At a high-level, an agent touch is an operation performed by the agent on a ticket. For example, it could mean that the agent:
- Added a public or private comment to the ticket
- Changed a ticket field, for example the ticket group
- Changed the status of the ticket, for example from Open to Solved
The method you'll use to determine agent touches will depend on your business needs. In this article, you'll learn about how you can use Explore to analyze agent touches in a few different ways. The examples assume you have some familiarity with creating Explore reports. If you need some help, see Creating reports.
This article contains the following sections:
Analyzing ticket comments
If you want to measure agent touches by the number of comments your agents make on tickets, then Explore has you covered. The Updates history dataset contains a built-in metric that measures this when you add it to a query. You can add attributes to slice this number however you want, for example by ticket ID and subject.
This simple Explore report displays a table showing your tickets and how many agent comments they received.
To create a report to analyze ticket comments
- Create a new report using the Support: Updates history dataset.
- In the Metrics panel of the report, add the Comments >
Agent comments metric. Explore displays the total number of agent
comments received for all of your tickets.Tip: The Explore Updates history dataset might contain a lot of data causing your reports to take a long time to return results. Consider adding a filter like Time - Ticket update > Update - Date to reduce and focus the results returned by the report. For more information, see Working with report filters.
- To break down the number of comments by ticket, add the attributes Ticket
ID and Ticket subject to the Rows panel of the
Explore displays a table showing all of your tickets broken down by the number of agent comments (each comment counts as one agent touch).
Analyzing ticket updates
Sometimes, you might want to report all updates made to your tickets, not just comments. For example, other updates might include changes to ticket status, priority, assigning the ticket to someone else, and more. Again, the Updates history dataset contains a built-in metric that measures this when you add it to a report. You can add attributes to slice this number however you want, for example by ticket ID and subject.
In this example, you'll create a report that shows all updates to all of your tickets. You'll then add attributes to slice this number by the date of the update, the ticket ID, and the person who made the update.
- Create a new report using the Support: Updates history dataset.
- In the Metrics panel of the report, add the Updates >
Updates metric. Explore displays the total number of updates made
for all of your tickets.Tip: The Explore Updates history dataset might contain a lot of data causing your reports to take a long time to return results. Consider adding a filter like Time - Ticket update > Update - Date to reduce and focus the results returned by the report. For more information, see Working with report filters.
- To break down the number of updates, add the attributes Update -
Date, Ticket ID, and Updater name to the Rows panel
of the report.
Explore displays a table showing all of events where at least one ticket update was made, broken down by date, ticket ID, and the name of the person who made the update.
Analyzing tickets with low agent touches
Explore records the number of agent replies it took to solve a ticket into the following brackets:
- One-touch tickets: Tickets that were solved with only one agent reply.
- Two-touch tickets: Tickets that were solved with two agent replies.
- Multi-touch tickets: Tickets that were solved with more than two agent replies.
In this example, you'll create a report that shows the number of one-touch tickets created each year by your agents.
- Create a new report using the Support: Tickets dataset.
- In the Metrics panel of the report, add the Agent replies distribution > One-touch tickets metric. Explore displays the total number of one-touch tickets in your account.
- To break down the number of one-touch tickets by year, add the attribute
Ticket solved - Year to the Rows panel of the
Explore displays a table showing the number of one-touch tickets in your account each year.
Can you please define what is counted as a "Touch" in the following metrics? I have reviewed our metrics and based on the initial percentages, it appears this only counts public agent replies.
Hi Chanteena! Taking a look at the formula used for the Agent replies distribution > One-touch tickets, Two-touch and Multi-touch metrics, I can confirm that these metrics are based on Zendesk's Agent Replies metric.
The Agent Replies metric will only include the number of public replies added to a ticket by an agent. This help article is a really a helpful resource for identifying what each metric means in Explore.
Hi! Is there a way to understand ticket touches for tickets submitted by an end user, not just those created by agents? I'd like to understand the number of tickets submitted to our team and the breakdown of touches between the ticket's submission and it being submitted as closed.
Hi Octave! These queries should return results for both tickets that were created by the end-user as well as the agent.
However, if you're ultimately looking to report on the number of touches per ticket for both the end-user and the agent, there are a few other comment-specific metrics that you could look at.
You'll find these metrics under the Comments section in Metrics.
Hi team - I'm just looking into tickets that may have multiple ticket touches, do you have any "How To" documentation on how to create a trigger for tickets with 4+ ticket touches, so I can create a view with all those tickets?
Hi Matt - If you'd like to create a View based on multiple touch tickets, the best way to accomplish this would be to first create a trigger that adds a tag based on the number of Agent Replies.
The Trigger itself would look something like this:
From there, you could then build a View that looks for tickets with the specific tag you're adding in that trigger and any other criteria you'd like to build into this view (i.e. ticket status).
Hi I have a query about what counts as an agent comment. Is this just public agent comments or would it include internal comments too?
Hi @... - The Agent Comments metric would include both public comments left by an agent as well as internal comments left by an agent.
If you'd like to only include public agent comments, you could either add a Filter to your query for Comment Public = True when using the Agent Comments metric, or you could build a custom Standard Calculated Metric that essentially combines Zendesk's out-of-the-box Public Comments and Agent Comments metrics together.
The actual query for the custom metric would look something similar to the below:
This is great Chandra thank you so much :)
Sorry another question. Is there a way of showing this result as a % average for a group? I have filtered for the group I need, but cannot seem to get an average % of agent public comments for the group or per agent.
Thanks in advance for your help
@... Happy to help! To find the average % of agent public comments for a specific group, you could build something similar to the following.
- Select either Agent Comments or the custom metric you created
- Select the exact metric you'd like to include in the denominator (ex: Ticket Created, Ticket Updated, etc).
It sounds like you're filtering for a specific group already but, if you're not, you could put the Group attribute in the Rows section. This is helpful when wanting to run an analysis for multiple groups at once.
From there, you can select Result Manipulation (the arrows on the right-side menu) and select Result Metric Calculation. This allows you to create a new metric based on two existing metrics. The actual query would look something like this - COUNT(Agent Comments)/COUNT(Tickets created) - depending on what metrics you've chosen to analyze.
Under Chart Configuration (the paintbrush icon on the right-side menu), you can select Display Format to specify a % as the format for this new metric. Hope that helps!
Hey Jackie - Wanted to follow up as I think I may have misunderstood what you were looking for last night. Were you hoping to understand what % of total comments each agent within a specific group accounted for?
If so, here's how you could analyze that:
Metrics: Select either Agent Comments or the custom metric you created
Rows: Updater Name (you'll likely need to filter your query by the Updater Role != End User, if you aren't already)
From there, you can select Result Manipulation (the arrows on the right-side menu) and select Result Path Calculation. You'll then select % of Total on Rows as shown in the below screenshot:
Hopefully that gets you closer to what you were looking for! :)
I think that there is a mistake, see this screenshot.
Hi Mirek Kokes, thanks so much for spotting that mistake. It's now fixed!
Hi guys, I am also trying to create a report to sum and average both agent and requestor comments in the ticket based on a given period of time. Our goal is to improve efficiency and find ways to lower the amount of back and forth it takes to get to a resolution and I need a way to measure that. When I look in Metrics for the word comments, this is all that comes up. Am I missing something? Thanks!
Kenny Rohan Mind confirming that you're in the Support - Updates history dataset rather than the Support - Tickets dataset? Based on the metrics included in your screenshot, I suspect you may be querying the Support - Tickets dataset instead.
That was it. Thank you very much!
Is there a way to measure average solve time for tickets with public replies only? It seems like there is no way to use the data from support tickets in support updates history.
Hey Landry Norman -
Applying the agent replies bracket filter on your report and excluding NULL and 0 should do the trick:
In my org, our agents aren't currently assigned tickets, so we all touch them. Is there a way to measure Unique touches per ticket. For example.
If a new ticket is run from Creation to Close and 10 agent touches were made, but only from 5 distinct individuals, how could I create a report showing that 5 agents touched that case, not that there were 10 touches...
More specifically I'd like to know what the Average number of Distinct agent touches were on tickets.
The goal is for us to trial a new ticket handling process and be able to show if the new process reduces distinct agent touches, increasing ticket "ownership" without it being tied directly to whether a ticket is assigned or not (As even an assigned ticket can still be touched by multiple people)
You can try using a custom metric to count the agents who made at least one update on a ticket. Here's a sample formula:
You the metric aggregator D_COUNT with the metric; with your data sliced by Ticket ID, your table should show the number of agents per ticket.
The D_COUNT option allows me to get a distinct count of agents/Admins on a case, so this works. I can break it down by tickets which gives me a good look into the tickets that have lots of people involved.
I'm also trying to break this down further to get the Average of the Distinct count of agents per ticket. So kind of a:
D_Count(Agents updating a ticket) / # of Tickets
Does that make sense?
To get that average, you will need to calculate the value through Totals while the data is sliced by Ticket ID. Select AVG for the totals aggregator and the table should display this at the bottom row.
By default, Explore will display a rounded value. If you need the average to be displayed with decimal values, then you need to change the display format for the metric (which will also affect the display format of all rows). Example:
Hi there - how do I make a report that shows the volume of outbound emails an Agent sends? We often create outbound emails from webinar sessions we host and I need to see how many we're sending out to customers. Thank you!
Riane Williams This help article should be helpful for your use case: https://support.zendesk.com/hc/en-us/articles/4408846777754-Explore-recipe-Tickets-created-by-agents-or-end-users. Specifically, the Tickets created by agents calculated metric.
Please sign in to leave a comment.