XM Discover Terms from A to Z

Overview of XM Discover Terms

Here you can find an organized guide of XM Discover features and frequently searched terms and phrases. Use this guide to browse the various features available and to get to the correct support page.

A

Account Administrator: An administrative user role that allows you to manage your XM Discover master account users and assets.

Account Owner: A Qualtrics employee responsible for administering customer accounts. Usually assigned to XM Success Managers and Technical Account Managers.

Alert: A notification that triggers automatically when certain conditions are met. You can get alerts when a certain topic occurs in your data, when a certain metric reaches a predefined threshold, or when a document fails to meet an intelligent scoring rubric’s target.

Associated words: Terms that are grammatically related in text, often displayed in a cloud or table widget. In the sentence “John rented a car,” relationships exist between “John –> car,” “John –> rented,” and “rented –> car.”

Attribute: A data field associated with a record that is not free-text (such as a satisfaction score, loyalty segment, customer name, and so on). Can also be referred to as a field.

B

Badge: An icon that appears in the document explorer to show enrichments relevant to a piece of feedback.

Bar widget: A Studio widget that presents data as a vertical or horizontal bar chart.

Book: A tabbed collection of dashboards that linked in some way, such as being created for the same target audience.

Bottom box: A metric that shows the percentage of respondents who gave the least favorable response in a survey.

C

Call transcript: One of data formats supported by XM Discover. Call transcripts contain conversational data, such as calls. Call transcripts are a type of conversational interaction.

Case: A way to track a variety of tasks and issues that need to be resolved. In a quality management scenario, a case is usually a coaching opportunity triggered from an agent-client interaction along with associated metadata and tags. For interactions that fall below a target score, a case is created automatically. Manual cases can also be created.

Category model: A rules-based hierarchical taxonomy used to organize sentences into topics. Queries can contain keywords, phrases, and attributes. Also called a category tree, classification model, or topic model.

Category reference: Allows you to reuse one category’s rules in another category. Works across models.

Category-derived attribute: An attribute that derives values from classification.

CB Autodetected Language: A premade enrichment that identifies the feedback language if the processing language is not specified. Used for languages not supported by the NLP engine. If processing language is specified, this attribute’s value is “not attempted.”

CB Chapters: A premade enrichment that divides semantically-related segments of a conversation into logical segments at the sentence-level, such as Opening, Need, Verification, Solution Step, and Closing. This helps to address the natural stages and nuances of a conversation and see the proximity of concepts within or outside of chapters.

CB Content Subtype: A premade enrichment that groups “noncontentful” documents into ads, coupons, and article links.

CB Content Type: A premade enrichment that distinguishes “contentful” documents from “noncontentful” ones. “Noncontentful” documents have content that isn’t as useful to customer feedback analysis, such as ads, coupons, and article links.

CB Document Word Count: A premade enrichment that counts the number of words in a document.

CB Effort: A premade enrichment  that measures the level of effort expressed by customers at each touchpoint of their journey. Effort is calculated at a sentence level on a -5 to 5 scale (using only whole, non-zero numbers), and is divided into 3 or 5 feedback bands: Very Hard, Hard, Neutral, Easy, Very Easy.

CB Emotion: A premade enrichment that identifies emotional sentences and specific emotions that appear in feedback and interaction data. Unlike classification models, this enrichment goes beyond a keyword approach and uses machine learning.

CB Emotional Intensity: A premade enrichment that measures the strength of emotions expressed by customers. Emotional Intensity is calculated at a sentence level on a 1 to 5 scale (using only non-zero values). It is divided into 3 bands to group feedback into: Low, Medium, and High.

CB Loyalty Tenure: A premade enrichment that detects mentions of how long a customer has used a service or owned a product (in years). Loyalty Tenure values are derived automatically from sentences with Tenure type.

CB Processed Language: A premade enrichment that identifies the language the feedback has been processed in. If a language is not supported by the natural language processing (NLP) engine, it is marked as “other” and not processed.

CB Reason: A premade enrichment that detects potential reason behind key events in a conversation, such as reason for contact and reason for empathy.

CB Sentence Quartile: A premade enrichment that identifies the part of the verbatim a sentence falls into. Possible values are 1, 2, 3, and 4.

CB Sentence Type: A premade enrichment that identifies sentences containing different kinds of intent specifically relevant to customer experience analytics (such as Churn, Request, or Suggestion).

CB Sentence Word Count: A premade enrichment that counts the number of words in a sentence.

Classification: The processing step in which XM Discover places sentences into a model’s categories based on their rules. This is a type of categorization.

Cloud widget: A Studio widget that presents data as a word cloud.

Co-occurrence: A type of analysis that helps you find topics and other data groupings that frequently occur together in your data. In Studio, you can run co-occurrence analysis in a network widget or in other widgets using the widget as filter feature. Also called overlap.

Coaching opportunity: In quality management, a coaching opportunity is created automatically using cases (when an interaction falls below the target) or manually. These can then be used for manager-to-agent coaching (in which a manager takes poorly scored interactions and coaches agents on them) or QA-to-manager coaching (which lets you review what a manager is coaching agents on).

Connectors: An XM Discover app that streamlines the process of pulling data from various sources to perform analysis across all of them. In addition, outbound connectors insert data enriched by XM Discover back into third-party services.

Contentful: If a customer feedback record expresses sentiment regarding a brand, it is useful for analysis and can reveal customer insights, it is said to be “contentful”.

Noncontentful: If a feedback record includes spam and advertising, and it does not contain information that is useful for analysis and can reveal customer insights, it is said to be “noncontentful”.

Content provider: In Studio, a content provider is the source of data and corresponds to specific instances of Designer.

Custom calendar: Map standard calendar dates to custom definitions of weeks, months, quarters, and years. These custom definitions are used in reports and alerts.

Custom math metric: A metric that enables you to mix and match different metrics and attributes using mathematical expressions.

D

Dark mode: A Studio dashboard theme with dark background to use in low light situations and lessen eye strain if you have a sensitivity to light. Can also be referred to as a dark theme.

Dashboard: Shareable report created in Studio that allows you to present your data as tables and charts (called widgets).

Dashboard personalization: A feature that lets you customize reports to the structure of an organization or market segmentation. In a personalized dashboard, widgets are automatically filtered according to the selected organization hierarchy and the currently logged-in user’s highest position in that hierarchy.

Derived attribute: Any additional data field created from existing attributes. For example, “Age Range” (values such as “<18,” “18-34,” “35-49,” “50+”) can be derived from the separate attribute “Age.”

Designer: An XM Discover app that serves as a platform for evaluating customer feedback. Designer lets you categorize data and perform other tasks.

Detractors: When calculating a satisfaction metric, detractors are a segment of survey respondents who are at risk of churning or posting negative reviews.

Digital interaction: One of data formats supported by XM Discover. Digital interactions are documents that contain conversational data such as chats. Digital interactions must follow Qualtrics conversational format, helping the system understand key details such as speakers, sequences, and conversational events. A similar conversational format is used for call transcripts.

Dimensional lookup: A derived attribute that maps text or numeric attributes to custom values.

Document: Feedback or conversational data that typically represent a single response or interaction with a customer (for example, a survey response, an online review, a social media post, or a transcript of a contact center call). Can also be called feedback or an interaction. When they contain call or chat data, it’s a conversational document.

Document Date: The primary date field associated with a document. This date can be used in XM Discover dashboards, alerts, and so on.

Document explorer: The viewer that launches when you click a data point in a widget. It displays attributes, topics, and verbatims related to each document behind that data point. This interface also displays conversational data such as call transcripts (with optional audio playback) and digital interactions.

Document ID: The unique system ID of the document. Unlike Natural ID, Document ID is generated automatically by XM Discover.

Drilling: The process of clicking a data point in a widget to zoom in on related data. This action can entail a further breakdown by a topic or attribute; it could launch Document Explorer; or it could show the customer comments associated with the documents.

Drivers: Key predictors of a particular outcome in your data. Use drivers to conduct investigations, generate hypotheses, and get criteria ranking suggestions when setting up intelligent scoring rubrics.

E

Embedded widgets: Studio widgets embedded in external tools using code snippets.

Empty verbatim: Customer feedback, especially survey data, that contains documents without text. XM Discover assigns a special NO_VERBATIM_TEXT verbatim type to documents with empty verbatim fields, which you can use to filter these documents. See Records Without Text.

Enrichment: CX data derived by the XM Discover NLU (natural-language understanding) engine. A variety of metadata added by the NLU engine to documents during processing at sentence or document level and stored in the database.

F

Feedback widget: A Studio widget that displays customer comments provided through one or more sources, typically filtered by one or more criteria. This interface is also used to display conversational data such as call transcripts (with optional audio playback) and digital interactions.

Filter: Criteria used to specify which documents should or should not be included in a dashboard or Designer.

Full-screen mode: A full-screen, auto-refreshing mode that lets you view a dashboard or book.

G

Global Other bucket: A collection of all sentences that pass the criteria for the root node of a model but not for any categories within that model. It represents “everything left over” after classification.

Group: A collection of users in Studio with whom you can share your dashboards, books, filters, and metrics.

Grouping: A dimension that calculations and metrics can be aggregated by, such as values of a topic or attribute, or segments of time.

H

Heatmap widget: A Studio widget that presents data as boxes, sized and colored according to the metrics you select. A heatmap is a good way of displaying hierarchical data as nested boxes when there are a lot of categories.

I

Impact rank: A ranking of drivers by their ability to predict a certain outcome. For example, a driver with a rank of 1 has more indicative power than a driver with a rank of 2 or more.

Inbox: A workspace for managing cases related to a variety of tasks and issues. Within the inbox, cases can be collaborated on and tracked to resolution. Also called Qualtrics Social Connect, or Social Connect.

Inbox template: A template that presents case information in a meaningful, actionable way. It defines the inbox topic the case will be routed to, as well as the case’s title, priority, assignee, and the relevant metadata fields.

Individual feedback: One of the data formats supported by XM Discover. This format is used for documents that present data as a single-row, or “flat” object. It may contain several verbatim fields, but usually these come from a single author (for example, when a survey contains more than one open-ended question).

Intelligent score: A score ranging from 0 to 100 that each document or interaction receives automatically based on the scoring criteria, or business logic, defined in a rubric. This score is a net new attribute that is appended to the document with a name that follows this pattern: [Rubric name] Score.

Intelligent scoring: A feature that combines categorization and custom scoring to create flexible rule-based scores to assess a variety of XM interactions.

Interactions: A page where you can view all the customer feedback, calls, and chats included in your projects.

Interactive Voice Response (IVR): Automated audio menu system employed by contact centers to facilitate navigation and resolution of issues without the need to interact with a live agent.

J

Job: A specifically defined flow of data coming in or out of XM Discover via one of the connectors. A job refers to a reusable dataflow setup (including its source, mapping, filtering, cadence and so on), while a run refers to each execution of a job.

K

Key performance indicators (KPIs): Metrics your business uses to measure performance. These vary from company to company. Many XM Discover features allow you to capture and report on KPIs, such as attributes and metrics.

Key Words: XM Discover uses natural language processing to combine all potential variations of words into their normal word forms, called Key Words. These are referred to as “Chiclets” and “Master Entities” in Discover documentation.

L

Label: A text tag you can apply to dashboards and books, filters, and metrics to make them easier to find.

License: A Studio license determines which permissions and data access level a user has.

Licensed viewer: A person who can view embedded widgets outside of Studio and interact with them (which includes drilling, exploring documents, and navigating to Studio dashboards). Requires SSO.

Line widget: A Studio widget that presents data as lines.

Local Other bucket: A collection of all sentences that pass the criteria for a node in a model (other than the root node) but not the criteria for any of its subcategories.

M

Map widget: A Studio widget that layers XM data on top of maps to provide the geographic context of customer interactions and feedback.

Master account: A collection of users in Studio, typically organized around one customer or internal team. The highest level of organization in Studio. Not to be confused with an individual’s user account.

Metadata: Any information that can be stored with a document, either from the source of origin (attributes) or generated by XM Discover as an enrichment (for example, topics, sentiment, effort, or CB enrichments.)

Metric: Any calculation to be performed on data to then be reused across dashboards (like NPS, average score, minimum, and so on).

Metric alert: A type of alert that triggers notifications when a certain calculation or metric reaches a predefined threshold or deviates from a target.

Metric widget: A Studio widget that lets you display important metrics or KPIs.

N

Natural ID: A unique identifier of your choice that you set when you create a job.

Natural Language Processing (NLP): The process of breaking down unstructured text into its related component parts and establishing relationships between those elements.

Natural Language Query (NLQ): The ability to process data queries in everyday human language, making your data easier to search and action.

Network widget: A Studio widget that presents data as a web of connections.

Neutrals: When calculating a satisfaction metric, neutrals are a segment of survey respondents who like a brand but don’t love it. Also called “passives.”

O

Object viewer: A Studio widget that lets you share the underlying structure of a categorization model or an intelligent scoring rubric.

Organization hierarchy: A framework for defining the breakdown of an organization for XM Discover. These definitions can be used to control access to dashboards and make navigating dashboards easier.

Outliers: A root cause exploration tool that lets you quickly discover which words are unique or unusual about any given data point.

P

Parent reporting: A reporting method that uses organization hierarchy to show data for levels above the current user’s node in the organization, starting at the user’s level and going all the way to the root node level. For example, parent reporting lets you compare your location to your region.

Participant: Identifies one side of an interaction in conversational data. XM Discover identifies participant types as representatives or clients using the CB Participant Type attribute and further identifies representatives as humans, chatbots, or IVR (Interactive Voice Response) bots using the CB Kind of Participant attribute.

Peer reporting: A reporting method that uses an organization hierarchy to show data for siblings, or hierarchy nodes that share an immediate parent with the current user’s node. For example, peer reporting lets you compare your location’s performance to other locations in your region.

Percent Parent: A standard metric that is calculated by dividing the value of a particular data point in a report by the value of its parent. For example, it may show the score for the “East Region” for November 2020 in a monthly trend vs. the value for its parent (the total value for all regions for November 2020 in that trend).

Percent Total: A standard metric that is calculated by dividing the value of a particular data point in a widget (for example, the “September 2020” value in a monthly trend) to the overall volume of the widget (the total for all months combined in that trend).

Period over period: A type of reporting that compares data over 2 periods: current and historical.

Permission: Defines what a user can do in Studio and other XM Discover applications. Can be assigned to users and groups.

Pie widget: A Studio widget that presents data as a pie chart or a donut chart. It is a useful way to compare items and visualize their relationship to the whole.

Precision: In classification, a measure indicating the level of accuracy of a model or category. It is calculated as a percentage of sentences returned that fit the scope of a model or category as compared to all sentences that were classified in the model or category. For example, if a “Price” category contains 100 sentences, but only 90 of them were actually relevant to price, precision is 90 / 100, or 90%.

Project: A project groups together feedback data, models, attributes, and settings within XM Discover.

Promoters: When calculating a satisfaction metric, promoters are a segment of survey respondents who are highly likely to recommend a brand.

Pulse: An XM Discover mobile app available for Android and iOS devices that gives access to dashboards.

Q

Quality assurance (QA): The standards a customer-facing organization should adhere to during customer interactions. A means of scoring interactions in order to improve interactions with customers in the future.

Qualtrics Contact Center Quality Management (CCQM): An accountability process you can use to improve the behaviors of representatives, with the right tools and processes in place to learn from the recorded conversations and coach on those that fail to meet a particular threshold.

Qualtrics conversational format (QCF): A data format that supports dialogue-specific metadata and unlocks conversational visualization and enrichments in XM Discover.

Quick translate: Option to get a quick translation of feedback in document explorer and the feedback widget. This will generate the translation from whatever the original language is into your choice.

R

Range rollup: A derived attribute that maps numeric or date attributes to custom ranges.

Rebuttal: A mechanism that lets representatives appeal quality management scores they disagree with. A rebuttal adjusts the Intelligent Score that was generated by the scoring engine and enables calibration flows. Rebuttals are performed in the inbox, but are visible in Studio for transparency.

Recycle bin: A temporary storage for deleted objects that lets you restore them in Studio.

Reference line: A horizontal line drawn on a widget intended to represent a reference point or target the other data points can be compared to. You can insert metric and event reference lines.

Representative: A person who represents the company/brand in a contact center scenario.

Request ID:  An ID available in widget stats mode. Mostly used by the product and Support for logging, tracing, and troubleshooting.

Root node: The top-level category in every classification model. If rules are applied to the root node, they work like a filter applied to an entire model.

Rubric: Defines the intelligent scoring criteria for documents. A rubric can contain a set of topics that should either be present or absent, with a relative weight (or automatic failure) assigned to each of them.

Rule lane suggestions: A Designer feature that analyzes the words you type into a box and suggests synonyms, related concepts, and common misspellings.

S

Satisfaction metric: A metric that shows the percentage of satisfied customers based on a question asking to rate a company or product on a numeric scale. The score is calculated by dividing all the positive responses by the total number of responses.

Scatter widget: A Studio widget that presents data as bubbles plotted along a horizontal and vertical axis.

Scorecard: In intelligent scoring, a scorecard is the result of scoring a document. A scorecard shows the document’s overall score plus individual criteria results (passed, failed, or ignored).

Scorecard alert: A type of alert that sends notifications when a document fails to meet an intelligent scoring rubric’s target. Note that you cannot edit these alerts directly; they are updated automatically based on the rubric’s definition.

Security Audit: A Studio page that shows the history of user activity in the account, including sign-ins, dashboard views, dashboard shares, and so on, as well as updates to account settings and actions on objects (creation, updates, deletions). Accessible only to authorized users.

Selector widget: A Studio widget that acts as a filter for other widgets, creating an interactive dashboard experience.

Sentence: In Discover, a sentence is any string of words with a subject and a predicate, regardless of whether it begins with an upper-case letter and ends with a period. A sentence is the basis for categorization, sentiment, effort, and emotional intensity.

Sentiment: A premade enrichment that evaluates how positive or negative an individual comment is. Sentiment is calculated at a sentence level on a -5 to 5 scale, divided into 3 or 5 bands to group feedback into: Very Negative, Negative, Neutral, Positive, Very Positive. It is available across all data sources.

Smart other node: A special category type that automatically excludes data captured by its peers.

Spine graph: A visual summary of an interaction with key metrics and events displayed in customizable vertical columns.

Studio: An XM Discover app used for creating customizable dashboards and sharing them with users across the organization. Dashboards can be configured to display a variety of XM reports, as well as custom content including text, images, and videos.

Studio homepage: The first page you see when you log into Studio. A page where you view and manage all dashboards and books you own or that other users have shared with you. You can also customize what’s displayed here.

System metric: A Studio metric that is provided out-of-the-box without the need for configuration. Standard metric examples include: Volume, Percent Total and Percent Parent.

T

Table widget: A Studio widget that presents data in a table.

Target and variance reporting: A reporting method that uses an organization hierarchy to quantify how each level of the organization performs on its targets, and how these targets vary from actual KPIs.

Target: In intelligent scoring, a target score is a benchmark that determines whether an interaction passed or failed the rubric criteria.

Text filter: A search field you can add to dashboards and widgets. Use it to search data that contains specific keywords.

Theme detection: Analyzes data and extracts prominent themes from it. You can use theme detection to add new or enhance existing categories in your classification models.

Token: A root form of a word that can be used in categorization and reporting.

Top box: A metric that shows the percentage of respondents who gave the most favorable response to a survey.

Topic (classification): A specific theme mentioned in the open-ended part of customer feedback that XM Discover captures using classification models. Can also be called a category or node.

Topic (inbox): In the inbox, a topic is the location where cases are stored. The title for topics can be customized or labeled such as “Coaching Opportunities” or “Calls to Review.”

Topic leaf: Any category found at the lowest level in the hierarchical structure of a model. It usually represents the most granular categorization level. Can also be called child node or child leaf.

V

Value metric: A metric that lets you create reusable numbers you can use with custom math metrics or widgets.

Verbatim: The part of customer feedback that contains comments, notes, or responses to open-ended questions. Also called feedback or interaction.

Verbatim alert: A type of alert that triggers notifications when a certain topic is mentioned in customer feedback.

Voice-to-text (VTT): The process of taking recorded speech and transcribing it into written text on an automated basis.

Volume: The number of documents represented by a widget, accounting for all related filters and widget definitions. Referenced as “N” in product. Can also be called total volume or document volume.

W

Widget: An individual chart, table, image, text block, or video used in Studio dashboards.

Widget as filter: A widget setup that lets dashboard viewers click a value in one widget to filter one or more widgets.

X

XM Discover: A Qualtrics app that includes Studio, Connectors, and Designer. XM Discover uses the world’s most advanced conversational analytics to help you understand what people are saying about your company, wherever they’re saying it.