Creating a "Customer Centricity Graph" from unstructured customer feedback
This whitepaper describes how Insaas and LMU Munich have used publicly available feedback on car insurance in Germany to develop a proprietary pipeline for calculating and visualizing customer opinions.
Summary
Business-to-consumer (B2C) industries, such as car insurance, need to focus on their customers’ needs in order to offer them the product they want. Since there are rarely touch points between companies and customers, companies need to get as much out of customer feedback as possible.
Currently, such feedback is mainly found in unstructured texts that are publicly available on the Internet, for example on comparison portals. Star ratings in particular are a popular way for consumers to comment on the overall quality of an insurance company.
However, this is only a very general approach to a complex topic. More differentiated information can be found in the rating texts themselves. To avoid manual analysis of this huge amount of data, an automated approach to information extraction and visualization is needed.
Together, Insaas and LMU Munich have developed a multi-stage procedure to solve this problem. The solution is able to identify topics and their polarity and to group this information in such a way that customer opinions can be displayed in the form of a graph, the so-called Customer Centricity Graph.
Using this graph, companies can identify the areas in which they perform better than their competitors and where there is room for improvement.
