Case Study$-$Decision making at Coca$-$Cola

The Coca$-$Cola $($Coke$)$ as the world's largest non$-$alcoholic beverage company, has built $15$ billion$-$dollar brands and also claims four of the top five soft$-$drink brands $($Coke$,$ Diet Coke, Fanta and Sprite$).$ Although it fell to the number$-$three spot in $2013,$ each year since $2001,$ global brand consulting firm Interbrand, in conjunction with Bloomberg BusinessWeek, has identified Coke as the number$-$one best global brand. Coca$-$Cola's executives and managers are focusing on ambitious, long$-$term growth for the company $-$ doubling Coca$-$Cola's business by $2020.$ A big part of achieving this goal is building up its Simply Orange juice business into a powerful global juice brand. Decision making is playing a crucial role as managers try to beat rival PepsiCo, which has a $40$ per cent market share in the not$-$from$-$concentrate juice category compared to Coca$-$Cola's $28$ per cent share. And those managers are not leaving anything to chance in this hot $-$ umm, cold $-$ pursuit!

In Coca$-$Cola's case, a glass of $100$ per cent OJ is possible only through 'satellite imagery, complicated data algorithms, and even a juice pipeline'. Standardisation is what it takes for Coca$-$Cola to make this work profitably. And producing a juice beverage is far more complicated than bottling soda.

Using what it calls its 'Black Book model', Coke wants to ensure that customers have consistently fresh, tasty OJ available $12$ months a year despite a peak growing season that is only three months long. To help in this, Coke is relying on a 'revenue analytic consultant'. 'Orange juice is definitely one of the most complex applications of business analytics,$\text{'}$ he says. To consistently deliver an optimal blend given the challenges of nature requires some $1$ quintillion $($that is $1$ followed by $18$ zeroes$)$ decisions.

There is no secret formula to Black Book, it is simply an algorithm. It includes detailed data about the more than $600$ different flavours that make up an orange and about customer preferences. This data is correlated to a profile of each batch of raw juice. The algorithm then determines how to blend batches to match a certain taste and consistency. At the juice bottling plant, 'blend technicians carry out Black Book instructions prior to bottling'. The weekly OJ recipe they use is 'tweaked' constantly. Black Book also includes data on external factors such as weather patterns, crop yields and other cost pressures. This is useful for Coca$-$Cola's decision makers as they ensure they will have enough supplies for at least $15$ months. One Coca$-$Cola executive says, $\text{'}$If we have a hurricane or freeze, we can quickly replan the business in $5$ or $10$ minutes just because we've mathematically modelled it$.\text{'}$

Discussion questions

Which decisions in this case information could be considered unstructured problems? Structured problems?