In the new global economy, reputation is an asset for a corporation and plays a key role for customer retention and acquisition. Online reviews often prove to be a reliable, complete and objective collection of opinions regarding a corporation. Socially engaged companies benefit from review mechanisms by having higher sales. Although extensive research has been carried out in this area, no single study exists which attempts to predict corporate reputation performance based on online collected data from media sources.

Trend

This genuine trend underlines the need to process huge amounts of clients’ opinions and reviews to extract valuable experiences from it that yield meaningful insights which can be used to boost corporate reputation.

In our current activities, large datasets are automatically scraped from well-known data platforms in order to derive reputation measures on four aspects with automated models for topic identification as well as sentiment prediction using machine learning and natural language processing.

Customer behavior

A rich survey dataset allows to additionally include information related to customer behavior in our models, besides the information that is extracted from ratings and reviews. Our complete data and model infrastructure allows for benchmarking and inclusion of new, original solutions that are developed and engineered in our state-of-the-art cmihva toolset that generates insights for evaluation of the reputation construct based on collected reviews for any retail organization that is willing to analyze or improve its performance.

There are at least four primary goals in our current activities:

1. To investigate which components of Corporate Reputation are crucial for its estimation

2. To develop methods to extract relevant information, for Corporate Reputation, from publicly available datasources, such as online review websites

3. To create models to predict relevant aspects of Corporate Reputation, such as delivery for instance, within a given review

4. To create models to predict sentiment polarity on predefined aspects of Corporate Reputation, within a given review