Tag pool settings

Reporting > Text Processing > Topic & Theme Tagging > Settings

The Settings screen defines the main specifications of your Medallia Text Analytics environment, including options related to sentiment analysis and comment processing. For information about configuring sentiment, see Configuring sentiment.

Preparing the system for comment processing

  1. Navigate to the Reporting > Text Processing > Topic & Theme Tagging > Settings screen.

  2. Turn on the Automatic Tagging Enabled property.
    Important: Both the Automatic Tagging Enabled Topic & Theme Tagging setting and the Automatic Processing Enabled Global Text Analytics setting must be enabled for Text Analytics to process any data.
    Warning: Disabling either Automatic Processing or Automatic Tagging causes topic alerts not to be triggered, even when a survey meets all alert conditions.

  3. In the TextAnalyticsVersion property, verify that the appropriate option is selected:

    Text analytics version

    Description

    TOPICS_ONLY

    The version of Text Analytics without sentiment.

    SENTIMENT_SUPPORT

    The version of Text Analytics with sentiment.

  4. (Optional) If you want to limit Text Analytics to work with comments of a certain length, indicate the number of characters you want to work with in the MaxCommentLengthToProcess property.
    Note: The default value is set to 16,384 characters, so generally this field is left blank.

  5. (Optional) If you want Text Analytics to report on comments from a particular date onwards, add a value to the ReportingStartDate property.

    • Use the format YYYY-MM-DD hh:mm:ss.

    • Leave this field blank if you want Text Analytics to report on all comments that are being parsed.

  6. Turn on the Comment Field Translation Dropdown Enabled property if you are running machine translations and building topics that will pull from these translated comments. Checking this box means that a dropdown will appear in reporting that allows you to filter for native, translated, or both types of comments.

  7. Provide a value for the MinimumImpactLocalSampleSize property to specify the minimum number of comments required to calculate an impact score. If left black, this value defaults to 10.

  8. Provide a value for the MinimumImpactGlobalSampleSize property. If left blank, this value defaults to 100.

  9. (Optional) Turn on the CalculateImpactForLowSample property if you want reports to display impact scores when the volume does not meet the value specified for the MinimumImpactLocalSampleSize property. If this box is checked, impact scores calculated using a low number of samples are highlighted in grey.

  10. Add a date in the AutoTaggingStartDate property that tells the system to process comments from a particular date onward. Use the format YYYY-MM-DD hh:mm:ss

  11. Click Save.

Properties

Deprecated manual mode
Text Analytics uses a hybrid method of automatic topic tagging supplemented with manual tagging. This property enables a legacy mode of manual tagging, and should not be used.
Automatic Tagging Enabled
When on, comments are tagged with topics automatically. For information about how this works, see Text Analytics processing.
Topic Alert Blocking Enabled
By default, when a company has an active topic alert, incoming surveys are processed by Text Analytics before the topic alert (or any alert with a lower priority than the topic alert) is evaluated. Text Analytics text processing provides Experience Cloud with the information required to evaluate the active topic alert, but the processing time required can delay survey data from appearing in reporting for up to 30 minutes. When on, survey records are blocked from appearing in reporting until Experience Cloud evaluates them against any active alerts (including active topic alerts). When off, surveys are allowed into reporting without evaluating them for topic alerts.
Important: This property only applies to companies with active topic alerts, and it should always be turned on. If you turn this property off and then turn it back on, reprocess the surveys completed during the disabled time range in order to trigger any missed topic alerts.
TextAnalyticsVersion
Version of Text Analytics to use.
Warning: If you switch to a different version after launching Text Analytics, there is high chance that related reports will show invalid data. Notify customers or disable capabilities appropriately.

Select from:

  • TOPICS_ONLY — The version of Text Analytics without sentiment.
  • SENTIMENT_SUPPORT — The version of Text Analytics with sentiment.
    Important: After you change the TextAnalyticsVersion property to SENTIMENT_SUPPORT, you cannot undo this action and go back to TOPICS_ONLY.
MaxCommentLengthToProcess
The maximum length of a comment. When you leave this field blank, the default value of 16,384 characters applies. Text Analytics ignores comments that have a greater number of characters that the specified amount.
Threshold for topic activation on partial publish
Period of time after which topics are activated on a topic publish flow, making those topics available in reporting. This enables activation of topics before the entire publishing job is finished.
ReportingStartDate
Surveys completed after the specified start time are loaded into reporting accessible areas (Such as Slug). Text Analytics reporting is available only for surveys completed after this time. When this field is blank, there are no time period restrictions on the survey data that appears in Text Analytics reports. These requirements apply:

  • You must run a backfill after changing this value. Until the backfill completes, comment highlighting might not work correctly.

  • Reports will include data from after the specified day and time (not at the day and time).

  • Use this format for the date: yyyy-MM-dd HH-mm-ss.

Comment Field Translation Dropdown Enabled
Adds a Translation dropdown to the Comment Field filter. Turn on this property if you are running machine translations and building topics that pull from these translated comments.
MinimumAutoTaggingWeight
Warning: This property is deprecated; do not use.
MinimumImpactGlobalSampleSize
The minimum number of tagged records required for the system to display the impact score; defaults to 100 if not specified. This value applies to:

  • Impact scores shown in reports

  • AA2 reports that use the company-impact calculation
  • Usages of the deprecated MII calculation dropdown option
Note: In certain contexts, records are organized behind-the-scenes into an outer (global) set of records and an inner (local) set of records. When this happens, the MinimumImpactLocalSampleSize value applies to the local records, and the MinimumImpactGlobalSampleSize value applies to the global records.
MinimumImpactLocalSampleSize
The minimum number of tagged records required for the system to display the impact score for a set of child records; defaults to 10 if not specified. This value applies to:

  • Impact scores shown in reports

  • AA2 reports that use the company-impact calculation
  • Usages of the deprecated MII calculation dropdown option
Note: In certain contexts, records are organized behind-the-scenes into an outer (global) set of records and an inner (local) set of records. When this happens, the MinimumImpactLocalSampleSize value applies to the local records, and the MinimumImpactGlobalSampleSize value applies to the global records.
CalculateImpactForLowSample
When on, Text Analytics reports display the impact score when the number of tagged records is below the minimums defined by the MinimumImpactGlobalSampleSize and MinimumImpactLocalSampleSize properties. When off, users see a dash (-) instead of the impact score when there are not enough samples.
WordGroups
Warning: This property is for internal use only. Define word groups from the Topics screen instead.
Word groups to use during topic building. In a topic rule, you can reference the name of the word group, and it will expand to include all words in the group. For example, you can group the words bee, honeybee, and bumblebee in the group bee using this syntax: {#bee: bee, honeybee, bumblebee}.

To define a word group, the syntax is {#groupname: word1,word2}. Each group belongs on its own line. Use a new line to separate word groups. You can only use alphanumeric characters or '-', '_', or '.' when defining word groups.

Topic Segments
Warning: This property is for internal use only. Define topic segments from the Topics screen instead.
Topic segments to use during topic building. To define topic segments, the syntax is "segment name;[comment field key, ...];{survey enumerated field key:[<seq number>, ...]}".
Competitors
Warning: This property is deprecated. Do not use.
CustomWords
Warning: This property is deprecated. Do not use. If you need to add customer-specific words, add a user feature instead.
MemberPositiveSurveyFields
Q-fields for questions that inquire about something positive. Move any positively-aligned Q-fields (such as Tell us how we exceeded your expectations) into the box on the right. When analyzing comments from these fields, Text Analytics takes the positive context into account.
MemberNegativeSurveyFields
Q-fields for questions that inquire about something negative. Move any negatively-aligned Q-fields (such as Tell us what we could have done better) into the box on the right. When analyzing comments from these fields, Text Analytics takes the negative context into account.
AutoTaggingStartDate
In systems with automatic tagging, only text from surveys completed after this time are analyzed and tagged by Text Analytics. When this field is empty, there are no time period restrictions on automatic tagging. These requirements apply:

  • Automatic tagging will begin after the specified day and time (not at the day and time).

  • Use this format for the timestamp: yyyy-MM-dd HH-mm-ss.

Use language detection
Important: This setting has been deprecated and removed. Depending on feature enablement, some instances may still see the option to disable language detection. For more information, contact your Medallia expert.

As of the 2026 H2 2 release, language detection is enabled for all customers as a fallback mechanism.

For more information, see Source language and language detection and Configuring language processing.
Use spellchecking
Warning: This property is deprecated. Do not use.
Enable MATCH processing
Important: This property applies only to the legacy version of themes that predates Themes with GenAI.
When on, the concept discovery engine is eligible to generate automated topics. After you enable this option for the first time:

  • The Topic Clusterings and Concept Clusters screens appear after you save and refresh the page.

  • The system generates the required tokens for concept discovery when you trigger text processing (parsing).

See Configure Theme Explorer for more information.

Supported languages for language detection

Text Analytics supports the following languages for language detection:

  • Afrikaans
  • Albanian
  • Arabic
  • Bengali
  • Bulgarian
  • Chinese (Simplified)
  • Chinese (Traditional)
  • Croatian
  • Czech
  • Danish
  • Dutch
  • English
  • Finnish
  • French
  • German
  • Greek
  • Gujarati
  • Hebrew
  • Hindi
  • Hungarian
  • Indonesian
  • Italian
  • Japanese
  • Kannada
  • Korean
  • Macedonian
  • Malayalam
  • Marathi
  • Nepali
  • Norwegian
  • Persian
  • Polish
  • Portuguese
  • Punjabi
  • Romanian
  • Russian
  • Spanish
  • Slovak
  • Somali
  • Swahili
  • Swedish
  • Tagalog
  • Tamil
  • Telugu
  • Thai
  • Turkish
  • Ukrainian
  • Urdu
  • Vietnamese