matthew | 2016-11-30 20:40:43 UTC | #1
Hi, the following page lists the dimensions for use in the aggregations and details queries :
https://developer.mypurecloud.com/api/rest/v2/analytics/dimensions.html
but when i try to add certain filters like purpose, or disconnecttype an error is returned saying
Dimension disconnectType is not valid in context AGGREGATION
Can you confirm that the list of dimensions on the above link is correct and all available for use in such queries?
my example query :
{ "interval": "2016-11-30T00:00:00.000Z/2016-11-30T23:59:59.999Z", "groupBy": [ "dnis" ], "filter": { "type": "and", "clauses": [ { "type": "and", "predicates": [ { "type": "dimension", "dimension": "purpose", "operator": "matches", "value": "Acd" } ] } ] } }
Matthew
tim.smith | 2016-11-30 20:54:12 UTC | #2
That list of dimensions is the complete list that may occur across all analytics resources. Currently, some of the limitations are documented on the pages about each type of query, but there are gaps in that documentation. The Analytics team is working on documenting which dimensions and metrics can be used with which queries.
matthew | 2016-11-30 21:25:31 UTC | #3
thanks for reply. on the following page :
https://developer.mypurecloud.com/api/rest/v2/analytics/index.html
it also says those dimensions are valid for use in the aggregations and conversation details :
If I can't use such filters then it seems i have to do a full export of the data, and do a manual calculation? With that said, can you confirm :
Time caller waited before abandoning a call: Segment Time from call segments where Purpose='Acd' disconnectType='Peer'
Time to Answer a call : Segment Time from call segments Purpose='Acd' disconnectType='system'
Can you confirm this logic is correct, and i can then do a export of the call segments, and do the calculation outside of purecloud.
thanks!
tim.smith | 2016-11-30 21:32:04 UTC | #4
I think your assumptions should be correct, but a more correct way of determining abandons vs assignments is to compare the ACD segment with the other segments. If the ACD segment is the last segment in the conversation (based on its end time) and the conversation is disconnected, then it was an abandon. If there are other segments that occurred after the ACD segment or the conversation is still active (analytics queries include active conversations), then it wasn't an abandon because something else happened.
matthew | 2016-11-30 22:09:38 UTC | #5
Thanks, thats helpful. In a future release would it be possible to include the conversation statistics in the results for each conversation? that would save these external calculations from being needed.
tim.smith | 2016-11-30 22:10:52 UTC | #6
There is a project being worked on that will provide metrics at the conversation detail level. I believe it's currently targeted for early next year.
matthew | 2016-11-30 22:21:12 UTC | #7
Excellent. That will be great.
thanks.
anon28066628 | 2016-12-01 15:40:15 UTC | #8
Is there information available about the PureCloud built-in reports and dashboards calculate abandons, ie do they look for 'acd' participants with a segment disconnected by 'peer', or do they look for the acd participant having the latest end-time of any segment?
tim.smith | 2016-12-01 15:42:25 UTC | #9
The PureCloud UI uses these metric definitions: https://help.mypurecloud.com/articles/metric-definitions/
anon28066628 | 2016-12-01 18:59:50 UTC | #10
Based on those definitions of Abandoned, Abandon %, and Answered (mentions of "Agent" and "Queue"), is it correct that customers who disconnect in the IVR don't count as Abandons?
If I wanted to calculate "IVR abandons", I could check if an IVR participant has a segment with disconnectType: peer. Or would it be better to check the end times of all segments and see if the IVR was the last one?
tim.smith | 2016-12-01 22:59:48 UTC | #11
anon28066628, post:10, topic:687
is it correct that customers who disconnect in the IVR don't count as Abandons?
Right. Abandons refers to conversations that were in an ACD queue and disconnected without an agent ever being connected. Calls disconnecting in the IVR were never in a queue.
anon28066628, post:10, topic:687
If I wanted to calculate "IVR abandons", I could check if an IVR participant has a segment with disconnectType: peer. Or would it be better to check the end times of all segments and see if the IVR was the last one?
I think you'll be ok with the first option, but the second is the most correct. A call could go to an agent and be transferred back to the IVR for a second time.
system | 2017-08-28 19:29:10 UTC | #12
This post was migrated from the old Developer Forum.
ref: 687