Approach
Before a call is scored it needs meet the following criteria:
Duration of at least 90 seconds
Agent clarity of at least 70%
Client clarity of at least 70%
The clarity criteria removes calls where the audio quality is poor or where there is no agent or client on the call.
The above criteria can be customized.
For each call, agents are scored in four major categories.
Agent Conduct
Communication
ComplianceCustomer Sentiment
Empathy Response (Previously known as ‘Customer Sentiment’)
Each category is broken down into a series of questions, with each question supported by a comprehensive set of rules and filters which can be customized for client specific requirements.
Each question will test for either positive or negative behavior. Agents gain points for positive behavior and lose points for negative behavior.
Overall, there are 33 questions categorized as follows:
Agent Conduct
2 questions testing for positive behavior
4 questions testing for negative behavior
Communication
13 questions testing for positive behavior
2 questions testing for negative behavior
Compliance
6 questions testing for positive behavior
0 questions testing for negative behavior
Customer SentimentEmpathy Response (Previously known as ‘Customer Sentiment’)
0 questions testing for positive behavior
6 questions testing for negative behavior
Each question is assigned a threshold and weight:
Threshold:
This is a hurdle (expressed as an integer) that must be met in arriving at a score.
For example:
A threshold of 2 may be assigned to the question, “Did the agent display empathy?”
The agent must show two signs of empathy in the call before gaining points for this question.
Thresholds can be customized to individual client requirements.
Weight:
Weights are used to provide weighted scores at the category level and total call level.
Not all questions are of equal importance, so each question is assigned a weight to indicate its relative importance.
For example, compliance related questions may be given a greater weight than questions relating to actions which may be not occur on every call (such as transferring a call).
Weights can be customized to individual client requirements but must sum to 1.
Scores and Scoring Levels
Scores are derived at three levels:
Question Level Scores
Category Level Scores
Total Call Level Scores
Question Level scores are “rolled-up” into:
Category Level scores and
Total Call Level scores
based on the weights assigned to each question.
Question Level Scores
Each of the 33 questions is given a score of 100 or 0
100 | Positive behavior is identified |
0 | Positive behavior is NOT identified |
The above approach permits both Positive and Negative scores to be summed to arrive at a score at the Category and Total Call level.
Category Level Scores
Each of the four categories is given a score
The category score is the weighted sum of the scores for the questions within that category
Category scores are indexed to 100 for ease of interpretation
Category scores will range between 0 and 100
Total Call Level Scores
Each call is given an overall score, being the weighed sum of the scores for the 33 questions.
The total score per call will range between 0 and 100
The structure and roll-up of scores into higher level scores is depicted in the diagram below:
