Everyone wonders about the secret formula for Coke. But at The Coca-Cola Company, whose brands like Minute Maid make it the No. 1 juice maker in the world, there’s a secret formula in an internal application called the “Black Book” that tells how to run the company’s global juice business.
According to Doug Bippert, Coca-Cola’s vice president of business acceleration, the “Black Book” is a Java-based application created with the help of outside firm Revenue Analytics, which combined the variables around agricultural fruit production, the taste of each juice brand, the orange-squeezing and containers, trucking, shipping and customer demand, to create a formula for what steps to take to synchronize Coca-Coca’s global juice operations.
The app pulls together what not long ago was knowledge held on spreadsheets and other means across at the various management divisions responsible for figuring out how to make a certain amount of juice — especially orange juice — every year. “Juice requires an agricultural supply chain,” said Bippert, about how the application plays a key part in guiding a multi-billion-dollar operation of staggering complexity. By contrast, making a soft drink like Coke is fairly simple.
“Mother Nature gives you different quantities and quality of juice each year,” Bippert said, speaking at the IBM Smarter Commerce Global Summit 2014 conference in Tampa this week. But a lot of research shows that consumers expect the brand they pick for their orange juice — and there are hundreds — to taste the way they expect each time. That means “master blenders” come up with formulas based on available oranges from Florida and elsewhere in the world to meet that expectation, and corporate buyers go out to get the quantities they can when available.
But at the end of the day, it’s about quality, says Bippert. And the ongoing taste-tests conducted by “expert sensory panels” with consumers indicate they are very picky about flavors and pulp qualities, something Coca-Cola tries to match through a scientific analysis of oranges especially.
Thus the “Black Book” came as a way to make contract buying and production “blend plan” decisions based on demand forecast to optimize supply-chain interactions. Dr. Jon Higbie, chief scientist at Revenue Analytics, said the data set is built using IBM’s ILOG Optimization Programming Language Studio and CPLEX decision-making toolkit. He said the app is one of the most complex optimization projects his company ever tackled, even more complicated than airplane production. “It’s been a blast building this thing,” he said.