This document discusses performance trends, root-cause analysis, and long-term planning.
Analysis use cases
SLA failure root cause. Analysts compare current periods against previous thresholds to pinpoint exactly when service levels dropped.
Short abandon tracking. The Queue interval dashboard is the primary source for identifying callers who disconnect quickly within the IVR. Short abandons are not included in the SLA calculation if you have configured the platform to eliminate them using a threshold.
FCR and training analysis. Managers track transfer frequency and Third-party time to identify training gaps or routing errors.
Trend identification. Planners use drillable data points to identify recurring daily or weekly call patterns.
Key metrics
Short abandons (from Call Queue Metrics (Historical)). Callers who disconnect within a narrow time window (for example, 5-10 seconds). A high volume of short-abandons often indicate a confusing IVR or technical issues
Third-party time (from Third Party Calls (Historical)). The amount of time that interactions are connected to a third party. Essential for tracking the cost and efficiency of external handoffs.
Queue Entries (distinct interactions) (from Call Queue Metrics (Historical)). The number of end-users that enter the contact center through a queue. This number doesn't include transfers. This number provides the most accurate view of actual customer volume and demand.
Strategic investigation: volume metric reconciliation
Sessions versus interactions. A session represents the total system touchpoint. Queue interactions include the entry plus other touchpoints like the Virtual assistant.
Entries versus interactions. A Queue entry counts a call regardless of transfers. A Queue interaction includes transfers and touchpoints.
Unique volume. Use Queue entries to see unique customer volume without transfer inflation.
Complex troubleshooting: identifying ghost SLA drops
Short abandon logic. Check for Short abandons; they are not included in Queue entries (both from Call Queue Metrics (Historical)) and strictly excluded from SLA calculations.
Transfer loops. Review the Transfers dashboard for the same Call ID across different queues; this indicates the interaction is "bouncing" between queues.
Deflection patterns. Review data in the Deflections dashboard. A callback deflection won't be identified in the perf va table because the virtual agent drops the call into a queue first.