Creating Intercom Reports
The purpose of this article is to describe how you can create and export Language IO usage metrics reports from Intercom. Language IO can filter metrics based on conversation attribute values.
In this article
Creating Reports
- Go to the Intercom sidebar and click Reports.
- To create a report, click + New Report in the top right corner.
- Select a template from the list, or click Create your own:
- Add a title and description for the report.
- Now that the report is created, create a chart. You can select a chart template or click Create your own:
- Add a title and description for the chart. Below the title and description, you can select the graph type that best represents your metrics:
- Add metrics (selecting attributes & filters):
- In the dropdown list, choose from the pre-selected list of metrics provided by intercom, such as “New tickets” or “New conversations”.
- To add filters, click on + Add filter, and select one or more filter from the list. Custom conversation attributes appear in this list. Attribute names that start with the prefix "LIO_" hold data that relates to the Language IO Translation integration.
Exporting Reports
- Intercom does not provide the ability to export custom reports as a CSV file. However, there is an option in the reports page named “Dataset export”. This options allows users to select from a list of datasets provided by Intercom. These datasets can in turn be exported as a CSV file.
- Similarly to how you can add filters to charts, you can also add filters to these dataset exports.
- When you are ready to export, click the Export CSV button on the top right to download the file. You can also schedule dataset exports with the Schedule button.
Available Report Types
Reports can be created for the following metrics:
- View the number of chats and tickets that required translation
- Keep track of which languages the agents are speaking
- Keep track of which languages the end users are speaking
- Create reports by group
Number of chats and tickets that required translation
Due to the way Intercom works, all tickets are considered conversations, but not all conversations are also considered tickets. Tickets have separate conversation IDs and ticket IDs, whereas conversations do not have a ticket ID. This means that tickets are included when using the ‘All Conversations’ metric, but conversations are not included when using the ‘All Tickets’ metric.
To compare these values:
- Add either the "All conversations" or "All tickets" metric, depending on the number you need (see above).
- Add the
LIO_Response_IDattribute as a filter. This attribute only has a value on tickets or conversations that had at least one translation. - Select “has any value” from the list of options for the
LIO_Response_IDattribute.
Keep track of languages agents and end users are using
Similarly to the way you can compare all conversations against conversations with a valid project ID, we can compare the use of different languages by using the LIO_Agent_Locale and LIO_Detected_Locale attributes. To do so, you must add each metric and filter manually.
Filters
You must add individual filters for each language or language pair that you want represented in the chart. The locale attribute values are text strings. This means that you can filter to match exactly one of the existing values (for example “French”), or you can use the “starts with” option to group all variations of a specific language within one metric (For example, starting with “Fr” could match “French” but also “French (Canada)”). You can apply multiple filters to a metric. This is useful if you cant to filter a specific locale pair.
When you add filters to an attribute, the available “exact match” options are based on existing values for this attribute, that you can select from a list.
Create reports by group
Intercom stores group identifiers as a custom attribute. This means that you can filter reports and metrics by a specific group (for example user groups).
To report by sub-groups within Intercom, add a filter to the desired metric that asserts a specific value for the selected attribute. You can apply multiple filters to a single metric, like in this example: