other information—like data on variability— was decidedly out of the ordinary.

The result was a detailed inventory model that forecast how much inventory Deere would need—no more and no less—and where that inventory should be stored. "The overall goal," says Barkman, "is to drop inventories, not service." But executing on the model would not simply be a matter of cutting inventories. Early in the project, Deere's managers had realized that the initiative would also require them to overhaul their distribution and logistics processes. The company could hardly ask dealers to slash their stocks without assurances that it would be able to whisk replenishments to them if the need arose. "It's actually not [simply] an issue of what we stocked," says Troyer, "but whether we could get product to our dealers when they needed it. We needed to improve our order filling and delivery."

Providing those assurances wouldn't be easy. The division's on-time delivery record was not likely to inspire confidence. In 2002, for example, it had only managed to get merchandise to dealers when they wanted it 50 percent of the time. Its record for delivering merchandise on its promised date wasn't much better—63 percent.

Supply chain overhaul
Still, Deere attacked the project with gusto, setting aggressive goals for improving order fulfillment and delivery performance. One of those goals, for example, was to cut the time it took to respond to dealer demands to two days. To do that, it has begun hiring third-party logistics service providers to stock some products closer to dealers to cut down on replenishment times (and transit costs).

Another goal was to boost efficiency. Here again, the SmartOps optimization tools proved helpful. Deere used the software to optimize its factory-to-dealer transportation. "You can't change the physics of moving product from Point A to Point B, but you can move it smarter," says Troyer. By loading its trucks more efficiently,Deere has improved truck utilization by 20 percent. Troyer projects that those savings alone will pay for the third-party services.

Through these and other initiatives, Deere has cut its order-to-delivery cycle from several weeks to five to seven days. Its on-time delivery performance has skyrocketed as well. The division that once struggled to deliver just half its shipments when the dealers wanted them now boasts a success rate of 88 percent. Likewise, the division now makes good on its delivery promises not 63, but 93, percent of the time.

With such consistent improvements, Deere had little trouble convincing dealers to cooperate with the new inventory program. Over the next few years, it gradually reduced inventory to the levels considered optimal for each dealer. In the meantime, it has adopted a new financing program designed to discourage dealers from carrying excess inventory. Today, the division finances only the recommended amount of stock for each dealer (as determined by the software). Dealers can purchase more vehicles and accessories if they wish, but they must finance those purchases themselves.

Less is more
It's fair to say Deere & Co. looks at inventory—and indeed, its supply chain—in a whole new light these days. "Historically, we had to increase our inventory to support higher sales," says Troyer. "But now we realize we have to turn our inventory faster and be more flexible. Today, we are much closer to having the right amount of stock at the dealers and ... in our warehouse for what we will need to replenish for the next couple of weeks."

That's not to imply that Deere rigidly adheres to the original SmartOps model's stocking recommendations. Rather, the SmartOps software continuously evaluates and adjusts inventories based on current sales and other factors. "Inventory reduction is a journey," says Troyer, "not a destination."

How far has Deere come on that inventory-reduction journey? Quite a ways, it seems. Incredibly, the program has produced inventory savings that approach the GDP of a small country—say, an Andorra or Guinea-Bissau. In 2001, the Commercial and Consumer Equipment Division carried $1.4 billion in inventory and projected it would be carrying $2 billion by 2005. Today, it maintains a total inventory of just $900 million—$1.1 billion less than its original estimate for 2005.

And Deere is saving more than just inventory expenses. If the division had continued on its original course, it would have been forced to expand its infrastructure to accommodate the mountains of inventory. At the very least, says Troyer, Deere would have had to enlarge some of its distribution facilities. But so far, that hasn't been necessary.

The inventory savings have also helped offset rising materials costs, like the steel and plastics used to make tractors. That alone has gone a long way toward helping the division and its dealers remain competitive in a tough market. "Our dealers now understand that," says Troyer, who acknowledges that the project's success depended heavily on the dealers' cooperation and support. "It's such a mindset change from the days when our philosophy was to bury them in inventory."