Impact of e-WOM on consumer purchase behaviour through Twitter sentiment analysis using vader and machine learning

With Internet penetration expected to reach 639 million in India by the end of 2020 and social media usage booming to share opinions, firms are looking for social media monitoring to get the customers’ pulse. This research paper investigates the impact of e-WOM on purchase behaviour and image by performing a sentiment analysis on 30,000 tweets using and Machine learning. It extracts real-time tweets of customers before the product launch of two brands, One Plus and OPPO, in the PAN-India market. It performs sentiment analysis to understand customer behaviour. This paper examined that tweets containing geopolitical situations like #Boycott China negatively influence purchase behaviour, addressed by sentiment analysis of tweets during OPPO’s product launch during an India-China border spat. This paper has also found out that tweets from Influencers and Peer groups impact consumers’ purchase behaviour. This paper also bridges the vital research gaps left in this domain, such as sentiment analysis on tweets containing emoticons, emojis, slang, local, regional languages, etc. Since social media conversations nowadays are mostly filled with emojis, emoticons slang, it is extremely important to consider it during sentiment analysis. This paper has also studied the significance of e-WOM factors on purchase behaviour and has provided a managerial perspective with a real-time example of decision making with sentiment analysis on product launch of two Chinese firms in India during a trending geopolitical situation.