Predicting Accurate CSAT through Call Quality

Case Studies

Here’s how our team helped a major UK based multinational banking & financial services company predict their overall customer experience with accuracy of over 90%.

Client Profile

The client was a major UK based multinational banking & financial services company with revenue over $60 billion. They were a universal bank with operations in retail, wholesale & investment banking, as well as wealth management, mortgage lending and credit cards. It has operations in over 50 countries and territories and has around 48 million customers.

Number of Agents: 4000 | Volume: 2,350,000 calls/month

The Business Challenge

The client was seeking to improve their customer experience for their contact center in the USA. The major challenge was to know if their quality framework was really capable of wisely sensing and predicting customer’s overall satisfaction levels on the brand and quality of service.

For example:

In the above example, the customer rated the agent low on survey questions related to overall satisfaction with quality of service, authority, and agent’s clarity of speech. Whereas the corresponding call quality score does not say so, thereby indicating that the call quality scores are not the true measure of overall customer experience.

Our Approach

The project was guided by the following set of broadly stated research objectives:

  • Identification of the factors that drive the overall customer experience through our home grown analytical technique called Key Driver Analysis (KDA),
  • Assess whether the attributes used to measure call quality are aligned to gauge the customer satisfaction as well,
  • Modification of QA structure to ensure that call quality scores predict the overall customer experience.

Our Solution

Key Driver Analysis:

A CSAT survey form consists of multiple questions which are either elementary (questions to assess agent behavior) or composite (bigger questions like overall satisfaction) in nature. Key Driver analysis was performed to identify the key survey questions which correlate to the overall CSAT. This phase was supported with the following approach:

During the analysis, it was found that the CSAT survey question “Overall Experience with Cards” was significantly driven by the question “Quality of Service on Phone” (considered as Level 1 factor) which was further driven by the another question “Quality of Service by Agent” (considered as Level 2 factor).

In order to control the overall satisfaction, the Quality of Service by Agent was further analyzed. This question was divided into seven sub-questions (e.g. did the agent take ownership, did the agent take answered clearly etc.). On analyzing the inter-relationship of these sub-questions using interdependence modeling we found that 7 sub-questions can be categorized into 3 logical segments as shown below:

Current QA (Quality Assurance) Form Alignment Study

On regressing we found several limitations in their current QA framework i.e. our team analyzed the following variable to conclude that their quality score were not aligned to survey framework:

  • Overall QA Score & Overall CSAT Score,
  • Overall QA Score & Drivers of CSAT (Customer Experience),
  • Section level QA score & Drivers of CSAT (Customer Experience) aligned.

Revision of QA Form

We designed a new QA form by carefully analyzing each of the CSAT drivers. Our analysts tested the set of new QA attributes over a sizable call volume to ensure that QA results generate a model that can predict the overall customer experience with accuracy of over 90%.

Download PDF

Share :

Share :