Service level target is a contact center performance statistic expressed as “X percent of conversations answered in Y seconds.” For example, a goal of answering 80% of voice calls within 20 seconds provides a service level target of 80/20. You can set the service level target for each media type in a queue. Analytics uses this service level target to calculate whether you are within the service level % defined by your service level agreement (SLA). For more information about setting the service level target, see Create queues or Edit or delete queues. Note: In the DNIS Performance Summary view, theService Level % is calculated by queue and will display as a – on the DNIS group row. To view individual queue service level %, click the + to expand the row.
The service level % is the percentage of interactions that meet your contact center's service level target. The formula to calculate the service level % is:
Number of Conversations Answered within SLA
The numerator is calculated by taking the total answered conversations and subtracting the conversations that were not answered within the SLA threshold.
Calculated by: (Answered Count – SLA Violation Count)
Total Number of Conversations
The Service Level Options configuration determines the denominator. With all the options below unchecked, the denominator is the count of answered conversations.
- Include Flowouts in Calculation:Flowouts are conversations that leave a queue without abandoning and without an agent handling the interaction. Typical scenarios include a customer on a voice call accepting a callback or an in-queue call flow transferring the call to another queue/flow. With this option enabled, the total number of flowouts is included in the total number of conversations for the SLA calculation.
- Include Short Abandons in Calculation:Short abandons are abandons that happen within a specific time threshold. For more information, see the Configure abandon intervals section in Configure contact center analytics options. When enabled, the SLA calculation makes no distinction between an abandon vs a short abandon. All abandons are considered in the formula. When disabled, the short abandons are subtracted from the abandon count in the denominator.
- Include Abandons in Calculation: Abandons are conversations where the waiting party disconnected without being answered or transferred.With this option enabled, the abandons are included in the total number of conversations.
As the service level calculation depends on an organization’s configured options, the final formula to calculate the service level is:
Note: With the Flowouts and Abandons enabled, the Flowouts and the Abandons are added into the denominator when calculating the service level. With the Short Abandons disabled, the Short Abandons are subtracted from the Abandons when calculating the service level.
The number of interactions routed into the queue for an agent to answer. Offered interactions are either answered, abandoned, or flow-out interactions.
A measurement of the number of assigned ACD interactions answered by an agent in a givenperiod. Answered increments when the agent is the first to connect to the interaction.
The number of times an end user abandoned an interaction in a queue. For example, the customer hangs up before an agent answers, or the system disconnects the interaction.
The average amount of timeagents spent handling interactions. This calculation includes talk time, hold time, and after-call work. For outbound calls it also includes dialing and contacting time. Calculated by: (cumulative talk time + cumulative hold time + cumulative after-call work + cumulative dialing time + cumulative contacting time) / Number of interactions handled Note: The interactions can span multiple time intervals. For example, a 45-minute call spans two 30-minute intervals. However, metrics are assigned a single timestamp and reported in the interval into which this timestamp falls. The timestamps are assigned based on when the metric calculation is complete. For instance, handle time will use the timestamp of after call work completion. In the example of a 45-minute call, the handle time falls in the second 30-minute interval because the handle time was calculated at the end of the call when ACW was submitted.
Scheduled agent count is calculated by:
- Summing up the scheduled on-queue time for all agents associated with the business unit and the selected planning group(s), and then dividing by the corresponding interval length.
Actual agent count is calculated by:
- Querying aggregated agent metrics for the agent routing status for all agents associated with the business unit and the selected planning group(s).
- Looking at the routing status per interval and count on-queue time when routing status is either interacting, idle, or communicating.
- Summing up the on-queue time for all agents, and then dividing by corresponding interval length.
Example: If an agent status is interacting, idle, or communicating for 7.5 minutes of a 15-minute interval, then the schedule shows 0.5 agent during that interval.
When an interaction finishes, the system marks it as completed.
Note: Because an interaction may be completed in a different interval than the interval in which it starts, Completed may differ from Offered.
The average amount of time an interaction waits in queue before an agent answers it. This metric does not include interactions that agents do not answer. ASA is directly associated with the service level and provides insight into the customer service your company provides. ASA is recorded in the interval in which the agent answered the interaction. Calculated by: Total Answered (tAnswered Sum) / Answered Count (tAnswered Count) Note: Raw values are provided in milliseconds.
This metric represents the total time that agents actually spend handling interactions:
- Predicted calculation is the total time predicted to be in Interacting routing status / total scheduled on-queue time in interval.
- Actual calculation is the total time in Interacting routing status / total actual on-queue time (Interacting + Communicating + Idle routing status time) in interval.
