As researchers, we listen to people and spot insights; we create tools and processes that help us gather the best possible information, and truly understand how social groups behave and as well as what they seek. But often we forget to pay attention to the places and platforms where consumers express their thoughts and feelings proactively and willingly; where they go to share their experiences and preferences. For a lot of people, social media has become an open space to connect with others, share their lives, and express their emotions and feelings.
Posts, pictures, videos and everything that people share on Social Media have the potential to be an authentic expression of today’s consumer habits and behaviors, and open roads to future insights. This is why, well-analyzed “social mentions” are an endless source of spontaneous information, and may complement or even replace traditional research methodologies.
Social Mentions as a Research Tool
To understand how ‘social mentions’ work, there are two essential terms to be considered, ‘Thick Data’ and ‘Big Data’. Both share the same procedures:
“Mentions on social media are detected through specific keywords with a common origin, which might be a brand, a category, an ingredient or a topic”
The only difference is their focus is: Thick Data explores and classifies, while Big Data confirms and quantifies.
As researchers and users of social networks, we knew that 75% of people find out what their “people “ and relatives do via social media, and that 1 of every 3 mentions in social networks are related to consumption occasions. By knowing this, we visualized the potential that these tools can bring to the research world; so, we wanted to prove how Social Listening can be useful for market research. We conducted a Thick Data case through segmentation of Beer and Flavored Alcoholic Beverages (FAB´s) consumption occasions in Brazil.
Case study: Beer and FAB´s consumption in Brazil
As we mentioned, we knew that 33% of the spontaneous mentions on social media referred to consumption occasions. These mentions contained spontaneous and unstructured information that could be key for any research, such as:
- What products and brands they are consuming
- When and where they are drinking it
- Who they are sharing with
- Whether they eat something or not
- … and the reasons and emotions behind the experience
Twitter, Facebook and Instagram were chosen to better understand the beer and FAB’s consumption occasions in Brazil ,and it is important to highlight that this country was the perfect match for this challenge, mainly because of its varied regions, cultures, geographic areas, variety of brands, slangs and conversation styles.
The first step in the study was to identify the mentions on these social media channels through more than 22 beer brands and 15 FAB´s brands, with both high and low penetration. More than 600,000 mentions were spotted and, after cleaning and discarding SPAM, ads, etc., 238,000 mentions were left and analyzed. Of all those mentions, a total of 16 consumption occasions were detected as the most important for the country.
In order to find the occasion with more potential for new product development, the project continued with an online quantitative validation of 1,000 millennial cases around Brazil. It was designed as an incomplete block, where each millennial answered about two occasions of the 16 consumption occasions that they actually had experienced with in the last two weeks.
Main Learnings from Social Listening
Firstly, Social listening is a very useful tool for actual research that can even replace typical methods like focus groups or online communities, as long as appropriate techniques are used to capture, clean and classify that data, either by using big data or thick data criteria.
While it is true that you can find spontaneous and unstructured data, you can also capture valuable insights as consumers are giving us their expressions in “Real time”, through arguments and variables that enable us to analyze their consumption occasions such as: time, place, brands, products, companions, etc.
The second learning is the importance of the procedure and techniques that you need to strictly follow, such as the analysis of small samples instead of big samples for data accuracy, the mapping of the facts, data description and emotions to classify what we needed: describe a population, a cluster, a category, a brand and a behavior. In this way it is like bringing the necessary structure to data that was originally un-structured.
Undoubtedly, the third learning is that Social Listening gives a better and fresher understanding of brands and how they interact in the consumer context: valuation, link drivers, barriers, relation with the competitive landscape, functional and emotional benefits, etc… (as long as it is analyzed both qualitative and quantitatively).
The outputs of this project can now be used strategically to bring innovation and disruption into white spaces in the market and/or to define current places were brands were not replaying due to indirect competition. Different customizations in products can be made in order to answer this, like introducing new flavours to your brands, changing your sensorial attributes, or making hybrid products (matching one or two categories at the same time).
