Overview
Records are the structured data units that Data Streams emits for every tracked interaction. Each record type captures one facet of the visitor's session: the page or event they triggered, the entities involved (product, user, campaign), and the environmental context (device, browser, location).
Records are delivered via the Root Stream, Custom Streams, the current Raw Data Export, and — for legacy field sets — the Classic Raw Data Export. This section documents the contents of each record type, the schema paths used on the record, the corresponding stream and export field names, and which channels each field is available in.
Record Types
Activity Records
Capture user-triggered actions in real time.
Pages — page view events with URL, title, and page-level metadata
Events — generic interaction events (clicks, custom actions)
Form — form interactions (submit, focus, field-level events)
Tracking Event — server-side / pixel-tracked events
Aggregation Records
Roll up multiple activity records over time or scope.
Sessions — visitor session aggregates (start, end, page count, duration)
Orders — transaction records with order-line items, totals, status
Customer Journey — touchpoint sequence across sessions
Entity Records
Describe the business objects referenced by activity records.
Product — product catalog data (SKU, name, category, price)
Users — visitor / customer identification and CRM attributes
Campaigns — campaign attribution (source, medium, campaign ID)
Media — media-asset interactions (video, audio, file)
Environment Records
Provide technical context for the visitor's device and connection.
Visitor Environment — device, browser, OS, platform, location, connection, visitor capabilities
Common Conventions
Every Record Contents subpage follows the same layout and uses the same column conventions. Read this section once; the subpages assume you know it.
Page Layout
Each record type has two parallel sections:
Field Definitions — semantic description of each field: name, schema path on the record, description, format, and notes. This is the canonical reference for what each field means.
Stream and Export Mapping — the same fields, mapped to their delivery channels. This is the reference for where each field is delivered.
The Name / Object column is shared between both tables and acts as the cross-reference anchor.
Schema Paths vs. Stream Field Names
Records on the wire use dot notation (hardware.deviceClass, browser.fullName, location.countryCode). The Custom Stream and Raw Data Export flatten the same paths to underscore notation (hardware_deviceClass, browser_fullName, location_countryCode). Both refer to the same data.
Mapping Table Columns
Column | What it means |
|---|---|
Name / Object | Human-readable field name, same as in Field Definitions. |
Available in Root Stream | Tick = the field flows through the Root Stream as-is. |
Field in Custom Stream | Underscore-notation field name in the Custom Stream output. |
Available in Custom Stream | Tick = field is selectable in Custom Stream config. |
File in the Raw Data Export | Which RDE file the field appears in (e.g. |
Field in the Raw Data Export | Field name in the current RDE. |
Available in Classic Raw Data Export | Tick = field is also exported in the Classic RDE. |
Deprecated Fields
Note
Some record types carry a legacy field set that is only available in the Classic Raw Data Export. These appear in a separate, callout-marked table after the modern field set. They are kept for backwards compatibility but should not be used for new integrations — use the modern fields above them instead.
Two record types currently carry deprecated subsets:
Visitor Environment — legacy
device.*fields and the entire Operating System (os.*) field set