Measuring Brand Health: From Survey Data to Social Media Data

Brand health is one of the key concepts in any brand’s performance, since regular brand health tracking helps to understand how market users perceive the brand, what the weaknesses and strengths of the brand are, and what steps can be taken to strengthen its position in the market. Today there are a wide range of methods to measure the brand health: for example, consumer surveys, measuring brand health through market indicators of sales volume, or recently appeared methods for assessing brand health through online indicators (web traffic, the number of visits to brand pages, etc.). The focus of this study is on the methods of measuring the brand health through online indicators; in the situation of the current mass popularity of social networks, social media is a huge platform with consumers of various brands, which cannot be ignored when measuring the effectiveness of the brand. The goal of this study was to develop a methodology for measuring brand health based on data from social media. The empirical object of the study was the brand "Tinkoff Bank", which is fully focused on remote bank service. To do this, we used one of the methods of data mining - web scraping and collected reviews about the brand "Tinkoff Bank" from three social media: the portal Banks.ru, the Google Play mobile app store and its equivalent App Store. Then using the topic modeling algorithm LDA (Latent Dirichlet Allocation), we analyzed the received text reviews for a finite number of interpreted topics and estimated the prevalence of each topic in the total number of comments in the selected month.