Question
What options do I have to report on different ticket comment activities in Explore?
Answer
The metrics Agent replies and Public comments both relate to how different actors add comments to a ticket. These two metrics have many differences including that they are in two different datasets.
Public comments
Updates history dataset
- The metric Public comments is within the Updates history dataset. Public comments represent all public comments on a ticket, including both end user and agent public comments.
Updates history dataset covers individual ticket events. This dataset is designed for narrow, targeted reporting rather than broad trends. For example, it can be used to filter by Updater attributes, such as Updater Name, to report on the number of updates made by individual users or agents. For more information see article: Counting agent comment updates
- Other metrics within this dataset are Internal comment, Agent comment, and End-user comment.
Agent replies
Tickets dataset
- The metric Agent replies represents all public comments added to a ticket by an agent, excluding ticket creation. For example, when an agent creates a ticket with a public comment, that first comment isn't counted. This may result in the public agent comment total being higher than the agent replies value.
The Tickets dataset covers ticket-level metrics, which are designed to make high-level, overview reporting more efficient.
- The metric Agent replies is calculated differently for messaging channel tickets and in addition to all public comments added to a ticket by an agent. It also calculates based on the number of active and messaging conversations that include at least one agent message. This may result in the agent replies metric being lower than the combined total of agent messages and public agent comments. For example, if an agent sends three messages during one active conversation session and then the conversation ends, this will be calculated as a single agent reply.
For more information, see the articles:
Messaging dataset
- The metric Agent replies represents the number of replies to the messaging ticket by an agent. Replies are defined as messages sent by the agent in response to an end-user message. For example, when a user sends a message and an agent replies, this is recorded as a single agent reply. Any additional messages sent by the agent in the same conversation will not increment the value until the user sends a new message.
- Other metrics within this dataset are Agent messages, Requester messages and Requester replies.
More context
When analyzing ticket comment activity, you can report on who posted the comment, the agent or the end-user, and you can report on how that comment was posted to the ticket, the comment is public or internal. Only agents can post internal comments. To learn more, see the article: Adding comments to tickets.
The following metrics allow you to report on ticket comments:
In the Updates history dataset, there are:
- Agent comment
- End-user comment
- Public comment
- Internal comment
In the Tickets dataset, there are:
- Agent replies
In the Messaging dataset, there are:
- Agent replies
- Agent messages
- Requester messages
- Requester replies
For more information, see the article: Understanding Explore datasets.
20 comments
Dominic
Hi Elissa,
How does one find the "Agent replies" metric? I am only able to find "agent replies brackets"?
Kindly,
Dominic
0
Chandra Robrock
@... If you go to the Tickets dataset, you should see Agent Replies as a Metric under the Activity section. Hope that helps!
0
Stéphanie Cettou
Hello,
it seems that for a defined period, my "Public comments" filtered on Admin and Agent are different with the "Agent replies" data.
Where the difference could come from ?
Thank you,
Stéphanie
0
Chandra Robrock
Hi @... - Would it be possible to confirm which Role attribute (i.e. Updater Role, Submitter Role, Assignee Role, etc.) you've filtered your Public Comments query by?
Also, what date attribute are you using to filter both the Public Comments query as well as the Agent Replies query?
0
Blanca
Hi Stéphanie,
Thanks for reaching out. So, the Agent Replies metric is the number of public replies added to a ticket by an agent and this metric is found under the Tickets dataset. While the Public Comments metric is the number of public comments on tickets both made by an Agent and an End-user and this metric is found under the Ticket updates dataset.
I hope this answers your question. Please feel free to reply for any clarification.
Cheers,
Blanca | Customer Advocate
Chat with our live support!
https://support.zendesk.com/hc/en-us
0
KROB
How do I create a report to count the total number of public replies an agent sent? It seems like this is based on a D Count metric instead of a total number. I've followed your recipe this way but it's not giving the correct data:
1
Nick Olson
Thanks for reaching out! Could you elaborate on what aspect of the data is showing as incorrect? Which ticket channels are you looking to report on total number of agent public replies for? Also, which recipe are you referencing so I could take a closer look at where there may be potential discrepancies?
I'm looking forward to your reply!
Regards,
0
Abed Islam
I want to emulate Tickets dataset reports within the Ticket updates dataset. What is the exact formula for Agent replies? In the Ticket updates data set, if I try the below where the updater role is either admin or agent:
I get a slightly higher number than if the updater role is simply not the end-user:
I then try to calculate which of these are opening posts (so they're not replies) via:
If I subtract the Agent opening public comments (above) from the Agent public comments formula (either of the first 2), I get a number higher than Tickets' dataset's Agent replies.
I'm at a loss here regarding these discrepancies. Please advise.
0
Amy Dee
Hi Abed! Thank you for providing these metrics! You're on the right track.
The Agent replies metric in the Tickets dataset counts all public agent comments excluding ticket creation. If an agent creates a ticket with a public comment, that first comment will not be counted. That means a count of all public agent comments in the Ticket Updates dataset will likely be higher than the Agent replies metric in the Tickets dataset.
I see you've accounted for that in your third metric, Abed. That's the approach I would take to align the Tickets and Ticket Updates results.
If your results are still off, there are a couple possibilities that come to mind. First, public comments by a system user may skew the results. The system user may count differently depending on which channel it's using, so there may not be an easy way to account for it. In the Ticket Updates dataset, the system user has a NULL role.
Second, the updater's role may have changed between when the comment was written and when you ran the query. The Agent replies ticket metric is logged at the time of the comment, and it does not change later. (It's just a number in the ticket's metric set; Zendesk doesn't track who made each agent reply.) Meanwhile, the Ticket Updates dataset may pull the updater's current role, rather than their role at the time of any given update. That could also cause discrepancies.
If you're seeing significant discrepancies that can't easily be explained, you may want to contact our support team. That would allow us to look through specific examples and narrow this down.
I hope this helps! Happy reporting!
0
Natalia Lutsevich
Are you sure about that? I did a report on Zendesk Explore and it turned out that Public comments include only agent's replies.
0
Alex Zheng
I will open up a ticket with you in order to investigate further.
Best regards,
0
Amy Dee
Hi Natalia! The pre-built Public comments metric in the Ticket Updates dataset uses this syntax:
It counts all updates that contain a public comment. It does not include any parameters for the commenter's role, so it will return all public comments, regardless of who made them.
There is a separate Agent comments pre-built metric that excludes end-user comments. It looks like this:
It's possible that you have a customized version of the public comments metric, or that your query has other filters that limit the updater role. My colleague can help you pin down the details.
In general, though, we can confirm the pre-built Public comments metric does include all roles.
I hope this helps! Happy reporting!
0
Stéphanie Cettou
Hello,
in the default query "Agent updates: Agent updates [default]" there is a filer on "Updater role" for Admin and Agent. This would meand that for this query "Public comments" would only refers to comments applied by Admin and Agents. Right? Excluding End-Users.
To summarize, it would be the equivalent of the Agent comments. Is this correct?
0
Alex Zheng
Right in this case since we have specifically selected for agent roles the public comments metric for this query would be referring to agent comments only.
Best regards,
0
Abie Paganpan
Hi,
I am trying to extract a report on the total replies sent by an agent on a weekly basis. However, I am confused with the term "Agent Replies" and Public Comments".
Can you help me explain this? Also, when I get the total number of tickets created for the week, it does not match the number of replies that I am getting.
0
Jupete Manitas
Hi Abbie, thanks for writing in!
The comparison between the two is that Public comments will give the total number of public comments from both end-users and agents. While there are specific metrics that you can use to count agent comments which are the Internal comment and Agent comment.
In your screenshot, you are using the Agent updates metric. Per the description below:
You can see that here. So there will be differences between Agent updates and Public comments since they query different data. We hope this helps!
0
D.Fitz
Hi,
I'm getting inconsistent figures between Agent Replies and Public Comments, despite filtering accordingly as per above. The figures are always a little out (never more than 10).
What's happening here?
0
Elaine
I recommend reaching out to Zendesk Customer Support to investigate the discrepancy you've identified in the report you created. This way, we can work together to identify and address the issue.
0
Tahlia DeMaio
I'm finding that all comment metrics are grossly under-reporting the number of responses for all of our Messaging, live chat, tickets. Should I not be expecting an "end user comment" to include every individual message a user "hits enter on" in the live chat window? Similar for "agent comments" -- does that include _every_ time they hit enter and send a snippet of a response to the chat?
0
James G
You're absolutely right. We can't expect an exact match between the number of messages sent by end-users via live chat or messaging channels and the ticket comments metric in Zendesk Explore. This is because the messages sent via live chat or messaging tickets are added as bulk ticket comments. For instance, if multiple messages are sent via the Messaging channel, they will be added as a single ticket comment in the ticket. This would cause the metric comments to count as a single comment instead of multiple messages. If we want to count the actual number of messages sent via live chat or messaging, we should use metrics like Agent messages in Messaging tickets metrics or Chat messages in Engagement metrics.
0