A utility client wanted to improve the performance of their call center operations. Traditionally the call centers had been run from an inside-out perspective. Hence, performance was measured only in terms of metrics that managers felt were important. The customer perspective was not considered. With the involvement of the market research department, a decision was made to understand the customer perspective and benchmark performance as seen by customers.

TRC (then The Response Center) was selected in 1999 to guide this process through proper design, data collection and analysis. Working with the client we identified customer satisfaction with the calling experience as the single most important variable to focus upon. This variable most effectively captures the customer perspective, and its variation over time would allow the client to understand call center performance as perceived by customers. However, tracking this variable over time alone would not be sufficient. It needs to be positively influenced. For this, a comprehensive customer satisfaction program needed to be created.

In consultation with the client we designed a customer satisfaction tracking program that used baseline and annual follow-up data collection efforts to monitor performance. The baseline wave of data collection was conducted in early 1999 followed by advanced statistical modeling (explained later) to understand the drivers of satisfaction with call experience. Suggested improvements were implemented in the following months and new measurements were made through the following year. Every year thereafter, statistical modeling has been conducted at the end of the year to identify key drivers. Early in the following year the client has implemented the suggestions and performance improvement has been continuously monitored.

Due to the success of the program, the scope of the study has increased over time to include multiple call centers with varying call types and different types of customers (business and residential). The scope of the questionnaire has also enlarged to include issue such as the Interactive Voice Response (IVR) system. This case study details how quantifiable improvement in the primary program was achieved through the use of advanced statistical modeling.

MODELING

Key driver analysis aims to understand the important influencers of a specified variable. For example, it is difficult to directly improve customers’ satisfaction with call experience. A key driver analysis would be required to identify the variables that influence satisfaction with call experience. In order to do this, the questionnaire needs to include adequate and proper questions. Working with the client we were able to design such an instrument and collect appropriate data through telephone surveys.

The specific statistical technique used in this case was Satiscan^TM a proprietary product that belongs to TRC. Traditionally key driver analysis uses regression analysis to identify the influencers of the variable of interest. While regression analysis is a pretty robust method, it does have some shortcomings. Primarily, it does not model the inter-relationships between the independent variables. Satiscan^TM does this and is hence able to provide a more detailed and accurate picture of the key drivers. For this program, the questionnaire was designed for the use of Satiscan^TM to model the data.

RESULTS

Based on the results from the Satiscan^TM modeling, several recommendations were made to the client. The most important of these were improvements in:

CSR taking responsibility for solving the problem
Customers having to call only once to resolve their problems

As it happens both of these issues are related. Improving performance on the former is one way of improving performance on the latter without necessitating a substantial increase in staffing resources. Further, the model also predicts that these improvements would lead to an improvement in the satisfaction with the CSR who answers the call. Finally, a clear improvement in satisfaction with the call experience (the variable of interest) should be seen.

After identifying the key drivers we engaged in a series of discussions with the client to discuss operational changes in their call centers and changes in training methods. These ultimately resulted in the implementation of these (and other) recommendations in an effort to improve performance.

Performance has been tracked steadily over the last five years and is currently ongoing. The improvement in performance in the variables of interest is as follows:

Overall call satisfaction has improved by 15% points
CSR satisfaction has improved by 14% points
First call resolution has improved by 18% points
CSR took responsibility has improved by 13% points

Data collection and modeling are continuing on different customer groups and call types. Recommendations continue to be made on a yearly basis and the client is very satisfied with the program.