Operational Indicators – Sevice Levels & ASA

By Turaj Seyrafiaan
In the last Indicators post we discussed one of the most important operational indicators: Average Handle Time or AHT. In this issue we continue with other operational indicators dealing with the efficiency of the center, i.e. “Service Level”, “Average Speed of Answer (ASA)” and “Occupancy Rate”. Although each indicator provides different information, they are inter-related and must be treated as a single group.
Service Level
Service Level – the most commonly used center metric – is defined as the percentage of contacts that are answered within a specified target time threshold. It is presented by two numbers such as “70 – 30”. The first number indicates the percentage of the calls, while the second number indicates the target time threshold in Seconds. The threshold is measured from the time that the calls arrive in the queue and does not include any time prior to that (e.g. time spent by the customers navigating the IVR menu). A 70 – 30 service level, means 70% of calls were answered within 30 Seconds.
In practice call centers set their overall target (both percentage of calls and the threshold) in conjunction with their Work Force Management (WFM) process in order to calculate their staff requirements and scheduling (more on that in later sections). A higher Service Level means faster service (answering the call) for customers. The actual Service Level – achieved percentages within that set threshold – is then reported on a daily / weekly / monthly report.
Service Level can be manipulated by an unscrupulous manager through reducing or ‘choking’ the incoming calls and by reducing lines etc. Remember service level is a measure of the percentage of calls answered within a defined timeframe measured over a period of time, so you must ensure that the period of time over which service level is measured is reasonable. When working with averages and service level is an average, the larger period of time over which it is measure the more periods that can fall below the target. For example service level measured over a 24 hour period can be missed for periods during the day and made up for during the evening or overnight period. Measured over a week could mean missing the service level Monday through Wednesday and making up the service level over the balance of the week.
A third factor or metric is often added to Service Level: that is Abandoned calls. Abandoned Calls can be defined as the percentage of callers who elect to hang up, or abandon, waiting in the queue and before an agent answers the call. A service level of 80/20/<3 would be answering 80% of the incoming calls answered within 20 seconds, with less than 3% of calls being abandoned. Average Speed of Answer (ASA)
While Service Level indicates the percentage of calls that were answered within the set threshold, it does not provide any information regarding the remaining calls! In practice, even when reaching the target Service Level, it is possible for number of calls to spend significant (and unacceptable) amount of time in the queue without impacting the Service Level. For this reason, it is important to look at this indicator that represents 100% of the callers. Average Speed of Answer (ASA) is the average wait time (in the queue) for all the calls answered, in Seconds.Obviously a lower Service Level (lower percentage of calls or longer threshold) produces a longer ASA. Combined with the Service Level, ASA provides a more complete picture of the flow of the incoming calls. For example an ASA of 18 Seconds over 100% of calls received along with the Service Level attained of 80-85% (at 20 Seconds threshold) indicates a significant delay in answering calls beyond the first 20 Seconds and therefore very poor service for the remaining 15-20% of customers). This is the effect of a long tail of outliers in any set of observations. This outlier effect on the customers’ view of the service is not intuitively obvious to many, both inside or outside the call center industry.
Occupancy Rate
Although not related to the customer wait time, Occupancy Rate is very much part of the WFM process and related to the Service Level. Occupancy Rate indicates the percentage of time that agents are occupied, performing call center activities (talking to customers and/or performing after call tasks). The inverse (100% minus Occupancy Rate) is the amount of time that agents are waiting for calls to arrive, also called Availability. It is easy to see that a higher Occupancy Rate indicates a more efficient call center (in terms of Work Force Management and labor costs), however, Occupancy Rates in the high 90% range also indicate extreme work load on the agents (leading to fatigue, poor performance and eventually high turnover) as well as lower Service Level (longer wait time for customers). Low Occupancy Rate on the other hand could indicate poor planning and/or scheduling (too many agents waiting for calls to arrive). Low occupancy can also lead to poor moral and agent dissatisfaction. The art of the Work Force Management process is to create a balance between the Service Level and the Occupancy Rate. Practice shows that for most (though not all) centers an Occupancy Rate of between 75 to 85 percent is optimal.
Target Service Level
Many contact center managers simply assume that a target Service Level of 80 -20 is the industry standard and therefore use that as their own target. While this may be the most common service level for customer service call centers, the fact is that there is no industry standard for the Service Level. Each center must set its’ own target based on its’ own customers’ expectations (mainly) and budgetary / staffing limitations. While there are centers that feel a 90 -10 results (90 percent of callers are answered within 10 Seconds) is not good enough for their particular sector. There are other centers that can reach an excellent customer satisfaction with 80 – 30 or even 70 – 30! (Setting the threshold target beyond 30 Seconds for customer service/support applications is not Recommended) Keep in mind that this is the time in the queue only and a customer may have already spent additional time in the IVR.
Other centers such as 911 or emergency set their standards as 100 – 3. If you are calling with an emergency then that level of service is appropriate. Similarly technical support centers often have target service level wait times of 3 to 5 minutes for free support. Once again, this is appropriate. Each center must define what is appropriate for their center and their customers. Your service level targets also have to be reasonable. There is no sense setting a target that is clearly not achievable. Targets must be reasonable and when they are not met the center manager must report on why the targets were not met!
Work Force Management (WFM)
We need to understand how these indicators are used in conjunction with the WFM process. In brief WFM is a series of activities related to forecasting call (and/or contact) volumes, and scheduling required and appropriate staff. Part of this process involves using Erlang formulas to calculate the required number of staff for a given forecasted call volume. The main equation (Erlang “C”) has 4 variables; a) Call Volume, b) AHT (total of Average Talk Time and After Call Work), c) Target Service Level and d) number of Agents.
The equation requires 3 out of 4 specific inputs while calculating the 4th one. In practice a user provides volume (based on forecasting), AHT (based on history) and target Service Level (based on center’s long term strategy) to calculate the required number of agents for any given staffing period. If, on the other hand, the number of staff is fixed (or has a limited range) the equation can be used to calculate the potential for the Service Level.

Erlang “C” Equation can also provide theoretical “ASA” and “Occupancy Rate” based on the provided inputs. Using a simple Erlang Calculators (which can easily be found on the internet for free) one can experiment in determining what should be the appropriate target Service Level.
But WFM is more than simply employing Erlang to determine your agents for a day-part. You must also schedule lunches, breaks, scheduled trainings and vacations, and deal with the 2-3% of staff that will not show up for their shift. If you do not account for all of the above in your schedule then you will not have the correct number of agents available when the forecasted calls are received and the center will miss the service level target.
Improving Operational Results
How does a center improve its results as they relate to the Service Level? The most effective approach is a robust WFM process. A competent WFM approach can provide the best and most achievable solution for the forecasted call volume. A well developed schedule ensures the adequate number of agents are available for any given period matching the requirements for achieving the target Service Level. Having said that, we must realize that nothing within a scheduling process can compensate for poor forecasting / planning or unrealistic AHT or Service Level expectations!
While a detailed schedule relies on an accurate forecast to deliver the number of calls, it is up to the agents (and management) to follow the schedule and be available for those incoming calls as predicted. Hence the emphasis on “Adherence” to schedule but that is another story all together.
Lastly, even with a great plan (forecast and schedule), call centers must be capable of tactical adjustments to their operations during the day as incoming call pattern for the day unfolds. An experienced WFM manager with focus on “Intra-Day” adjustments can significantly (and positively) influence the final outcome for the Service Level and the Occupancy Rate.
A little known tip is to use the doubling point to know when, where and what adjustments to make. This is the point of any day where half the calls have arrived. This is derived historical norms for each day of the week. What point of each day has half the volume arrived? Are there the right agents doing the right activities? What adjustments can the center management make?
Balancing Act
Service Level, ASA and Occupancy Rate all provide a view of how efficiently a center is operating. While Service Level and ASA focus on the wait time for the customers, Occupancy Rate is an indication of the wait time for the agents (waiting to receive a call). An experienced WFM manager (or a contact center manager) can provide a balance between needs of the customers (i.e. better Service Level) and needs of the organization (i.e. higher Occupancy Rate) while staying within the boundaries of the center’s limitations such as operating budget and resources.
In the next issue, we will discuss the financial indicators such as Cost per Call (or Cost per Minute) and overall operating expenses.

4 thoughts on “Operational Indicators – Sevice Levels & ASA

  1. Simon Houston says:

    Hi, ASA does not capture 100% of calls presented. This metric only captures calls answered as suggested by the name Average Speed of Answer. By definition ASA is the average amount of time callers spend waiting in queues before being answered.

  2. Sailesh Katara says:

    This has helped me on my presentation and given me many talking points and some innovation idea to improve the process.

    Thank You

  3. […] Moll: Ausbildung im Dialogmarketing Band 3, Bildungsverlag Eins, 1. Auflage 2008, Seite 16 und The Taylor Reach Group INC schreibt im Blog, Zitat: “Target Service Level: Many contact center managers simply assume that a target […]

  4. Dwight Williams says:

    Very useful article. Look forward to the next one as we are a growing inbound contact center.

Leave a Reply

Your email address will not be published. Required fields are marked *