Case Studies

Milking Demand

by Monica Mehta

Land O’Lakes uses Oracle’s Demantra software to anticipate market demand and increase margins.


Land O’Lakes

    Headquarters: Arden Hills, Minnesota

    Industry: Consumer goods

    Employees: 9,000

    Revenue: US$12 billion in 2010

    Oracle products: Demantra Demand Management, JD Edwards EnterpriseOne solutions, Oracle Transportation Management

Barry Libenson

Vice President and CIO

Length of tenure: Two years

Education: MBA from Duke University’s Fuqua School of Business

Personal quote/mantra: “Always do the right thing, even when it is not popular.”


Since the 1920s, the image of a seated Native American maiden holding a package in her hands has become synonymous with the word butter. That’s when the Minnesota Cooperative Creameries Association was formed, later renamed Land O’Lakes. The cooperative established a new standard of quality in the dairy industry with the systematic inspection and certification of butter from its member creameries. Today, the co-op is one of the largest producers of butter and cheese in the United States, handling 12 billion pounds of milk annually and generating revenues of US$12 billion.

Land O'Lakes

Launch the Slideshow

Despite that generous slice of the U.S. dairy goods market, Land O’Lakes’ dairy business is subject to extreme shifts in seasonal demand. When colder temperatures drive Americans indoors and the winter holidays give them reason to fire up their ovens, production of Land O’Lakes butter spikes by as much as 65 percent. During November and December, grocery stores around the country require 1 million pounds of butter per day to meet customer demand.

Historically, this fluctuation in demand required planners to stock excess inventory to make sure they could meet customers’ needs and service levels. But in 2010, Land O’Lakes deployed Oracle’s Demantra Demand Management to give production and deployment planners the ability to optimize production planning, significantly reduce inventory on hand, and effectively route products to markets where demand is highest—all while maintaining high service levels. These improvements in operational efficiencies have led to better use of raw materials (in this case, perishable milk), increased working capital to fuel growth, and higher profit margins.

“Figuring out how much inventory we need to meet customer demand—and understanding where product will be needed at a specific point in time—is critically important to the business,” says Barry Libenson, vice president and CIO at Land O’Lakes. “The Demantra solution is the key driver that we use in order to make that happen.”

Responding to Demand

While the Land O’Lakes brand of butter is nationally recognized, the company is not limited to producing dairy products. With the acquisition of Purina in 2001, Land O’Lakes became a major producer of premium feed for animals—a business that nicely complements the cow-powered dairy operation. The company also owns Winfield Solutions, an industry-leading farm seed business and the #1 distributor of crop protection products in the United States. Demantra software also provides insight into the ebb and flow of the seed business.

But it’s the seasonality of the dairy foods retail business that has always been a forecasting challenge for Land O’Lakes. And the stakes can be high—the butter and cheese production process has multiple steps to get from the dairy farms to the grocery. Once a stick of butter is made and packaged from cream produced at one of Land O’Lakes’ butter plants in California and Pennsylvania, it is quickly shipped to a distribution center. From there, it is shipped again to the customer’s distribution center, and then forwarded on to grocery stores, food service companies, and restaurants.

Because dairy products have a limited shelf life, this process must be as efficient as possible—so keeping enough finished product in inventory is essential for making the business successful. However, keeping too much in inventory introduces two challenges. First, excess inventory that does not make it to market eventually has to be dumped, taking a bite out of the company’s profit margins. Secondly, maintaining a product surplus shifts valuable corporate resources into inventory on hand, limiting capital that could otherwise be invested in other business needs.

“The more working capital we have that’s not tied up in inventory, the more advantageous it is for us to operate,” says Libenson. “Decreasing inventory can mean millions of dollars that we can use for acquisitions or other growth initiatives.”

Land O’Lakes executives sought a demand-sensing tool that could help them forecast short-, medium-, and long-term demand more accurately. Having implemented Oracle’s JD Edwards EnterpriseOne enterprise resource planning (ERP) system, management added Demantra software to extend the ERP functionality.

“We were striving for a best-in-class tool that would get us to the top quartile or better in inventory management,” says Libenson. “One day of inventory for us is millions of dollars of potential value, so anything that allows us to push product upstream or to have a higher degree of accuracy around inventory has a huge impact on our business.”

Connecting to the Future

Getting the most value from a lean inventory is the goal of the Demantra system. And that all begins with smart use of the dairies’ raw materials. “A major aspect of demand planning is how to most effectively leverage the milk that we have,” says Libenson. “How much milk gets turned into butter versus cheese? When it’s turned into cheese, what kind of cheese are you going to produce? A number of variables need to go into these decisions.”

That data lives in Catalyst, the Land O’Lakes JD Edwards EnterpriseOne ERP system. An extract, transform, and load technology routes the data into the Demantra system, capturing core information about shipment history; open customer orders; and key reference master data for products, customers, and related hierarchies. From there, Demantra software uses sophisticated statistical modeling and forecast consumption, order lead time, and consumption allocation algorithms to determine a very accurate forecast, particularly for the next four weeks. The planners use JD Edwards EnterpriseOne’s product and deployment planning system to plan production and deploy product so that the right product is available for the right customer at the right time.

“We are improving all parts of the forecasting accuracy,” says Yone Dewberry, vice president of logistics and planning at Land O’Lakes. “The better you can forecast, the better can make those decisions.”

The long-term reports, which forecast demand 18 to 24 months out, help managers make sure their milk sheds are going to yield enough milk and give them time to plan which type of product will get made, be it butter, cheese, or milk powder. The medium-term reports, which forecast demand two to six weeks out, drive the initial production plan. The short-term reports determine which customers will need shipments within the next two weeks. Supply chain managers rely heavily on these reports to make decisions about which Land O’Lakes facilities should be involved in the production and distribution of the product. These decisions have significant impact on the company’s margins, because reducing the distance between the factory and the customer translates into real cost savings. (See sidebar “Oracle Transportation Management.”)

“We have product that ships to the East Coast, the middle of the country, and the West. For the most part, we want to produce product in the location where we need it, because it’s a lot less expensive to ship locally,” says Libenson. Knowing exactly where the product needs to go geographically allows us to do transportation planning as well, which has an incremental benefit of saving on fuel and transportation costs.”

As a result of the improved forecasting accuracy, Land O’Lakes managers have also been able to decrease the amount of excess inventory, which has led to higher profit margins. “You don’t end up with inventory that you have to get rid of,” says Libenson. “You want to have just enough product on hand to meet the demand in order to capture the most value.”

Icing on the Cake

Based on the results from using the Demantra tool, Land O’Lakes is able to store and sell products much more efficiently, cutting total inventory of finished product on hand by four days, which adds up to an inventory reduction in the millions of dollars. “In the winter months, when you’re talking 1 million pounds of butter saved per day, that’s a lot of saved capital,” says Libenson.

But narrowing inventory on hand has its own risks. When available inventory is reduced, there is always the threat that spikes in demand will lead to stock-outs or interruptions to delivery. However, with the new production and inventory management system, managers were able to raise their already high service level for customers. “The tool worked really well for us, and we’re really happy,” says Mark Thome, vice president of information technology at Land O’Lakes, who oversaw the Demantra IT implementation. “The users are delighted with the tool—they are actually giddy using it.”

Libenson says Land O’Lakes is in the top 10 percent of the industry in terms of inventory management, because it was able to go from a 40 percent to a 70 percent rate of short-term forecasting accuracy. Most companies are only able to maintain a 50 percent rate of short-term forecast accuracy, he says. Land O’Lakes worked with Demantra developers to develop and improve this capability in the tool for the consumer packaged goods industry.

“Now, we have a 70 percent chance of getting the right product to the right places, whereas before we had only a 40 percent chance,” says Dewberry. “With the collaboration of our IT team and the Oracle development team, we now have a tool that allows us to compete with the leading companies that forecast on demand sensing.”

This all translates into savings of operating capital that can be redeployed to expand the business. For example, Land O’Lakes management is looking at expansion into international markets. The vast majority of the company’s US$12 billion yearly revenue is domestic, and Libenson sees a huge opportunity in exploring rapidly developing markets such as Brazil and China. “We want to leverage our equity and balance sheet, not only to make improvements domestically but to expand our operations internationally,” says Libenson. “There’s a gigantic opportunity in the international markets, in terms of our seed business as well as our dairy business. All of that is capitally intensive and will require major investment, but the potential returns are significant.”

Libenson says the Demantra system has become a “primary linchpin” for Land O’Lakes’ business operations, because the business can trust the tool to deliver reliable, actionable information.

“The Demantra solution is a best-in-class predictive analytics tool for demand sensing and planning,” says Libenson.

“It has sophisticated data modeling capabilities, and it offers the best predictive analytics available. The confidence we have in the information it provides allows us to manage our inventory in a way that we never would have been able to do without the platform.”

Photography by Shutterstock