PDF File: 

crm7_4_167_wp_purdue_csula.pdf

Purdue University, Center for Customer-Driven Insight

Untested and unquestioned as the sole determinant measure for years, service level today is only oneamong several metrics directing strategic decisions, affecting design, implementation and the success ofoptimized contact centers.

How do you run a call center efficiently and effectively to maximize caller satisfaction given the resource and budget constraints that most businesses face? What are the critical variables in providing efficient and effective call center management? Service level is used in virtually every call center (for a solid discussion of what a service level is, see [1]).

Service level is defined as the percentage of calls answered within a specific time, and appears also to be the focus of how industry experts believe call centers can be run well.[2] Eighty percent of calls answered within 20 seconds is probably the most common form of service-level goal.[3,4]

There is no empirical work on how important service level really is for call center efficiency and effectiveness. It is used as a common metric because:

1. It is very easy to measure since it is automatically reported in every major technology.
2. It has been around a very long time. It was built into the first automatic call center switches and has
3. There is a certain face validity to thinking that if you answer a greater percentage of calls within a certain period of time, this indicates good service.

If service level is important, there should be a relationship between service level and caller satisfaction and other variables of call center efficiency and effectiveness. If service level is virtually a meaningless measure, then the industry has generally been wasting time, effort, and – more importantly – money that could have been used to build up call center operations that have an impact on important outcomes like caller satisfaction.

Research has been done to understand the relationship between call center service level and caller satisfaction.

While there is no real agreement about what metric(s) are important in running a call center (aside from service level), the following metrics appear to be the most widely cited (e.g., [5]): abandonment rate, adherence to schedule, average speed of answer, talk time, work time after call, inbound calls per eight-hour shift per customer service representative (CSR) – the average number of calls per CSR handled per shift – calls closed on the first contact, calls blocked, queue time, percentage of calls closed on the first contact, percentage of calls blocked, abandonment time, total calls offered, CSR turnover and caller satisfaction.

Hypotheses

The focus of customer contact today is on the relationship of service level to a call center’s efficiency and effectiveness. If it is, indeed, an important measure, service level should closely relate to call center efficiency and effectiveness:

H1 – There should be a positive relationship between service level and caller satisfaction.

H2 – There should be a negative relationship (opposite to predictions) between service level and abandonment rate. If service level increases, we would expect that consumers would have no reason to abandon a call.

H3 – There should be a positive relationship between service level and adherence. Service level means that more agents are in seats ready to answer calls.

H4 – There should be a negative relationship between service level and average speed of answer (ASA). As service level improves, the total number of calls answered can be increased and customers will not wait for as long a time for a customer service representative.

H5 – There should be a negative relationship between service level and average talk time. Better service level represents a call center’s ability to handle a greater number of calls and reach a higher degree of productivity. The average total time for one customer connected with the CSR should decrease.

H6 – There should be a positive relationship between service level and average work time after call. The more calls answered, the more likely more will need after-call work.

H7 – There should be a positive relationship between service level and inbound calls per eight-hour shift per CSR. Higher service level means more calls answered in a specific time frame, and therefore more calls answered during shifts.

H8 – There should be a positive relationship between service level and percentage of calls closed on the first contact. As service level improves, the call center can handle more calls and if the call center regularly closes calls on first contact, the number closed on first contact should increase.

H9 – There should be a negative relationship between service level and percentage of calls blocked. If service level increases, customers can get through to CSRs more easily and more calls can get connected to the call center without being blocked.

H10 – There should be a negative relationship between service level and average queue time. Better service level means more calls can be answered quickly so that the amount of time customers wait in a queue to be connected to a CSR should decrease.

H11 – There should be a positive relationship between service level and average abandonment time. In this way, better service level equates to more calls answered and fewer consumers hanging up.

H12 – There should be a negative relationship between service level and total calls offered. As service level improves, the number of total calls offered will go down.

H13 – There should be a negative relationship between service level and CSR turnover. Poor service level leads to more complaints since callers cannot get through, increasing stress which leads to more turnover.

Methodology

Data in this study is taken from the Purdue University Benchmark database of call centers (www.benchmarkportal.com) for the year 2004. This database contains 2,974 call centers related to 30 industries. Variables are measured according to the operational definitions provided by Anton.

Results

There are wide differences in service-level performance across industry groups, as shown in Figure 1:

The correlation between two variables (reflecting the degree to which the variables are related) (see Figure 2), were completed between service level and the variables of interest.

The results show that service level was significantly related to six variables;

1. As service-level performance decreased (for the purposes of this study, higher service-level performance equals poorer performance – i.e., it took longer for centers to answer 80 percent of calls), caller satisfaction increased.
2. As service-level performance decreased, abandonment rate increased.
3. As service-level performance decreased, average speed of answer increased.
4. As service-level performance decreased, percentage of calls blocked increased.
5. As service-level performance decreased, average queue time increased.
6. As service-level performance decreased, average abandonment time increased.

Among these six variables, only four related to service-level performance in the manner predicted: abandonment rate, ASA, percentage of calls blocked and average queue time. It is logical that since fewer calls can be answered quickly, more callers would abandon the call (abandonment rate). As wait time increased, more calls are blocked by the call center because the queues become filled, causing callers to wait longer.

If any or all of these variables were related to customer satisfaction, then service level would be an appropriate proxy and would be important to measure. However, only one of these (ASA) is related to caller/customer satisfaction, indicating that service level is not as good as the single variable alone in predicting/causing caller/customer satisfaction.

Two variables had negative relationships (opposite to predictions) with service level: caller satisfaction and average abandonment time. No correlation was found between service level and adherence to schedule, average talk time, after-call work time, inbound calls per shift, calls closed on first contact, total inbound calls annually or turnover, although it was predicted that there would be significant and clear relationships. A follow-up regression analysis showed that only two of the variables – ASA and first-call resolution – were determinant (causal) of caller satisfaction.

The determined issue is the presumed centrality of service level in the call center professional literature. Not only were many of the presumed relationships virtually nonexistent, but when they were found, they were small and some even opposite of what would be expected. This disparity is particularly important in terms of caller satisfaction.

If we assume there really is nothing more important to the effectiveness of a call center than caller satisfaction, we would assume that if no other relationship were found; yet, in fact, there should be one between service-level performance and caller satisfaction. Interestingly, a relationship was found, but in a direction opposite of that predicted. As service-level performance decreased (it took longer to answer 80 percent of calls), caller satisfaction increased.

Discussion

In understanding the relationship between call center service-level performance, caller satisfaction and measures of call center efficiency and effectiveness, the findings of this research show that service level is actually not that important in the operation of a call center. It is not a critical variable for caller satisfaction and is not important for call center efficiency and effectiveness. Indeed, the negative relationship between service level and caller satisfaction found (as service-level performance decreased, caller satisfaction increased) suggests that call centers that focus on reaching service-level goals may in fact make it harder to achieve any customer satisfaction goals.

The percentage of calls closed on the first contact and the average speed of answer are found to be the two significant predictors of caller satisfaction. This suggests that if call centers want to measure and manage something that really makes a difference, they would measure only average speed of answer and those closed on first contact.[6]

Managerial and Practical Implications

This study indicates service level is not a key metric for call centers; certainly not as important as the consultants and professional magazines suggest. Firms using service level as a standard for call center efficiency and effectiveness waste resources on a comparatively unimportant factor. The new rulers of call center management are speed of answer and first-call resolution.

Endnotes

1. Slifer, D. (2004), “Service-level management: Guaranteeing customer satisfaction,” Available at www.seneca.com/news-whitepapers.html
2. Cleveland, B. (2002b), “Easy access: Ten guidelines to speed-dial your way through the customer service gridlock,” Operations and Fulfillment, August, 26-29.
3. Anton, J. (1997), Call Center Management by the Numbers, West Lafayette: Purdue University.
4. Cleveland, B. (2002a), “Twelve Traits of the Best- Managed Call Centers,” retrieved Dec. 5, from http://www.incoming.com.
5. Klenke, M., and Trickey, P. (2002), “Super balance,” Operations and Fulfillment, November, 31-36.
6. Feinberg, R., Kim, I. S., Hokama, L., De Ruyter, K. D., and Keen, C. (2000), Operational Determinants of Caller Satisfaction in the Call Center,” International Journal of Service Industry Management, 11(2), 131-141.