Improve CSAT Ratings Leveraging Statistical Modeling and Predictive Analytics
The retail giant was scoring very low on the customer satisfaction (CSAT) review for the last 4 years. This was a significant departure from the scenario 10 to 15 years ago when the giant ranked the highest on customer satisfaction.
The retail giant took up the issue seriously and conducted an internal survey with around 8000+ consumers. To its immense surprise, the ratings decreased further and they had in fact reached the lowest customer satisfaction rating.
One of the top retail organization based out of United States with a chain of hypermarkets, discount department and grocery stores across the world.
It has more than 5000 stores in 20 countries and is continuously expanding in new geographies.
A team of Data Scientists, Data Engineers, and Business process consultants was assembled and tasked with decoding the factors driving the low ratings. The team went through gigabytes of customer satisfaction data along with the data on purchase behavior, staff behavior in correlation with the respective store locations for 12 months prior. The key objectives of the analysis were
1. Identify key factors that impact customer satisfaction scores across the stores.
2. Establish a relationship between key influencing factors and customer satisfaction scores.
The Data Scientists on the team used R-Studio tool for Data Analysis and came up with a list of 15 attributes like Duration of Relationship, Product Line, CSAT Score etc., that were found to be the key drivers of customer satisfaction.
The team decided that predictive analytics can help improve customer satisfaction and built a CSAT prediction model using statistical modeling to achieve the same.
The Operations team now have visibility into real-time predictions of customer satisfaction ratings. The system automatically flags cases likely to result in low CSAT scores.
The operations team can then immediately initiate corrective actions by getting in touch with the customers through their preferred channel of communication like Email or Phone Call. Following are some of the impacts generated by analytics implementation:
1. The CSAT Prediction tool predicts low customer satisfaction scores with an accuracy of around 75%.
2. The survey results took a positive turn with a decrease in poor ratings when compared Month-to- Month and quarter-to- quarter pre- and post-implementation.