Structuring a big data strategy
A comprehensive collection of articles, videos and more, hand-picked by our editors
By 2050, the Food and Agricultural Organization of the United Nations (FAO) predicts the world's population will hit 9 billion, a 34% jump over current totals. The increase will have serious implications for water and energy resources and almost certainly for the agricultural industry. To keep up with population growth, the FAO believes farmers will have to produce more food in the next 50 years than they have in the last 10,000.
"This is all a productivity game," explained Karen Oerter, director of marketing and customer insights at WinField Solutions LLC, a wholly-owned subsidiary of Arden Hills, Minn.-based Land O' Lakes. "If we're going to feed those people, it's all about understanding what you're going to need to do to get the right products on the right acre."
For Land O' Lakes CIO Barry Libenson,who recently left to become CIO at Safeway Inc., this is also an innovation game. In order to help the crop protection and seed distribution subsidiary target customers more effectively, Libenson and his IT team are wrangling "dirty data," or data that can contain errors. By combining internal customer data with reams of external data, his team is helping the business predict what crops will yield the highest value where -- down to the acre.
The new initiative not only involves integrating diverse data from disparate sources but also putting sophisticated and easy-to-consume analytics right into the hands of the company's internal sales and marketing team, as well as into the hands of the cooperatives and farmers it distributes to -- Land O' Lakes' external customers.
"Big data has given us the opportunity -- and the need -- to start thinking in new ways," Libenson said.
Digging through the dirt
That's saying something for an industry known to be rich in data but sometimes short on data-based action. While new farming equipment is outfitted with sensors, the data it generates has done little to change the day-to-day operations for the family farms that are Land O' Lakes' chief customers, according to Oerter.
"Do you know how framers transact? They don't carry their credit card, they usually don't have their wallet," she said. "They transact by walking into the co-op and saying, 'Hey, I need a load of fertilizer delivered over there in the corner of Grandma's 40 acres.' That's a transaction. You try to figure out how to put that in your database."
- WinField Solutions is a wholly-owned subsidiary of the dairy cooperative Land O' Lakes.
- WinField Solutions is the largest independent distributor of seed and plant health products in the United States.
- The project of integrating external and internal data sources was the winner of this year's Gartner Business Intelligence Excellence Awards.
- In 2012, WinField Solutions saw a 60% increase in earnings -- one year after deploying new analytics tools.
Other data sources suffer from a similar lack of standards, Libenson adds. Take the crop-yield data U.S. farmers report to the U.S. Department of Agriculture in order to receive subsidies. By piecing together yield data, a non-relational data source made publicly available by the government, Winfield Solutions can uncover what's growing where. Figuring out who is doing the growing, however, isn't simple, as the same reports do not require farmers to enter their names.
"One of the running jokes is that Bruce Springsteen apparently plants more corn than anyone else in the world, which tells you a little bit about the average age of the growing community," said Libenson, but not much else about the identity of those growers.
"That's sort of the nature of unstructured data: The lack of precision, the fact that there's likely to be erroneous information stuck in with the good stuff, that there are no strict restrictions," he continued. "It can make for a lot of work to massage it or get it into a format you know is reliable and meaningful."
In the past, this kind of "messy, ugly" data, as business counterpart Oerter calls it, provided limited insight for the business. Today, Libenson and his team are making better use of it by triangulating crop yield data with GPS coordinates -- another external data source -- to tease out who the grower is.
"We were not extracting information off of the Internet previous to this, other than simply looking it up and loading it into an Excel spreadsheet," Libenson said.
Tactics such as using geospatial data or even satellite imagery are known as precision agriculture, and while this practice benefits the farmer, it also gives the WinField Solutions' sales and marketing team a more detailed picture to help target product sales.
"Time is everything," said Oerter. "If you can be ahead of the competition, give your sellers confidence, keep your growers happy, everything works out fine."
Reaping what you sow requires having the right data mining tools
Leveraging external data sources -- such as social, census and meteorological information -- to drive deeper insight has been a Land O' Lakes vision for more than three years, conceived initially by the chief operating officer, said Oerter and Libenson. Because so much of external data doesn't fit easily into a relational database, the big question was how the IT department could deliver.
Libenson and his team eventually found Endeca Latitude. Acquired by Oracle in 2011, the data discovery platform can pull together non-relational and relational data sources on the back end and provide an easy-to-use data visualization tool on the front end. The investment was a game changer for Land 'O Lakes, according to Libenson, and not just because the technology enabled deeper data insights.
"Like a lot of technology organizations, I think historically, the business didn't necessarily see us as delivering things that were going to empower them" he said. "This really changed the relationship between the technology organization and the business because we were delivering something to them that was unlike anything they'd ever gotten."
Today, WinField Solutions' sales and marketing teams are armed with company-issued mobile tablets and can pull up these visualizations and build a window into buying behavior right in the field.
"We buy a lot of iPads in the WinField division and a lot of Android tablets," Libenson said. "As a matter of fact, we had a significant planned increase in mobile technology spend from 2012 to 2013 partially driven by tablet absorption and the need for additional bandwidth for the sales organization."
Planting new seeds with customer-delivered analytics
The on-the-spot analysis can get quite sophisticated, Oerter said in a recent video interview she did with Oracle. "[The sales team] might be able to go in and identify customers that purchased a certain seed, and then maybe based on the current weather patterns, might want to consider a certain fungicide or insecticide that's associated with that crop."
This really changed the relationship between the technology organization and the business because we were delivering something to them that was unlike anything they'd ever gotten.
Barry Libenson, CIO, Land O' Lakes
And, for a group of customers who have long relied on their farming smarts to survive, the data is proving compelling -- at least once they see it.
"When I took the product out to one of our really big co-ops, [the manager] said, 'We don't need it,'" Oerter said. He doubted the data, and so Oerter took him through a demo, quickly identifying that one of his best customers hadn't ordered the half-million dollars in seed he did the previous year.
"Everyone was waiting for the customer to come in," she said. "What this analytical tool helped us to do was to go out and, with confidence, make the call."
According to a 2012 Gartner Business Intelligence platform customer reference survey, Land O' Lakes' WinField Solutions is far from alone when it comes to bringing data to the customer: Thirty-six percent of the 1,364 respondents indicated they already had externally facing business intelligence applications.
Kurt Schlegel, an analyst at Stamford, Conn.-based Gartner Inc. and lead author on the survey report, believes the trend will grow stronger. "This percentage will increase substantially as delivering analytics externally becomes an expected component of every business relationship. Needless to say, this also represents a major career opportunity for CIOs and IT leaders," he wrote.
Putting analytics into the hands of customers provides a layer of transparency that Schlegel called the future of analytics. "There's going to be no place for poor performance or inefficiencies to hide, and I think that's going to relevant to every industry," he said. Schlegel believes Land O' Lakes is headed in the right direction, and Libenson couldn't agree more.
"With big data applications, the goal is to empower the user to solve a fairly complex problem based on a lot of different data and without having to rely entirely on the IT organization all of the time," Libenson said. "I think we've been successful … in doing that.
"At the end of the day, I'd rather my people spend time doing other things than having to build reports and dashboards all day long," he said. "They have a lot to do, and tools like this really, really go a long way toward freeing up our resources to work on other stuff."