Insights recipe: Reporting on multi-select fields using ticket tags

Zendesk is removing Insights reporting on February 5th, 2021. If you are using Insights, you’ll need to move to Explore, the latest generation of reporting for Zendesk products, before that time.

At this time, multi-select fields are not synced to Insights. To report on these fields, you will need to use tags instead.

Tags can be tricky to use in reports, since one ticket may have any number of tags at the same time. (For more information, check out this article on Insights tag reporting.) This is especially true for multi-select fields, since one ticket may have any number of tags from the same field at the same time.

Fortunately, as long as the field is set up effectively, you should be able to capture your field tags with one dynamic custom metric.

Skill level: Beginner

Time required: 20 minutes

Ingredients:

• Multi-select fields with consistent tag labels
• One custom metric

Procedure

To report on multi-select fields, you will need to complete the instructions included in the sections below. 

• Add consistent tag labels
• Create a tag metric 
• Build your report

Add consistent tag labels

This approach works best if the multi-select field has the same label at the start (or end) of each tag. For example, in this multi-select field called "Favorite Genres," each tag starts with the string "genre":

 

This makes it much easier to report on the tags as a group. It also makes the tags easy to identify in searches, business rules, and other workflows, so it's a useful best practice to follow.

 

At this time, multi-select fields do not add these labels automatically. They need to be added manually.

 

Create a tag metric using the label

Most of our documented recipes use attribute values. That allows them to be more precise, but they're limited to those values. This recipe uses attribute labels, which allows you to use text strings instead of specific values, making the metric more flexible.

Note: These metrics use the Advanced Metric Editor. For more information, check out our guide on creating custom metrics in Insights.

Create the following metric in the metric editor:

• SELECT IFNULL((SELECT # Tickets WHERE Ticket Tag [Ticket Tag] ILIKE "your_field_label%"), 0)

This counts the number of tickets where the ticket tag starts with the string "your_field_label" and has any number of characters after that. The "ILIKE" part means the text is not case sensitive. For more information on this type of metric, check out GoodData's documentation on Filtering with the LIKE clause.

Note: You can't simply copy and paste in the metric editor. The color-coded items need to be selected from the Elements list on the right:

• # Tickets is under Metrics 
• Ticket Tag [Ticket Tag] is under Attribute Labels > Ticket Tag 

The your_field_label part is typed. It won't appear in the Elements list. You should use the same label that you used in the field. 

For the "Favorite Genres" example above, the field label would be "genre":

Note: This recipe is designed for labels at the start of the tag. If you put the labels at the end of the tag instead, switch the % and field_label in the ILIKE part: Ticket Tag [Ticket Tag] ILIKE "%your_field_label"

The metric may not make much sense on its own; it's a lot easier to understand in the context of a report.

Build your report

To get started, use the new custom metric under WHAT and Ticket Tag under HOW. This is a simple way to see what the metric is doing.

Here is the "Favorite Genres" example:

The report only shows tags that start with "genre," even though there are plenty of other tags in this account. There aren't any extra filters or tag lists here.

As long as the "Favorite Genres" field continues to use consistent tags, they will all show up in this report without any changes to the metric.

This metric also allows you to use ranking filters. For example, here are the top five "genre" tags from the example above:

Note: One ticket may have any number of multi-select field tags. If the same ticket has multiple tags, it will be counted on multiple lines. You can see it if you add Ticket Id under HOW:

In this example, ticket 825 is counted twice, ticket 829 is counted once, and ticket 831 is counted three times. This means the total count on a multi-select tag report will usually be higher than the total number of tickets. 

Restrictions

Ticket tag data is stored in a separate dataset. It is not directly connected to ticket events or properties. This means there are limits to this type of reporting.

Since tags are not connected to event data, there is no way to report on when tags are added or removed. You can only see whether tags are currently present on a ticket. This means you can't use Date (Event) to filter this report; you must use a ticket date like Date (Ticket Created) or Date (Ticket Solved).

The metric in this recipe works with ticket tags, because the # Tickets metric is correctly connected to the tags dataset. If you want to include other metrics in this chart, you would need to make sure those metrics connect to tag data as well. That can get complicated, and the details depend on the metric.

Using ticket tags does limit your reporting options a bit. However, this approach gives you the flexibility to report on multi-select fields with one dynamic metric.