One topic stood out at the recent ESOMAR Fusion conference in Dublin. That was the application of AI & Machine Learning in market research. The usage of these technologies in market research is three-fold:
- Automation, Intelligence and Insight
A large part of any research study is shifting through data to identify insights. Using Machine Learning to recognise entities in images, researchers at REDc were able to identify the underlying themes that were being shared on social media sites around the Irish abortion referendum.
Identifying the emerging themes that motivated the pro and against abortion camps allowed messaging to be tailored to a specific tone and narrative.
- Social Media: Positioning and Targeting
Using social monitoring tools to understand brand sentiment and track brand health are now established approaches.
However, can social listening help brands launch a product?
Suze, an aperitif French based product, were looking to re-launch a forgotten brand. Turning to social media analysis, they gleaned what category consumers were mentioning the brand in and the context to which they mentioned it.
Looking at the related keywords, Suze was able to establish the key competitors to be Jagermeister, Campari, Fireball, Aperol, Monkey Shoulder and Cynar. Through the analysis around the context the brands were mentioned in, Suze learned that the current established brands skewed towards either masculine or feminine connotations, with Suze somewhere in the middle ground.
The connotation around the middle ground identified the consumers of Suze being young/trendy bohemians. A demographic that Suze decided to integrate as part of their targeting and brand strategy.
- Uncovering Driver of Satisfaction
Satisfaction surveys are long, cumbersome and pre-defined with a set of fixed attributes. As a result, consumers are not willing to fill out a survey that is longer than a few minutes. They are easily bored by rating scales, and sometime provide nonsense answers to open-ended questions.
However, social media and chatbots have set a new norm in the way consumers provide feedback.
Using AI & Machine Learning can satisfaction surveys be revamped while yielding meaningful and actionable insights?
Using a mobile satisfaction survey, SEAT asked respondents a total of 5 questions, with 3 being text reply questions and 2 using a likert scale. Applying machine leaning techniques, SEAT was able to create clusters from open-end text as well as understand the drivers of satisfaction.
SEAT was also able to create a pipeline that automates and clusters text information from new customers, allowing them to react in a just-in-time manner with consumers that have a negative level of satisfaction with the brand.
How Does All This Impact Market Research?
AI & Machine Learning has been viewed as something foreign and even threat to market research. However, the rise of technology and the different forms in which data is generated means that AI & Machine Learning is the most effective tool in analysing this data – from the aspect of volume, time and cost.
The cases above illustrates that AI & Machine Learning play a complimentary role to the market research industry and should be viewed as any other analysis methodology tool. As consumers and brands become more technology orientated, generating unstructured data will become more common. Therefore, the need for AI & Machine Learning will likely increase in the future.
This post was originally published on Research World Connect
By Shamvir Singh, Associate Director Innovation & Analytics
If you would like further information, please get in touch via: