Harvesting grapes has traditionally been back-breaking labor, sometimes literally.
Grapes are usually grown on t trellises, which stand low to the ground. Workers must get on their hands and knees in order to pick the grapes that grow
In 2007, designers at Sun World, a Bakersfield, Calif., farm, came up with a new type of trellis, one several feet higher, that would allow workers to stand while harvesting grapes, according to Steve Greenwood, the company’s director of budgets and reporting. It would certainly be easier on the workers’ backs, but would it also make them more efficient?
Until recently, Sun World execs wouldn’t have had any reliable way to know. The company had been using Cognos TM1 for years for financial reporting but hadn’t thought to turn the power of data analytics software to their vineyards, according to Greenwood.
That changed when a new ownership group took over in 2006, Greenwood said. The new owners wanted to track a number of key performance indicators (KPIs) to better manage Sun World’s operations, and it was Greenwood’s job to make it happen.
He didn’t have far to look. He turned to analytics software from the company’s incumbent financial reporting vendor, Cognos (since acquired by IBM), and set to work. After integrating labor, energy and pricing data from a number of disparate databases into the Cognos analytics engine, Greenwood wrote and applied several algorithms that would give the new owners the numbers they were looking for.
Among them was the number of boxes of grapes harvested per hour across its multiple vineyards. Specifically, Greenwood began tracking and analyzing harvesting rates among the company’s teams of workers to see which teams were doing the best and which were falling behind.
Execs used data analytics tools for short- and long-term profitability
With the new analytics capabilities, execs were able, on a daily basis, to identify teams that were less efficient than the average and began instituting incentive plans to improve their performance.
Timeliness was crucial, Greenwood said. In the grape business, the harvesting season lasts only about six weeks, from early July to mid-August. In order for analytics to make an impact on the current crop, it is needed quickly, before the short season is over.
Data at Sun World is integrated into the Cognos in-memory analytics engine every five minutes, Greenwood said, so farm managers can analyze it first thing each morning and any time throughout the day. So if a farm manager identifies an inefficient team, he or she can take immediate action. Such insights gleaned in September, for example, couldn’t be applied until the next harvest season.
But execs were also interested in more long-term analysis. They wanted to introduce the new trellis but weren’t sure whether the return on investment in the new equipment would be worthwhile. Analytics gave them the answer.
The company began testing the new trellis and, with their new analytics capabilities, started to compare the number of boxes harvested per team using the new trellis to those using the old, low-to-the-ground trellis. Turns out workers harvesting grapes from the newly designed trellis were significantly more efficient than their co-workers picking from the old trellises.
The analysis gave the execs the validation they were looking for. Since deploying the newly designed trellises, the company has reduced the cost of each box of grapes harvested by $0.50, Greenwood said. With 6 to 7 million boxes harvested every summer, that’s an annual savings of $3 million or more.
“We’ve been able to validate the cost savings for these trellises,” Greenwood said.
Sun World’s new analytics capabilities also help the company measure the effectiveness of various irrigation technologies, he said. The company has reduced its water usage by nearly 9% since 2006.
“We kind of guessed at [those things] before, but we’re confident in the numbers now,” Greenwood said.
Selling execs on data analytics software wasn’t easy
But while the company is enjoying the fruits of its analytics labor now, some Sun World execs were skeptical at first, he said.
“One of the biggest things we’ve had to overcome is fear,” Greenwood said.
With little to no experience with data analytics, some execs simply didn’t trust the new numbers. They lacked “technology savvy,” as Greenwood put it. Unfortunately, there was no quick fix. The data analytics software simply had to prove itself for resistance to breakdown, a process that took time.
Greenwood’s patience eventually paid off. As the benefits to the bottom line became apparent, adoption of analytics technology at Sun World increased, and Greenwood has since tailored the analytics software to meet the needs of various user groups.
Executives, for example, use high-level dashboards to track the company’s sales and financials. Farm managers, meanwhile, track and manage operations with daily, interactive crop cycle reports.
IBM’s Turbo Integrator feeds Sun World’s data from multiple source systems into the Cognos analytics engine, which then generates various reports and facilitates further analysis by end users, Greenwood said.
Of course, Sun World is an early adopter in an industry that, in some cases, has been using the same farming techniques for generations and is not always keen to adopt new technology such as data analytics. But Greenwood thinks that is likely to change.
“Corporate farming is a very tough, competitive business,” he said. “I think companies that don’t engage this type of analytics technology are going to find themselves in trouble.”