Text Analytics FAQs

Last updated
Save as PDF

Gainsight NXT

Text Analytics FAQs

Below are some frequently asked questions about Text Analytics .

How can I enable the multi-channel source support in Text Analytics?

The multi-channel source support is available as a paid add-on feature. For assistance on how to enable this in your org, contact Gainsight Support.

I have the Text Analytics feature in my org, but still don’t see the Text Analytics tab in Salesforce App's Navigation Menu, are there any more configurations required for this?

The NXT Admin of the org needs to perform the following actions, to ensure the accessibility of the feature to the intended audience.

Grant access to the Text Analytics Navigation menus in the permission bundle to the intended users.
In the same permission bundle, also grant the module level Survey Analyze access to the intended users.

I have configured the initial data setup in the Text Analytics Admin configuration, but why the widgets in the Text Analytics dashboard are empty?

Text processing considers the data that belongs to the Active company records. Hence, ensure that you are collecting the textual data or notes for an Active Company record.

Why doesn’t the number of entries displayed in the Revenue Matrix widget match with the number displayed in the Companies widget?

The Revenue Matrix displays a maximum of 50 companies per quadrant and, If the company record is NOT assigned with a value from the respective revenue field, such company records will not be visible in the Revenue Matrix widget. There is no such restriction in the Companies widget.

My org has an External NPS® survey with many external responses. Why am I not able to see those entries in the NPS® tab of the People widget?

You may have an NPS® collection populated with the external NPS® scores (NPS® upload through CSV). It is important to note that such entries will be populated to Text Analytics People widget NPS® tab, only if a genuine Company Person record exists for each external participant record (in the CSV entry).

I have added a topic in the Text Analytics Admin page. Why am I not able to see it in the dashboard?

It takes nearly 4 hours to process the text data and build analytics. One of the reasons could be that particular topic has not been processed.

How much time will it take to process the text data and to build analytics?

Once the data is configured, it takes nearly 4 hours to process the text data and to build analytics. The analytics are refreshed every 4 hours. However, the time required to process the text directly depends on the volume of the text.

Can I see the Timeline tab in the People widget drill down view?

No, you can only see the Survey tab in the People widget drill down view.

How many Topics or Keywords can I add in the Text Analytics Admin page?

There is no limit. You can add as many Topics or Keywords as required. The results will be better if you add more Topics or Keywords.

How many months of historical data can I select while configuring the Time Frame?

You can select a maximum of 18 months historical data.

Can I apply global filters on the selected view in the Text Analytics home page?

Yes, once you apply global filters, all the Widgets display the Sentiment score, NPS®, Number of Mentions and the Number of Comments based on the filtered criteria. For more information on the global filters, refer to the Text Analytics Home article.

Can I share my view with others?

Yes, You can share your view with others (Public, Users and User Groups). You can assign View, Edit or Admin access to the shared users.

How is data permission managed in Text Analytics with certain limitations?

In Text Analytics manage data permission to enhance security and ensure compliance. Administrators can apply permissions at both company and other MDA objects to ensure that users access only authorized data. For more information on configuring these permissions, refer to the Data Permission article.

However, certain field limitations in Text Analytics are not governed by Data Permissions. The fields listed below in the CX Center Text Analytics View do not adhere to Data Permissions settings:

Activity Id
Activity Type
Survey Participant Id
Language Code
Text
Translated Text

Note:

Users can access only the data permitted by their assigned permissions. Any data linked to restricted MDA objects is excluded from Text Analytics, ensuring that sensitive information remains secure and visible only to authorized personnel.
Data permissions are not supported for Key Driver analysis along with a summary in the Highlight section. Thus, all company and other data types are included in the aggregated analysis.
CSAT Survey Response, NPS Survey Response, and CX Center Text Analytics are the MDA objects that support data permission in Text Analytics.

Natural Language Processing

How is sentiment computed in Text Analytics ?

Sentiment in Text Analytics is computed using state of the art Natural Language Processing, primarily at 2 levels: Keyword and Comment/Post.

Keyword: Text Analytics computes the sentiment associated with a particular keyword within the text data. This is what is shown whenever the keyword is highlighted for sentiment.

Comment/Post: Text Analytics computes the overall sentiment for the post/comment (Survey responses/Timeline activities). This is not simply an average of the sentiment for the keywords within the post, but the overall general sentiment within the post.

How are suggested keywords computed?

Text Analytics analyzes the keyword mention patterns in the entire text data, to derive a similarity score for the keywords. The keywords with a similarity score beyond a certain limit for any keyword within the topic are shown.

Can I add or remove keywords I do not like?

No, you cannot currently add or remove the keywords derived from the NLP (Natural Language Processing) algorithms.

Can a keyword belong to multiple topics?

Yes, a keyword can belong to any number of topics.

Can I use the Text Analytics for sentiment analytics of non-English text ?

The sentiment analytics in Text Analytics is ALWAYS performed on english text. Yes, you may feed-in a non-English text. But such texts are first translated to English and then submitted for the text processing workflow. In other words, sentiment analytics is always in english. In the drill down pane, the user has a choice to toggle to the original non-English text; however, the original text won’t render the highlighted keywords with the respective sentiment color. That will only be available with analytics on english text.