Imagine your best friend is a manager of a contact center that handles telephone and e-mail response. Her organization's high work-volume season is underway. And, as times are hard, she is dealing with a reduced budget and lost frontline agents whom she cannot replace. Training is postponed, morale is low, and her best agents are leaving the organization. Customers are waiting longer for calls to be answered. What's more, she's struggling to maintain her service level: company-allowed percentages of abandoned calls and calls/e-mail messages answered within a certain timeframe. What would you advise her to do? Here are your options:

A. Maintain the service level at all costs. Schedule overtime and shorten lunch hours. Put the kibosh on all transfers and promotions, and ask supervisors to go back to the phones. Agents who man the phones by day can answer e-mail in the evening. In short, by toughing it out, she can maintain service level.

B. Ease service-level goals and let the chips fall where they may. Customers understand and are patient with companies going through tough times. They won't mind waiting another minute or so.

C. Choose a middle ground. Maintain current goals. Be forgiving of herself and her staff when goals are not met.

D. Tell her nothing - you have no idea what she should do.

The correct answer is "D."

You are missing the information you need to provide wise counsel. Before she takes action, she must know how customers perceive service level and how experiences impact their loyalty. As her budget tightens, she should get smart about service level.

You could ask her how she selected her service-level goals. Maybe she set arbitrary goals based on the fabled "80/20" industry standard (80% of calls answered within 20 seconds). Maybe she "upped the ante" in previous years: making goals tougher to achieve to demonstrate a commitment to continuous improvement. Perhaps she invested in benchmarking. Or maybe the goals were in place when she was hired.

Unless service-level goals are aligned with customer expectations, she has work to do. Possibly, the new degraded service-level performance is still within the customer's tolerance. Advise her to find out. Meeting high service-level goals is costly. And she has to know that adhering to them is money well spent.

SERVICE-LEVEL ELASTICITY STUDIES OF TELEPHONE RESPONSE

You may strongly suggest that she conduct a service-level elasticity study. Through the study, she would learn from customers what they are willing to tolerate. Once she determines how long customers will wait on the telephone without perceiving degradation in the quality of service, she can reset the goals, if necessary. Literature has long established that customer satisfaction is based on perceived (not actual) service delivery.

A service-level elasticity study involves several steps:

The customer survey. Managers design and implement an event survey that captures how the customer perceives wait time, agent and call center performance, and his levels of satisfaction and loyalty.

One question on the survey, measuring perceived wait time, might ask: "How long would you estimate you waited on the telephone before speaking with an agent?" This question would be followed with choices like: (a) 0 to 30 seconds, (b) 31 to 45 seconds, (c) 46 to 60 seconds, (d) 61 to 90 seconds, (e) 91 to 120 seconds, (f) 121 to 240 seconds, and (g) greater than 240 seconds.

The survey could ask the following three questions: (1) Was your wait shorter, longer or as long as you expected? (2) How satisfied are you with the service you received? And (3) how satisfied are you with our company (or organization)? The first question measures expectations. The second and third questions measure the customer's satisfaction with the agent and the organization. The survey can use five-point Likert-type response scales on these questions.

Managers would select the survey sample to reflect calls answered at times when the service level was high and when service level was low to capture a wide range of customer experiences.

The analysis. Analysts use data from the phone system to determine customer wait time. The time is paired one-to-one with the results of the survey; and analysts conduct multivariate statistical analysis. (If actual data is not available, analysts use the average readings for the timeframe in which the customer called. While this is not as accurate, analysts can draw conclusions.) Insights derived from this pairing can answer the following questions:

• What is the perceived versus actual wait time?

• Does wait time affect caller satisfaction?

• Does caller satisfaction affect customer loyalty?

• If wait time impacts satisfaction, is the impact stronger, weaker or equal to the impact of how well the agent handled the call?

• How long will most customers wait for an agent to answer?

• If callers believe they have waited too long, what can the agent do to rescue the call?

The conclusion. Following are examples from companies who have conducted service elasticity studies. The numbers are unique to these companies, but they show what you may discover.

Possible study results:

• Satisfaction depends more on the quality of the experience with the agent than on wait time.

• Wait time may not impact how the customer rates the agent.

• Your customer will tolerate a wait of approximately 120 seconds.

Our results of service elasticity studies conducted for telephone response indicate that if customers estimate a wait time much longer than actual, they have exceeded their tolerance level.

This "exaggeration" is easy to understand. Next time you are in a line, look at your watch to mark the time when you enter a queue. Don't look at your watch as you progress through the queue. When a rep begins to help you, estimate how long you have been waiting. If your wait is "beyond your tolerance level," your estimated time will probably be longer than the actual time and you will feel angry and frustrated. If your wait is "within your tolerance level," your estimate will be closer to (or lower than) the time elapsed. And you will not feel angry and frustrated.

ELASTICITY OF E-MAIL RESPONSE

E-mail response is different from telephone response. Poor service can affect customer loyalty in either case. However, poor e-mail response means more work for the contact center in the form of: (a) customers calling to say, "Didn't you get my e-mail?" and (b) customers sending subsequent "Where are you?" e-mail messages that must be handled. You must know when customers are sufficiently satisfied with the e-mail response and, therefore, are not generating e-mails.

You can study e-mail service elasticity using the same methodology and analysis you would use for a telephone response survey.

Imagine your best friend is a manager of a contact center that handles telephone and e-mail response. Her organization's high work-volume season is underway. And, as times are hard, she is dealing with a reduced budget and lost frontline agents whom she cannot replace. Training is postponed, morale is low, and her best agents are leaving the organization. Customers are waiting longer for calls to be answered. What's more, she's struggling to maintain her service level: company-allowed percentages of abandoned calls and calls/e-mail messages answered within a certain timeframe. What would you advise her to do? Here are your options:

A. Maintain the service level at all costs. Schedule overtime and shorten lunch hours. Put the kibosh on all transfers and promotions, and ask supervisors to go back to the phones. Agents who man the phones by day can answer e-mail in the evening. In short, by toughing it out, she can maintain service level.

B. Ease service-level goals and let the chips fall where they may. Customers understand and are patient with companies going through tough times. They won't mind waiting another minute or so.

C. Choose a middle ground. Maintain current goals. Be forgiving of herself and her staff when goals are not met.

D. Tell her nothing - you have no idea what she should do.

The correct answer is "D."

You are missing the information you need to provide wise counsel. Before she takes action, she must know how customers perceive service level and how experiences impact their loyalty. As her budget tightens, she should get smart about service level.

You could ask her how she selected her service-level goals. Maybe she set arbitrary goals based on the fabled "80/20" industry standard (80% of calls answered within 20 seconds). Maybe she "upped the ante" in previous years: making goals tougher to achieve to demonstrate a commitment to continuous improvement. Perhaps she invested in benchmarking. Or maybe the goals were in place when she was hired.

Unless service-level goals are aligned with customer expectations, she has work to do. Possibly, the new degraded service-level performance is still within the customer's tolerance. Advise her to find out. Meeting high service-level goals is costly. And she has to know that adhering to them is money well spent.

SERVICE-LEVEL ELASTICITY STUDIES OF TELEPHONE RESPONSE

You may strongly suggest that she conduct a service-level elasticity study. Through the study, she would learn from customers what they are willing to tolerate. Once she determines how long customers will wait on the telephone without perceiving degradation in the quality of service, she can reset the goals, if necessary. Literature has long established that customer satisfaction is based on perceived (not actual) service delivery.

A service-level elasticity study involves several steps:

The customer survey. Managers design and implement an event survey that captures how the customer perceives wait time, agent and call center performance, and his levels of satisfaction and loyalty.

One question on the survey, measuring perceived wait time, might ask: "How long would you estimate you waited on the telephone before speaking with an agent?" This question would be followed with choices like: (a) 0 to 30 seconds, (b) 31 to 45 seconds, (c) 46 to 60 seconds, (d) 61 to 90 seconds, (e) 91 to 120 seconds, (f) 121 to 240 seconds, and (g) greater than 240 seconds.

The survey could ask the following three questions: (1) Was your wait shorter, longer or as long as you expected? (2) How satisfied are you with the service you received? And (3) how satisfied are you with our company (or organization)? The first question measures expectations. The second and third questions measure the customer's satisfaction with the agent and the organization. The survey can use five-point Likert-type response scales on these questions.

Managers would select the survey sample to reflect calls answered at times when the service level was high and when service level was low to capture a wide range of customer experiences.

The analysis. Analysts use data from the phone system to determine customer wait time. The time is paired one-to-one with the results of the survey; and analysts conduct multivariate statistical analysis. (If actual data is not available, analysts use the average readings for the timeframe in which the customer called. While this is not as accurate, analysts can draw conclusions.) Insights derived from this pairing can answer the following questions:

• What is the perceived versus actual wait time?

• Does wait time affect caller satisfaction?

• Does caller satisfaction affect customer loyalty?

• If wait time impacts satisfaction, is the impact stronger, weaker or equal to the impact of how well the agent handled the call?

• How long will most customers wait for an agent to answer?

• If callers believe they have waited too long, what can the agent do to rescue the call?

The conclusion. Following are examples from companies who have conducted service elasticity studies. The numbers are unique to these companies, but they show what you may discover.

Possible study results:

• Satisfaction depends more on the quality of the experience with the agent than on wait time.

• Wait time may not impact how the customer rates the agent.

• Your customer will tolerate a wait of approximately 120 seconds.

Our results of service elasticity studies conducted for telephone response indicate that if customers estimate a wait time much longer than actual, they have exceeded their tolerance level.

This "exaggeration" is easy to understand. Next time you are in a line, look at your watch to mark the time when you enter a queue. Don't look at your watch as you progress through the queue. When a rep begins to help you, estimate how long you have been waiting. If your wait is "beyond your tolerance level," your estimated time will probably be longer than the actual time and you will feel angry and frustrated. If your wait is "within your tolerance level," your estimate will be closer to (or lower than) the time elapsed. And you will not feel angry and frustrated.

ELASTICITY OF E-MAIL RESPONSE

E-mail response is different from telephone response. Poor service can affect customer loyalty in either case. However, poor e-mail response means more work for the contact center in the form of: (a) customers calling to say, "Didn't you get my e-mail?" and (b) customers sending subsequent "Where are you?" e-mail messages that must be handled. You must know when customers are sufficiently satisfied with the e-mail response and, therefore, are not generating e-mails.

You can study e-mail service elasticity using the same methodology and analysis you would use for a telephone response survey.