How Coca-Cola Controls Nature’s Oranges

Only one of the below launched a campaign to conquer the world… but how?

We think about big data as a fairly new invention, and the brands that come to mind are usually large or newly founded tech companies that harnesses the power of digital technology as the core competency of their business. What we often ignore is that many industries that have existed for hundreds of years have actually started tracking data much earlier on, and their ability to utilize and understand data is becoming even more fundamental to their continual survival.

implementation-of-sap-bi-in-coca-cola-16-638Coca-Cola is one of those companies that has been very conscious of the power of data analytics to create and capture value for their customers, their suppliers and vendors alike. As the world’s largest beverage company, with more than 500 brands and 3,500 products sold worldwide, Coca-Cola has about 250 bottling partners with 900 bottling plants, and employs over 700,000 system associates worldwide. This entails enormously complex production and operation systems that relies on a robust data analytics system to forecast supply and demand. Coca-Cola also built a system of Point of Sales data from retail channels like Walmart (Walmart alone contributes $4 billion in Coca-Cola sales annually) to build customer profiles, create centralized iPad reporting and enable cooperative planning, forecasting and restocking processes within their supply chain, based on these collective data.

On the value creation front, Coca-Cola also uses big data to manage quality control. One of the notable examples is how Coca-Cola leverages a complex algorithm called the Black Book model to control the production of orange juice to ensure that the taste of the juice is consistent all year-round, despite peak orange growing seasons of just 3 months. The Black Book model combines various data sets such as weather date, expected crop yields, satellite imagery, regional consumer preferences, cost pressures, detailed data on the 600 different flavors that make up an orange, and many other variables such as acidity or sweetness to advise Coca-Cola’s factories on how to blend the orange juice respectively to create the consistent taste, down to the pulp content. Bob Cross, inventor of Coke’s Black Book juice model, calls it “one of the most complex applications of business analytics. It requires analyzing up to 1 quintillion decision variables to consistently deliver the optimal blend, despite the whims of Mother Nature.”

Coca-cola Freestyle vending machine
Coca-cola Freestyle vending machine

Another important example is how Coca-Cola is using their omnipresent vending machines to gauge movements in consumer demand. In 2009, Coca-Cola invited Segway inventor Dean Kamen to help design the next generation of their vending machine. The result was the Coca-Cola Freestyle machine, which could dispense well over 100 combinations of carbonated and non-carbonated soft drinks. They also produced a corresponding mobile app, with more than a million downloads, that allows customers to name and save their favorite combinations and connect to the Freestyle machines to automatically pour them the drink. Having poured more than 5 billion servings and thousands of flavor permutations, the data they’re generating is a fountain of marketing knowledge in helping the beverage giant shape product offerings for itself and its foodservice customers. According to Jennifer Mann, VP-general manager of Coke Freestyle, “Before Freestyle, Caffeine-Free Diet Coke was available in less than 1% of our dispensers in the U.S., now with Freestyle it’s available in every dispenser, and it’s become a top-five brand in the afternoon daypart. So there was a huge unmet demand we were able to fill.”

It is evident that Coca-Cola had been successful in leveraging data to create and capture value, the challenge ahead definitely lies in the exponential increase and complexity of data that is available for them to analyze in relation to the data analytics talent they are able to recruit. Whether Coca-Cola can continue to stay ahead of the game relies on not only their brand equity, but their ability to not just look backwards and learn from data, but to turn this data into predictive analytics to anticipate future trends.

 

Reference:

  • http://sloanreview.mit.edu/article/coca-colas-unique-challenge-turning-250-datasets-into-one/
  • http://adage.com/article/dataworks/coke-freestyle-dispenses-insights-soft-drinks/295661/
  • https://datafloq.com/read/coca-cola-takes-refreshing-approach-big-data/425
  • https://en.wikipedia.org/wiki/Coca-Cola

Previous:

Nike: A Full Stack Athletic Performance Company

Next:

Counter-Strike: Global Offensive — One of Valve’s Data Habits

Student comments on How Coca-Cola Controls Nature’s Oranges

  1. I have only seen the FreeStyle machine in a very limited number of locations. Why do you think it hasn’t become more prevalent? Do you think that Coke was able to accurately predict the demand for the FreeStyle machines? Also, I’m curious what the impetus was for creating this machine. You mentioned that caffeine-free diet Coke was only available in less than 1% of dispensers. I would assume this meant their isn’t a strong demand for this drink; low availability would correlate with low demand. Thus, is it logical for Coke to make such drinks available?

  2. In theory, the FreeStyle machines are an exciting innovation, but yet in practice, they are pretty confusing to use and I have only seen one in the iLab. To be honest, they look much more exciting than they actually are. How many people actually prefer a mix of different drinks? I wonder where they exist and how the data is used in future decisions. I would like to see Coca-Cola address the customer demand for healthier drink options alongside the multitude of carbonated beverages.

  3. @Jennifer and @Jess — the FreeStyle machines are great, and I’ve seen a ton of them, particularly in movie theaters. I assume they’re not in wider distribution yet because they’re more expensive to install. The real value of these machines is not to allow customers to mix Sprite and Diet Coke (yikes!), but rather to gather really granular data on customer preferences by time of day, flavor mix (maybe Cherry Coke is most popular fruit Coke, but Orange Sprite is most popular Sprite beverage) at low cost — Coca Cola produces all of these syrups anyway, having them in one dispenser is easy, but making the investment to bottle and market a certain combination could be costly without the data to back up demand first.

Leave a comment