Skip to main content
Docs
Feature Requests Onboarding Trust Center Status Page

Community

Connector Improvement: Zendesk Support Connector Improvement - Ticket Metrics Endpoint missing

Answered
Sandeep Sreedharan User
  • Edited

Zendesk ticket metrics endpoint below are missing with the synced data using Fivetran. There is no way to calculate some of the key metrics - ticket resolution time, average waiting time etc which are important for customer support teams performance reporting. Please find the zendesk ticket metrics end point details : 

https://developer.zendesk.com/api-reference/ticketing/tickets/ticket_metrics/

Please sign in to leave a comment.

Comments

3 comments

Date Votes
  • Official comment
    Fraser User

    Hi Sandeep - the Ticket Metrics are all recreated in the zendesk__ticket_metrics model in our Zendesk Model Package

    Cheers,

    Fraser

  • Zach Brak User

    The goal for everyone here is to have ticket metrics accurate and readily accessible. I agree with taking over end-to-end analysis of the ticket metrics in your own model, as there are issues with Zendesk's own reporting of these metrics.

    However, I got our company to onboard Fivetran in the hopes it would help us make this data more readily accessible. Still, I have found the ticket_metrics is not available in the transformation quickstart.

    I will be forced to continue using stitchdata to move the ticket_metrics endpoint over until I can get a working and reliable model - something I am not getting out of this service right now.

  • Marcello Molinaro User

    Hey @Fraser - We've found that not all metrics can be recreated in the Zendesk package.

    We only get a transcript for the entire conversation for messaging tickets on the ticket_comments table, which means we have bot messages, customer messages and agent responses all bundled up together. Therefore, Fivetran's dbt package doesn’t help, because it relies on ticket_comments for reply time (which works for web / email, but not for chat nor messaging). Although we could transform the data to separate the individual messages, in order to identify the first reply by an agent, we’d have to rely on matching agent name from the chat transcript and on the users table, which is not a robust solution.