I am currently working on a series of executive dashboards and I have found a significant issue. When I query the /api/v2/analytics/queues/observations/query (Aggregates) I see 45 abandoned calls for yesterday. However, when I pull the raw conversation details for the same period and filter for ‘abandoned’, I only find 38. This is a nearly 20% difference. Can someone explain why these two endpoints do not agree on the basic count of abandoned interactions? It makes our reporting look very unreliable to the leadership team.
I took over our org six months ago and found the same thing. The issue is how the system defines an ‘abandon’. Aggregate metrics often count short-disconnects or calls that abandoned in the IVR before even reaching the queue, depending on your filter.
Detail records require you to check the segment level. You are likely missing calls that were transferred out of the queue but the customer hung up during the transfer leg.
Hey! I am not a developer but my team sees this all the time when the internet is slow. We have agents who say the call just vanished. I think the system gets confused when the connection drops. It is very annoying for the supervisors who have to explain the numbers to the boss.
It is a common point of confusion for those new to the platform. The Analytics Aggregate API uses a different data set than the Conversation Detail API. Specifically, the ‘tAbandoned’ metric in aggregates includes calls that disconnect during the alerting phase, whereas your detail query might only be looking for segments with a specific disconnect type.
I recommend reading the ‘Metrics Definition’ article in the Resource Center which explains the exact calculation for each endpoint.