Pending Status Is Hiding SLA Breaches
Northwind sells B2B software and receives about 5,200 tickets a month. Its SLA policy promises a 2-hour first response and 24-hour resolution during business hours. Agents are trained to set a ticket to Pending whenever they reply and are waiting on the customer. In a sample month, 78% of tickets entered Pending at least once, and 31% entered Pending within four minutes of the first public reply. The SLA dashboard shows 97% compliance, but the customer success team reports that key accounts wait three or more days for answers. A 30-ticket audit finds Pending durations of 6–9 days on tickets that were never followed up, and a macro named 'Waiting on customer' sets Pending and sends a generic reply. Leadership wants to remove Pending from all SLA calculations. You need to separate legitimate customer-wait time from avoidance behavior without punishing agents for truly waiting.
1. Quantify the problem and determine whether Pending is being used as designed. What reports and fields would you examine?
Read the worked answer
Build a report at ticket-event level, not just current status. For each ticket, calculate: time in Pending, number of Pending transitions, who made each transition, time from Pending to next agent reply, and total agent-owned time. Join the audit log for status changes and the macro usage log if available.
Segment by macro, agent, and customer. A legitimate pattern is: public reply asks a specific question, status goes Pending, customer replies, status returns Open, agent responds. A suspicious pattern is: status goes Pending within minutes of a generic macro, no specific question, no follow-up after 48 hours, and the ticket eventually closes without a resolution.
The 31% figure is the key: if nearly a third of tickets enter Pending immediately after a generic reply, the macro is being used as a workload shield, not as a customer-wait marker. Compare the 97% SLA number with median time-to-substantive-answer and with the 6–9 day Pending durations. That gap tells leadership the dashboard is measuring a paused clock, not customer experience.
2. Design controls so Pending remains a legitimate pause but cannot hide neglect.
Read the worked answer
Keep Pending as a pause, but make it a controlled state. First, change the macro so it cannot set Pending without a public reply and a required 'waiting reason' field (for example, customer_info, approval, third_party). Second, add automation: after 48 hours in Pending, send the assignee a reminder; after 5 business days, set status Open, reassign to the group lead, and add a private note. Third, build a manager view of tickets pending more than 72 hours with no outbound reply after the Pending event.
Restrict who can bulk-set Pending, and review macro usage monthly. Do not simply remove Pending from SLA calculations: agents will switch to Resolved or Closed to stop the clock, and you will lose the distinction between customer wait and agent neglect. Instead, separate contractual SLA reporting from operational 'agent-owned time' reporting. Coach on the difference between waiting for a customer and parking a ticket.
3. How should leadership report true responsiveness without redefining SLA in a way that misleads customers?
Read the worked answer
Report two layers. The contractual layer is the SLA policy as written: 97% compliance, with a footnote that Pending pauses the clock. The operational layer should include first response time, time to first substantive answer, agent-owned time, number of agent touches before resolution, and percentage of tickets pending more than 48 hours without follow-up. Present them side by side so leadership can see that the contract metric and the customer experience metric answer different questions.
If customers are promised a 24-hour resolution, a 6-day Pending period is still a failure even if the SLA timer was paused. The fix is governance and follow-up automation, not redefining the contract retroactively. Publish the operational dashboard to the support director weekly, review the top 20 aging Pending tickets, and set a target such as 'no ticket remains Pending more than 3 business days without an agent touch.' That gives agents a legitimate wait state without letting it become a hiding place.