Using NLP to analyze social media comments and reduce the time to market by 2X
The customer is a leading multi-national CPG company that wanted to listen and monitor the social media buzz around its products and generate insights about buyers’ sentiments by analyzng the reviews and comments. The customer was also looking to uncover pain-passion points for own and competitor brands, and identify specific aspects of products that need to be improved. The marketing team wanted to leverage these insights to optimize marketing efforts, whereas the R&D for product innovation.
We gathered data from different social media channels and internet platforms with brand mentions using scraping tools and APIs. NLP-based sophisticated architectures and best-in-class sentence embedding techniques were used to create a powerful filtering mechanism to focus on meaningful and relevant comments. Brand extraction and disambiguation models were created which removed the ambiguity between words similar to the brand name. These state-of-the-art deep learning models were trained for each brand using the labeled brand data and brand dictionary. We worked on refining the keywords through an in-built aspect recognition system and used techniques such as LDA, Double Propagation, Attention, and Latent Semantic Analysis for keyword extraction.
The solution enabled the product R&D team to focus on product innovation. It also resulted in an increased response time to customers as it allowed brand teams to address customer pain points and dissatisfaction while targeting campaigns to relevant audiences.
Lesser time to market
Reduction in time to launch a new brand
Increase in prediction accuracy