Understanding the “Voice of the Customer” is necessary for responding to customer needs and improving service, but it can be difficult and time consuming to identify the most actionable feedback. A major insurance company utilized a customer survey that covered 11 different topics. The goal was to gain actionable business insight that could be used to improve customer loyalty.
The text data in every survey was different so training data could not be extracted from the survey results. Elder Research used a text analytics technique that required only a few labeled cases to evaluate the survey data. Best-practices were used for creating features from the text and then combined with structured data to create a loyalty score based on the survey text comments.
The semi-supervised approach provided more focused analysis of the survey comments that mattered for understanding customer sentiment and resulted in more focused action and improved customer loyalty.