Service Level

[wannabe21]

I was curious what the difference would be between a 90% service level answered in 20 seconds rather than a 90% service level answered in 40 seconds. Obviously the hold time, but would it be more beneficial to decrease Service Level or increase ASA?

[JoeB]

Excellent question that demonstrates the need to fully understand the meaning and significance of call center metrics and not rely on arbitrary rules of thumb.

The two figures in service level are not directly correlated, i.e. increasing one and decreasing the other by the same amount does not yield identical service level.

The time threshold in service level should be set to below the abandonment sensitivity: the time that most callers are wiling to wait for service. Each audience has different needs and expectations that influence their abandonment patterns, and there is no “standard” or “best practice” that you can use. Rather, you need to determine the behavior of your audience by analyzing wait times and abandonment rates using ACD data.

The percentage determines the minimum number of callers that will experience wait times that are below that time threshold, assuming that the rest will abandon. Therefore, if the wait time sensitivity is 30 seconds and you target service level is 80/30, you assume that 20% will abandon. Please note that this is an example of a poorly selected service level: the time threshold should be shorter than the abandonment threshold, e.g. 80/20 or even 70/20.

Another consideration in setting service level is that it determines how many callers will be placed in a queue. In certain environments it is advantageous to minimize the number of such incidents, and the service level should focus on answering most of the calls within the target service level even is the queue wait time is below the abandonment sensitivity. Also, remember that ASA represents wait time for all calls, whereas service level defines the wait time for a predefined percentage of that group.

When I look for opportunities to optimize a call center operation I always emphasize that there is no significant benefit in answering most of the calls faster than callers expect. Therefore, if your callers tolerate 40 seconds wait time, there is no real advantage in answering the calls in 20 seconds, and you should increase the percentage of callers that receive service within their acceptable window.

The staffing modeling of call centers and the non-linearity of many of the processes involve numerous factors and cannot be expressed in a set of rules of thumb. To fully understand the impact of changing service levels, including changes in abandonment rates, how many callers will be placed in a queue and for how long, what are the maximum anticipated wait time and so forth, you need a call center modeling software, which will also determine the staffing requirement to meet those goals.

I recommend that you download a call center modeling software (e.g. http://www.diagnosticstrategies.com/easyerlang.htm) and use it to try different service level combinations to see their impact and resource requirements in your call center. Additionally, you may want to read these articles on call center planning, service levels, and call abandonment.

http://www.diagnosticstrategies.com/SL_elasticity.htm
http://www.diagnosticstrategies.com/...c_modeling.htm
http://www.diagnosticstrategies.com/abandonment.htm

I hope this helps.
Joe