Big Beer Uses Big data
The beer business is one that has always had a certain amount of “feel” to it. From the selection of the best hops to the art of brewing and forming a recipe to the right brands and packages to put into a convenience store or a bar to sell the most product, the experience of the farmer or the brewer or the sales person has carry a substantial weight. It is an industry that will likely never be fully automated as it is, at its heart, and people and a relationship based business. But that hasn’t stopped the world’s largest brewer, Anheuser-Busch InBev, the owner of brands like Budweiser, Stella Artois, Michelob Ultra, and Goose Island, from deploying artificial intelligence and machine learning to a number of their processes to improve product quality and sell more beer.
“The question is not going to be where we deploy AI, but where is it not going to be deployed, because we see it in so many different fields,” said Tassilo Festetics, ABI’s Global VP of Solutions . The company has deployed the technology to a number of areas already. A few of them can be listed below:
- Smart Barley – the platform started and managed by ABI aggregates farming data from all of its participating barley growers across 7 continents. The aggregation of massive amounts of data has allowed for algorithms to develop that can predict things like crop yields, weather affects, and optimal growing patterns to help improve overall yield and profitability for both the producers and the buyers (ABI.) [2,3]
- K-Filtration – The K Filter is an important step in the brewing process, providing a final “filtration” step near the end of the process that ensures the product is up to optimal quality standards. ABI has partnered with “Pluto7” and Google Cloud that uses machine learning to improve the accuracy of the filtration process, “which helps to reduc[e] costs, increase efficiencies, and perhaps most importantly for beer aficionados, ensure taste. The Pluto7 solution combines TensorFlow, Cloud Machine Learning Engine, Cloud SQL, and BigQuery.” 
- LOLA – Intelligent Sales Recommendations – in the past, sales reps would prioritize brands and packages based on what their boss told them to push, what was selling well in the rest of their territory, or what the retailer wanted. Now ABI has developed a proprietary tool known as LOLA, which aggregates zip code level market share data, account level volume data, and existing in account sales to create intelligent, algorithm driven recommendations and sales objectives for brand / package assortments that will maximize overall portfolio sales for the account. Sales reps are compensated can be compensated based on their execution against these objectives.
It is clear that ML has endless application for ABI and the rest of the beer business, but getting people internally up to speed was not easy. Festetics said “In this course about machine learning you could hear a pin drop after the lesson, because everybody was still processing. And that’s, I think, the important part — that you really continue learning and you continue to build those capabilities inside of your company.” . Integrating new technology into a business where production has traditionally been viewed as an art and sales have been done by “gut” is a process that takes time. It started with a further reliance on big data to make decisions. That data was, however, still reliant on the utilization and interpretation by humans. ML and AI takes this big data to a new level – truly removing the human decision making from processes that have a lot of pride attached to them. The true “art” comes from being able to explain to employees that they technologies are not replacing their capabilities, but simply enhancing their ability to do what they want to do anyways – produce the best beer and get it into the hands of the most people.
Student comments on Big Beer Uses Big data
Thanks for the article, Kyle. I definitely agree that major elements of this industry will remain with people itself but I do see where AI can make positive differences that are actually welcome by the end customers. The LOLA application seems like something we would have enjoyed in the Beer Games simulation last year as I think this could really help supply chain predictions as well as the intended sales recommendations.