Written on 15 Feb 2022.
Ivan Gordeliy is an Assistant Professor of Marketing at EDHEC Business School. His research focuses on word-of-mouth marketing and user-generated content. He will be sharing his expertise with EDHEC’s Master’s in Marketing Analytics students. Here, he gives us a brief introduction to text analysis.
I will teach Behavioral Insights from Text Analysis. An essential methodological element of contemporary marketing analytics is the quantitative analysis of textual data. Students will be introduced to the concepts, techniques, and uses of text mining, text analysis, and natural language processing in the context of marketing and word-of-mouth marketing. The course will also cover the analysis of user-generated content ‒ such as product reviews and the reactions of consumers or influencers on social networks ‒ as well as how to gather textual customer intelligence. Our Master’s in Marketing Analytics students will learn how to collect, analyse and interpret textual data and extract actionable marketing insights from that analysis.
My research focuses on word-of-mouth marketing and user-generated content. My current interests are in consumer wellbeing, in issues of trust, and the impact of deceptive behaviour. I use machine learning and natural language processing techniques to analyse user-generated content to understand how behaviour affects language and how language impacts consumer behaviour and perception.
Nowadays, consumers are more and more involved in the co-creation process. Interaction between brands and consumers is less and less one-directional. A lot of that interaction is or contains textual data (product reviews, social-media interactions, customer-service interactions), which have not traditionally been analysed quantitatively by marketers. And yet, over the past decade, such methods have been developed and perfected. They are becoming an essential skill in the modern marketing analyst’s toolkit.
Students will learn how to quantitatively analyse textual data and how to draw managerial insights from this analysis. They will learn how to read cutting-edge professional (and scientific) literature on the topic. They will learn ways of quantifying textual data and analysing it. More specifically, we will cover the use of Python libraries for text processing, such as NLTK, spaCY, scikit-learn, and SciPy.
Interpreting textual data, quantifying textual data, text mining on social media, sentiment analysis… we may use the Behaviour Lab. One potential option is for students to participate in collecting data using lab facilities and analyse that data as part of their group project. Students will work in groups throughout the course to complete a project in which they set a question of managerial importance, collect data, analyse it and draw managerial insights from that analysis.
As the field is rather new, I haven't been recommending any particular resources. We will be reading some recent papers on the topic from the top marketing journals (Journal of Marketing, Journal of Consumer Research, Journal of Marketing Research). The course material will be sufficient to master the subject.
I expect course participants to understand the concepts and models of quantitative analysis of textual data. Furthermore, we expect the students to gain experience in working with Python tools to collect, quantify and analyse textual information. Another takeaway from the course should be skills in reading and interpreting cutting-edge literature on the topic.
My estimate, based on job postings on LinkedIn, is that at least 80% of job listings related to marketing analytics require at least one of the skills students will acquire on this course.
As the natural language processing market has been growing exponentially of late, I believe the percentage of jobs requiring the skills acquired on the course will continue to grow.
Furthermore, demand for these skills is higher than supply at the moment, in my opinion. Thus, mastering those skills will give our students an edge in the jobs market.
I expect demand for these skills to continue to grow with the rapid increase in the availability of textual data.