Yo, I am a supervisor for thirty agents and I am seeing some weird behavior in our reports. When I watch the real-time dashboard, I see agents with ‘Time on Hold’ over five minutes. But when I pull the historical interaction report for the same day, those same calls show much shorter hold times. My team is complaining that their stats look wrong. Why would the real-time view show different hold times than the final interaction record?
I deal with this in our bot analytics too. The real-time view is just a snapshot and it sometimes counts things differently while the call is still active. Once the call is wrapped up, the analytics engine re-processes the entire interaction and that is what shows up in the historical report.
It is super confusing for the agents but there is not much you can do about it.
I have seen this on my BYOC trunks when we have signaling issues. If the trunk does not receive the SIP message that the agent took the call off hold, the real-time timer will just keep running even if the audio is flowing. The historical report usually fixes itself because it looks at the actual media segments, not just the signaling events.
It is a known quirk of how the edge servers report to the cloud.
From my experience in our German org, we have found that the ‘Time on Hold’ metric in real-time includes the initial queue wait time in some specific configurations. When the interaction is finished, the system separates the ‘Queue Wait’ from the ‘Hold Time’ correctly. You should check your management unit settings to see how the metrics are being aggregated for your live dashboards.
It is usually just a display configuration issue.