Social Media Analytics At Coca-Cola: Learning From The Best

Social media has become a powerful instrument and today it's inseparable from marketing. Aside from allowing for more brand awareness, social content creates many useful data points around them, and if collected and analyzed in the right way, it may reveal some interesting and useful customer insights. Wes Finley, former Global Digital Marketing Operations Lead at Coca-Cola, who was among the keynote speakers at our Big Data & Marketing Innovation Summit in Miami, shared with us how to extract value from social media data.

Coca Cola has over 700,000 associates and 23 million retail outlets across 200 countries, as well offering over 3,500 products. But with this kind of organizational structure, how can you ensure your marketing strategies are evenly executed across both local and global markets? Wes says that Coca-Cola has managed to appear as one brand voice across all their markets by acknowledging these their unique features and demographics. Essentially, their local markets work like franchises, where marketers have their own budget and manage their own campaigns. At the same time, there is also a global brand focus, which is reflected in events like the Olympics and various World Cups.

One of the difficulties of his job, Wes admitted, 'was getting all that data from all those local marketers and bring it together in a cohesive way, to tell the story and to really drive every one of the regions forward in their marketing.' Regarding social media strategies at Coca-Cola, there was a lot of data to work with. In 2015 alone, Coca-Cola produced over 120,000 pieces of content, where each had unique data points in the form of impressions and engagements, and these helped the company to learn more about different types of consumers, their demographics, and behavior.

From 120,000 pieces of content, Coca-Cola needed to analyze over 20 billion impressions, as well as all the data that was generated from daily mentions of the brand outside marketing campaigns, which came to around 35 times a minute globally.

Why do you need to analyze all of this data?

Firstly, because it helps to create one powerful brand narrative across all segments, and, secondly, because 90% of consumers, today, base their purchasing decisions on social media content. With such large amounts of data available, Wes says, there must be machine-learning solutions, which allow assumptions to be made and help to differentiate data.

Using learning and recognizing algorithms, it is easier to find out whether Coca-Cola is mentioned 35 times a minute because someone tweeted 'the drink is horrible' or these mentions are positive. Wes says it's critical to identify these risks and opportunities as they can then be pushed to the marketers, who can then decide on whether they need to resolve a PR problem or further promote the content.

To ensure marketing campaigns reach targets and create value across all segments evenly, Coca-Cola has 37 Social Centres which represent connected and networked resources. Using Salesforce, these teams collaborate and extract social data related to risks and opportunities. Social Centres have three main purposes: to listen, to engage, and to create content.

As for the actual data analysis, social centre teams use a variety of tools. Among them is Gaze Metrix which allows them to identify company's brands in visual identity systems, for example, photos. If people tag a product in a photo, the system can differentiate what the product is. Marketers can then apply images to Gaze Metrix and from there, they can see whether people have been taking pictures with Coca-Cola brands on Instagram. This data can be then added to the company's listening software to measure if conversations are going up or down, even if people were not mentioning the brand. Wes pointed out that it's similar to having a facial recognition feature, but for brands.

Additionally, listening and analyzing content can tell marketers when they need to take action. For instance, Coca-Cola had a movement known as 'Bring back Surge' (a discontinued 90's brand) which had over 1 million followers. Wes said it all started when a group of people showed up in front of the office with a billboard saying 'Dear Coke, we couldn't buy Surge so we bought this billboard instead.' He added that it was important to listen to this kind of conversations as these help the company to understand more about changing preferences, ways customers buy products, and they also indicate what they want. Surge was brought back firstly to Amazon, and once it proved to be successful online, Coca-Cola decided to relaunch it for mass distribution.

The value that social media analytics can create should never be underestimated. With the right set of tools and customer first approaches it can allow companies to learn more about hidden risks and opportunities, as well as allowing brands to appear more trustworthy in the eyes of a customer - because they are heard, engaged, and informed, at all times.