As the world's largest beverage company, Coca-Cola serves more than 1.9 billion drinks every day across more than 500 brands, including Diet Coke, Coke Zero, Fanta, Sprite, Dasani, Powerade, Schweppes and Minute Maid.
Big data and artificial intelligence (AI) are the power of everything the company does – the global director of digital innovation, Greg Chambers, said : “Artificial intelligence is the basis for everything we do. We create intelligent experiences. Artificial intelligence is the core that makes this experience possible. “
WHAT PROBLEM IS ARTIFICIAL INTELLIGENCE THAT HELP RESOLVE?
Offering soft drinks worldwide is not a “one-size-fits-all-affair”. Coca-Cola products are marketed and sold in more than 200 countries.
In each of these markets there are local differences with regard to tastes, sugar and calorie content, marketing preferences and competitors with which the brand is confronted.
This means that in order to stay at the top in every area, it needs to collect and analyze huge amounts of data from different sources to determine which of its 500 brands is likely to be well received. The taste of their most famous brands will even vary from country to country, and understanding these local preferences is a hugely complex task.
HOW IS ARTIFICIAL INTELLIGENCE USED IN PRACTICE?
Coca-Cola serves a large number of drinks through vending machines every day. On newer machines, the customer will usually communicate via a touchscreen display, allowing them to select the desired product and even adjust it with different shots of different flavors. The company has started to equip these machines with AI algorithms to promote drinks and flavors that are likely to be well received at the specific locations where they are installed.
The vending machines can even change their “mood” depending on where they are located – with machines in a mall with a colorful, fun personality, those in a gym more focused on achieving performance and those in a hospital seem more functional.
Coca-Cola also uses AI to analyze social media and to understand where, when and how its customers like to consume their products, and which products are popular in certain places. With more than 90% of consumers making purchasing decisions based on social media content , it is essential to understand how his billions of customers are discussing and interacting with the brand on platforms such as Facebook, Twitter and Instagram. Coca-Cola analyzed the involvement with more than 120,000 social media to gain insight into the demographics and behavior of its customers and the customers who discuss the products.
Another application of AI was securing the proof of purchase for the company's loyalty and reward schemes. When customers were asked to manually enter 14-digit product codes on bottle caps in websites and apps to verify their purchases, the recording was understandably low due to the cumbersome nature of the operation.
To encourage more customers to participate in these schemes, Coca-Cola has worked on the development of image recognition technology that can verify purchases by taking a single smartphone photo.
WHAT TECHNOLOGY, TOOLS AND DATA WERE USED?
Coca-Cola collects data on local beverage preferences via the interfaces on its touchscreen machines – more than 1 million of them are installed in Japan alone.
To understand how its products are discussed and shared on social media, the company has set up 37 “social centers ” to collect and analyze data for insights using the Salesforce platform. The goal is to create more content that has been shown to be effective for generating positive engagement. In the past, the process of creating this content was carried out by people; However, the company has actively looked into the development of automated systems that create advertisements and social content based on social data .
It also uses image recognition technology to target users who share photos on social media, showing that they can be potential customers. In an example of this strategy in action, Coca-Cola targeted advertisements for its Gold Peak brand iced tea to those who placed images suggesting they enjoyed iced tea, or in which the image recognition algorithms saw logos of competing brands . Once the algorithms had determined that specific individuals were probably fans of iced tea and active social media users who shared images with their friends, the company knows that targeting these users at ads is likely to be an efficient use of their advertising revenue.
For purchase verification, the standard image recognition technology proved insufficient to read the low-resolution dot-matrix printing used to stamp product codes on the package. Coca-Cola therefore worked on developing its own image recognition solution using Google's TensorFlow technology. This made use of convolutional neural networks to enable machine recognition of codes that could often appear differently, depending on when and where they were printed.
WHAT WERE THE RESULTS?
By analyzing the data from vending machines with AI algorithms, Coca-Cola can better understand how the buying behavior of its billions of customers around the world varies.
It uses this to inform new product decisions – for example, the decision to launch Cherry Sprite as a bottled product in the United States was made because the data showed that this would probably be a winning initiative .
Using computer vision analysis and natural language processing of posts on social media, as well as an in-depth learning-driven analysis of statistics for social engagement, Coca-Cola can produce social advertisements that resonate more with customers and encourage the sale of its products.
By applying TensorFlow to create convolutional neural networks, scanners could recognize product codes from a simple photo, increasing customer engagement with Coca-Cola's various loyalty programs around the world.
MAIN CHALLENGES, LEARNING POINTS AND TAKEAWAYS
- If you sell hundreds of different products in multiple countries, perceptions and customer behavior can vary greatly from market to market. If you understand these differences, you can tailor specific messages to different markets, instead of relying on a clear approach
- When dealing with global brands, user data from social media or generated through your own systems (such as vending machines) is huge and messy. AI offers a viable method for structuring this data and collecting insights
- Computer vision technology, such as image recognition tools, can analyze millions of social media images to help a brand understand when, how and by whom its products are enjoyed
- . brands that have fully invested in AI are starting to use it to design new products and services.
This is an edited extract from Artificial Intelligence in practice: how 50 successful companies used AI and machine learning to solve problems , by Bernard Marr, with Matt Ward (published by Wiley, April 2019).
About the authors: Bernard Marr is the founder and CEO of Bernard Marr & Co and an internationally best-selling business author, futurist, keynote speaker and strategic adviser for companies and governments. He is one of the world's most respected voices and a renowned expert when it comes to topics such as artificial intelligence and big data. Marr advises many of the world's best-known organizations on strategy, digital transformation and business performance. He is the author of Big Data in Practice: How 45 Successful Companies Use Big Data Analytics to Deliver Extraordinary Results and Big Data: Use SMART Big Data, Analytics and Metrics to make better decisions and improve performance, both published with Wiley.
Matt Ward is the research leader for Bernard Marr & Co. Matt has a background in investigative journalism and in recent years has worked closely with Bernard Marr on the latest technological topics. Matt is an expert and experienced writer in the field of business technology and artificial intelligence, where he has collaborated with companies such as IBM, Intel, Citibank and NASA.AI, coca cola, Kunstmatige intelligentie, marketing, social