High Variation in CSat Data Covers Up
Possible Insights to Improve Customer Satisfaction
Context
A chain store, with nearly 300 stores in 5 states, has an active customer satisfaction improvement program; however, in recent quarters, efforts to improve satisfaction have stalled.
Previously, the chain’s managers have used dashboard-type reports that present average customer satisfaction scores and other summary results. These dashboards are not providing actionable insights to generate improved customer satisfaction.
The chain’s managers wonder if alternative analyses of existing data might provide insights to drive additional improvement.
Question
Do new analyses of existing data reveal possibilities for enhanced improvement of customer satisfaction?
Some stores have customer satisfaction scores that are highly variable, ranging from a low of 10 to a high of 100
Results
A statistical analysis called “Analysis of Variance” indicates that excessive variation occurs in individual stores’ scores.
A graph of customer service scores for individual stores demonstrates the excessive variation in customer scores within stores: Some stores’ scores vary by up to 90 out of 100 possible points.
For example, one store has 10 separate customer service reviews and an average score of 75.2, which the chain classifies as a low score. The store’s satisfaction scores vary from a low of 10 to a high of 100, which is 90% of the possible range of scores.
Insights
High variation in customer satisfaction scores is bad. The extremely low scores reduce the average customer satisfaction scores. But more importantly, excessive variation means that stores can’t predict how each new customer is going to react to the store’s goods and services.
Recommendations
Research indicates that, for every one customer that complains, 26 other customers have the same complaint but don’t voice their dissatisfaction. Identifying why one customer is dissatisfied will provide insight into why other customers also may be dissatisfied. However, the current customer satisfaction survey doesn’t generate the correct data to define why customers are dissatisfied.
The best approach for the chain store to understand the low scores is to go straight to the customer to find out directly why customers who gave low scores are dissatisfied with their experiences.
The chain managers might consider contacting a random sample of customers who submitted low scores to explain that understanding why the customer was dissatisfied is important to the chain’s managers and to ask for the customer’s input to improve customer satisfaction.
Being contacted by the store will show the customers that their opinions are important to the chain and asking for their input to solve a problem will make the customer feel like the chain’s partner in solving a problem.
The information from these interviews will help the chain’s managers to more clearly understand what really is important to customers rather than what the chain thinks might be important to customers. These insights can be used to change products and processes in the stores as well as to update the customer satisfaction survey to be more responsive to customers’ needs and desires.