Theme Explorer standard report

Important: This report is superseded by the TA Topics/Themes Explorer module, which is part of Medallia Alchemy Experience Reporting and can be configured to display more metrics for analyzing theme data, such as Volume Change and Impact on Net Sentiment.

Theme Explorer is a reporting application for investigating the most popular themes that arise from text feedback. The system uses machine learning techniques (rather than humans) to create the themes. Consequently, Theme Explorer presents an unbiased view of the issues customers discuss.

For information about how to analyze non-traditional data sources with Text Analytics, see Import data for Text Analytics.

Note: This topic explains what Theme Explorer is and how companies can use it. See Configure Theme Explorer for guidance on adding Theme Explorer to your system.

Overview video

Watch this video to learn how Theme Explorer can help you uncover valuable insights:

When exploring your themes, you can toggle between Theme Explorer's two views: the Bubbles view and the List view.

You can toggle between the Bubbles view and the List view.

Additionally, a phrase panel displays the phrases associated with the selected theme. When you hover over a bolded word, a tooltip explains how the word relates to the theme:

For example, if the word consultant is associated with the agent theme, this message will appear when you hover over the word consultant in a phrase: This word either directly matches or is closely associated with agent.

The Bubbles view displays the themes that capture the most feedback. Click a bubble to view the themes that are related to it. When you are ready to zoom back out, click outside the bubble, or click one of the navigation breadcrumbs.

The List view is an all-inclusive list of topics. In the List view, Theme Explorer displays the following information for each theme:

  • Volume — Number of records captured by the topic.
    Note: When topics capture multiple phrases from the same survey record, those phrases only count once towards the volume count.
  • Positive — Percentage of positive phrases.
  • Negative — Percentage of negative phrases.
  • Mixed — Percentage of phrases that are both positive and negative.
  • No opinion — Percentage of phrases that do not imply any positive nor negative opinion. For example, the phrase I canceled the service because I moved out conveys both a positive and negative opinion.

Searching for insights

When you search for a word in Theme Explorer, Text Analytics surfaces the themes that are conceptually related to the word. It finds those themes by searching through their concept clusters. When the word you searched matches a word in a concept cluster, the theme associated with the concept cluster appears in the search results.

Note: Theme Explorer does not search the comments. It searches only the themes' concept clusters.

Watch the following video for advice on searching for insights:

Effective searches (what to search for)Ineffective searches (what not to search for)
  • Root versions of nouns, verbs, and adjectives (such as bag instead of bags, or help instead of helping)
  • Multiple words separated by single spaces (such as cashier bag) — Displays the themes that apply to all specified words.

    For example, if you search for cashier bag, Theme Explorer returns results that match both the cashier and bag themes.

  • Phrases (such as Not in stock) — Theme Explorer searches one-word concepts that were generated from comments. It does not search the customer comments themselves.

  • Hyphenated words (such as over-the-counter) — Theme Explorer does not return results for hyphenated words because it does not create hyphenated concepts.

Exporting data

Click Export to download the statistics associated with themes and phrases.

Store is the selected theme, so theme and phrase exports are associated with Store.

Note: This export is not customizable.
The exported data corresponds with the selected theme. For example, in the previous image, Store is the selected theme.

  • The Theme Export (left side) generates a .csv file that includes the selected theme and its related themes. Since Store is the selected theme, a Themes Export would show information about the Store theme, and the themes related to Store (such as Manager Store or Store Refund).

  • The Phrase Export (right side) generates a .csv file that includes all unique phrases from the selected theme. Since Store is the selected theme, a Phrase Export would show information about unique phrases from the Store theme.

The theme discovery process

The theme discovery engine uses machine learning techniques (such as unsupervised clustering) to generate themes. During theme discovery, the system performs the following actions:

  1. Processes text at the phrase level.

    The parser analyzes the raw text (e.g. customer comments) at the phrase level to identify parts-of-speech and the root form of each word. For example, each of the phrases in this comment would be analyzed separately:

    This comment contains two highlighted phrases

  2. Creates concept clusters of words that frequently appear together (words that are conceptually related).

    For example, friendly, staff, salesperson, and rude might constitute a concept cluster because they are used similarly in the text.

    The name of the concept cluster is salesperson because it is the most frequently used word in the concept cluster.

    Although concept clusters are not visible, they form the underlying foundation for the themes the system creates.

    Themes are visible in Theme Explorer, but concept clusters are not visible.

    Parts-of-speech and synonyms do not influence the process of grouping words into concept clusters. Also, the title of the cluster is typically the most frequently-occurring noun people use when discussing the concept.

    In the Phrases panel, Theme Explorer highlights all the words that are part of an underlying concept cluster in bold when you select a theme:

    In this example, store, busy, and wait are highlighted in this comment: "The store is always too busy and the wait too long."

  3. Creates up to 5,000 themes from the concept clusters.

    The top ten themes (the themes that apply to the most records) are displayed in the initial Bubble view. The List view displays the full list of themes.

    When you click a theme, the system reveals new level of related themes. In the following image of the List view in Theme Explorer, Year, Channel, Price, Internet, and Ad are the themes with the greatest number of records that are also related to Program, so these themes appear under the Program theme.

    Several themes in the Program theme

    Theme Explorer displays up to 10 levels of themes. The number of levels varies depending on the number of records available.

  4. Adds emerging themes to the system.

    At the beginning of each month (typically on day 3), Text Analytics analyzes the records from the previous month and adds any significant new themes. You can have up to 6,000 themes.