Metrics and attributes for live chat



image avatar

Rob Stack

Zendesk Documentation Team

Edited Mar 20, 2025


0

41

45 comments

Hey Kevin Kiser! With the Incremental Chat Export API, you can find all the pieces of information needed for a chat to find the create metrics that you would see in Explore.

- Ticket brand: With the zendesk_ticket_id property, you can find the associated brand_id from the Ticket object for the given ticket via Tickets API

- Engagement assignment: this can be calculated from the assigned and accepted properties in the engagements object from the Chat object

- Chat completion: this can be calculated as Gabe mentions here

- Chat type: there is a property called type in the Chat object 

- Chat first reply time (sec): this can be retrieved from the response_time object of the Chat object

- Chat no reply time (sec): this can be retrieved from the abandon_time property of the Chat object

0


Hi, I'm seeing a lot of fields here regularly used in internal explore reporting calculations that I'm not seeing in the API docs.

- Ticket brand

- Engagement assignment

- Chat completion

- Chat type

- Chat first reply time (sec)

- Chat no reply time (sec)

To Ramona's point, if these fields aren't in the API then is there equivalent formulas for how to re-create the metrics that depend on them using the fields available to us in the API? For context, we are having issues with Explore not matching our tableau reports. A mapping towards what is available in the API would be especially appreciated! 

Thanks.

1


Hi CJ,
 
Unfortunately, as what you have mentioned, the concurrency dataset is still broken and it's not yet planned on when it will be fixed. However, we have a planned new dataset that should work similarly. Please wait for the update as early as Q1 2023.

-2


Are all concurrency metrics still broken? It's been something like 6 months and there's been no updates or even notes in this article that the entire concurrency dataset is broken. 

0


Hi Kai,
 
If you'll closely check how Chat average reply time is extracted, it is the base value for the metric. It only means that we cannot modify the results being given by that metric even we use some calculated metric.

-1


Dane

Thanks for the explanation. Many of our customers start entering a question before the agent joins, so it is probable that the first reply time is included in most chats - and thus the parallel lines happen - especially since the first reply is usually the longest reply time. We know that we would need more agents to handle incoming chats. 

Is there any way to get a "clean" view on average reply times during a chat? Without the first answering time?
So that we can see if the agents are efficiently working - and to see if the amount of chats one agent is assigned is neither too small nor too big? 

 

Thanks!
Kai

0


Hi Kai,
 
Chat average reply time is calculated as the time it took for agents to reply to visitor comments. Chat first reply time is also included if the a visitor has already sent a message before an agent replied. 
 
Just in case you're assuming that there's a data discrepancy on the query you have created, you can contact support directly and we'll investigate further.
 
Hope this helps.

-1


is it possible that the Chat first reply time (sec) is part of Chat average reply time (sec)? 
When I draw a graph for both (by month), the lines are parallel. 
Something I wouldn't expect unless the first reply time is part of the average reply time. 

Is there a way to get a realistic average answering time within a chat when Chat average reply time (sec) contains the time until the first answer? 

 

0


@Collin, 
 
User information is not included on chat dataset. I understand your use case and how essential it is for data reporting. I'll treat this as a feedback for the continuous improvement of our Explore reporting. I'd also recommend creating a Community post separately for that with your use case to help get more visibility and votes on the idea. Then, others can share their use cases to further drive demand for that feature. 

-2


Is it possible to report on a visitor's email and name if the option to auto-create tickets is disabled? The only similar attributes that I can find in the Engagement dataset are based on the ticket requester. It's odd that I can clearly see user info in the History tab of Chat, but that data is inaccessible in Explore.

2


Sign in to leave a comment.