MB-260 key terms

A single record that merges data about one person or company from multiple source tables, built by matching and consolidating rows so downstream segments and measures use one consistent view.

A configured connection to an external system, such as a CSV file, Azure Data Lake, or Microsoft Dataverse, from which tables are imported into Customer Insights.

The process of bringing raw records from configured sources into the Customer Insights data lake so they can be transformed, unified, and analyzed.

A named collection of rows and attributes imported from a source. Tables are the input to unification, enrichment, and segment building.

A single field or column within a table, such as email address or total spend. Attributes carry the values used for matching, filtering, and segmentation.

The attribute, or combination of attributes, that uniquely identifies each row in a source table. Unification requires a reliable primary key.

A label applied to an attribute that tells the system what the data represents, such as email, phone, city, or date, enabling correct matching and formatting.

The storage format of an attribute, such as string, integer, decimal, boolean, or date-time. Correct types are required for measures and segment conditions.

An automated review of imported tables that reports row counts, null percentages, distinct value counts, and value distributions so data quality issues are found early.

A summary of issues found in a table, including missing values, duplicate keys, and invalid formats, used to decide whether data is ready for unification.

The configured frequency at which sources are re-imported and downstream unification, segments, and measures are recalculated, keeping profiles current.

A refresh mode that imports only new or changed rows since the last run rather than reloading the entire table, reducing processing time and cost.

A source-side mechanism that records which rows were inserted, updated, or deleted so incremental refreshes can retrieve only modified data.

The managed storage layer where ingested tables and unified profile data are held. It underpins unification, enrichment, segments, and exports.

An isolated Customer Insights instance with its own data, configuration, and permissions, typically aligned to development, test, or production stages.

A set of permissions assigned to a person or group that controls whether they can view, edit, administer, or export data in the environment.

A recorded permission from a customer allowing their data to be used for specified purposes. Consent settings can restrict which profiles enter segments or exports.

The policies and controls that determine how customer data is collected, used, retained, and shared, including consent, privacy, and access restrictions.

The degree to which source data is complete, accurate, consistent, and timely. Poor quality leads to failed matches and unreliable segments or measures.

The process of detecting and removing or merging duplicate rows or records so each real-world customer appears once in the unified output.

Configuration that defines which attributes to compare and how closely they must agree for records to be treated as the same customer during unification.

A numeric confidence value produced by comparing attribute values during matching. Higher scores indicate a stronger likelihood that two records describe the same person.

A cutoff value applied to match scores that determines whether candidate records are automatically merged, flagged for review, or kept separate.

The step that decides which value from competing source records wins for each attribute in the unified profile, using rules such as most recent or most frequent.

The stable key assigned to a unified profile so segments, measures, and exports can reference the same customer consistently across the system.

The end-to-end process of importing tables, matching records across sources, consolidating attribute values, and producing unified customer profiles.

Adding extra attributes to profiles from external providers or internal sources, such as demographics or brand affinity, to improve segmentation and personalization.

A configured connection to a data provider whose attributes are appended to unified profiles, usually matched on identifiers such as email or postal address.

A defined group of unified profiles that meet specified attribute and behavioral conditions, used for analysis, activation, and targeted marketing.

The logical conditions, such as attribute comparisons and set operations, that determine which profiles qualify for membership in a segment.

The current set of profiles that satisfy a segment's rules after the latest evaluation, refreshed whenever underlying data changes.

A calculated business metric defined over customer data, such as lifetime value or purchase frequency, that can be analyzed and used in segmentation.

The configuration that specifies a measure's name, output attribute, entity, calculation type, and filters, determining what value is computed for each customer.

A machine-learning model trained on historical customer data to score profiles, for example estimating churn risk or likelihood to purchase.

The numeric output a trained model assigns to each profile, expressing estimated likelihood or risk and usable as a segment condition.

A configured target, such as a file, Azure storage, or Dataverse, to which segments or profile data are written for activation in other systems.

The scheduled or on-demand transfer of unified profiles, segment membership, or measures to an external destination for use in campaigns or analytics.