Data Analysis – Inferential Statistics

 Scenario:

Your organization is evaluating the quality of its call center operations. One of the most important metrics in a call center is Time in Queue (TiQ), which is the time a customer has to wait before he/she is serviced by a Customer Service Representative (CSR). If a customer has to wait for too long, he/she is more likely to get discouraged and hang up. Furthermore, customers who have to wait too long in the queue typically report a negative overall experience with the call. Youve conducted an exhaustive literature review and found that the average TiQ in your industry is 2.5 minutes (150 seconds). 

Another important metric is Service Time (ST), also known as Handle Time, which is the time a CSR spends servicing the customer. CSRs with more experience and deeper knowledge tend to resolve customer calls faster. Companies can improve average ST by providing more training to their CSRs or even by channeling calls according to area of expertise. Last month your company had an average ST of approximately 3.5 minutes (210 seconds). In an effort to improve this metric, the company has implemented a new protocol that channels calls to CSRs based on area of expertise. The new protocol (PE) is being tested side-by-side with the traditional (PT) protocol.  

Download the database. Each row in the database corresponds to a different call. Column variables are as follows.

  • ProtocolType: indicates protocol type, either PT or PE 
  • QueueTime: Time in Queue, in seconds
  • ServiceTime: Service Time, in seconds

Perform a test of hypothesis to determine whether the average TiQ is lower than the industry standard of 2.5 minutes (150 seconds). Use a significance level =0.05. 

Evaluate if the company should allocate more resources to improve its average TiQ.

Perform a test of hypothesis to determine whether the average ST with service protocol PE is lower than with the PT protocol. Use a significance level =0.05. 

Assess if the new protocol served its purpose. (Hint: This should be a test of means for 2 independent groups).

Write a 250-word summary of your conclusions. Include your calculations in the summary.