Skip to content

Web Analytics Source Mappings

Purpose

Source mappings define how canonical Web Analytics metrics are obtained from the GA4 Data API, across the six properties in scope.

The canonical metric definitions describe the organization's meaning of each metric. This document describes how those metrics are represented by GA4.

A GA4 field must not be treated as canonical simply because it exists in the API response, and — as covered in the field matrix — a GA4 field must not be treated as equivalent to an existing canonical metric simply because the names look alike.

Mapping Structure

Each mapping records:

Property Description
Property Which of the six GA4 properties the field was observed on (or "All" where uniform)
Source Object GA4 API surface providing the field (standard report, real-time report, cohort report)
Source Field Field exposed by the GA4 Data API
Canonical Metric Organizational metric to which the field maps
Mapping Type Nature of the semantic relationship
Confidence Confidence in the mapping
Measurement Type Cumulative, rate, or derived average
Time Semantics Bounded period unless otherwise noted
Historical Availability Whether historical values can be retrieved
Notes GA4 specific limitations or interpretation

Mapping Types

Direct

The source field represents substantially the same concept as the canonical metric.

GA4
sessions
        ↓
sales.web_sessions

Semantic

The source field represents a concept that is sufficiently similar to the canonical metric but needs a transformation, not just a rename, before it matches the canonical definition.

GA4
userEngagementDuration   (a sum)
        ↓
sales.web_avg_engagement_time   (an average)

Conditional

The mapping is only valid via a different request shape than the one used for the rest of the domain, or only under defined conditions.

GA4 cohort report
cohortActiveUsers / cohortTotalUsers
        ↓
sales.web_retention_rate

Unresolved

The field has been identified as in scope but no GA4 source currently provides it.

Organic search query
        ↓
(unresolved — requires Search Console, not GA4)

Unresolved fields remain visible here rather than being silently dropped, so the gap stays deliberate rather than accidental.

Current Mappings

Acquisition

Property Source Object Source Field Canonical Metric Type Confidence Notes
All Standard report activeUsers sales.web_users Direct High GA4's default "Users" figure
All Standard report newUsers sales.web_new_users Direct High First-time users only
All Standard report sessions sales.web_sessions Direct High Standard session count
All Standard report sessionSource / sessionMedium dimension on sales.web_sessions Direct High Session-level traffic source
All Standard report firstUserSource / firstUserMedium / firstUserDefaultChannelGroup dimension on sales.web_users Direct High Acquisition-time source, distinct from session-level source

Engagement

Property Source Object Source Field Canonical Metric Type Confidence Notes
All Standard report engagementRate sales.web_engagement_rate Direct High Not comparable to sales.engagement_rate (social) — see field matrix
All Standard report userEngagementDuration sales.web_avg_engagement_time Semantic Medium Divide by sessions to get the per-session average shown in the GA4 UI
All Standard report screenPageViews sales.web_page_views Direct High Combined page and screen views
All Standard report pageTitle / pagePath dimension on sales.web_page_views Direct High Page-level breakdown
All Real-time report audienceName dimension on real-time user counts Conditional Medium Not part of the standard lifecycle reports; real-time only, not queryable historically

Monetization

Property Source Object Source Field Canonical Metric Type Confidence Notes
All Standard report conversions sales.web_key_event_count Direct High API field name unchanged; GA4 UI now labels this "Key events"
All Standard report totalRevenue sales.web_revenue Direct High Expect sparse or zero data for non-ecommerce properties
All Standard report ecommercePurchases context for sales.web_revenue Semantic Medium A transaction count, not itself a revenue figure

Retention

Property Source Object Source Field Canonical Metric Type Confidence Notes
All Cohort report cohortActiveUsers / cohortTotalUsers sales.web_retention_rate Conditional Medium Requires a cohortSpec request, a different shape from every other mapping in this domain

Not Yet Sourced

Fields identified as conceptually in scope for this domain but not currently obtainable from GA4:

Concept Reason unmapped Likely source if pursued
Organic search keyword / query performance GA4 has not exposed search query terms since Google's "(not provided)" change; this data does not exist in the GA4 Data API Google Search Console, via the Search Console Search Analytics API — a distinct source with its own auth flow, property linking, and field matrix, not an extension of this one

This is recorded here deliberately rather than mapped to an approximate GA4 field, since no GA4 field represents this concept even approximately.

Measurement Semantics

A direct mapping does not mean that source values can automatically be aggregated together.

Metric Measurement Type Time Semantics
Users Cumulative (period aggregate) Bounded period
New Users Cumulative (period aggregate) Bounded period
Sessions Cumulative (period aggregate) Bounded period
Engagement Rate Rate Bounded period, non-additive
Average Engagement Time Derived average Bounded period, non-additive
Page Views Cumulative (period aggregate) Bounded period
Key Event Count Cumulative (period aggregate) Bounded period
Revenue Cumulative (period aggregate) Bounded period
Retention Rate Rate, cohort-based Bounded period since acquisition, non-additive

Rate and derived-average metrics should never be summed across properties or time periods; they must be recomputed from the underlying counts whenever a blended figure is needed across more than one property.

Historical Availability

GA4's standard reports can be queried for historical date ranges through the Data API, but this is not unlimited: GA4's underlying event-level data retention is governed by a configurable retention setting (typically 2 or 14 months) at the property level, independent of whether the aggregated report metrics themselves remain queryable for longer. Relying on live queries against GA4 for long-range historical trend reporting is therefore risky in the same way it would be for social platforms with no native history.

GA4 Data API
    ↓
Current period query
    ↓
Scheduled extraction
    ↓
Web analytics metric snapshot
    ↓
Historical trend (warehouse-side, not GA4-side)

As with Social Media, the absence of unlimited native history does not prevent historical reporting — it means history must be accumulated from the point at which scheduled extraction begins, not assumed to be retrievable retroactively from GA4 itself.

Governance

Mappings should be reviewed whenever:

  1. GA4 changes its API or its Data API field names.
  2. GA4 changes a metric's underlying calculation (as happened when "conversions" was relabelled "key events" without a field rename).
  3. A canonical metric changes definition.
  4. A new property (website or app) is introduced.
  5. Mapping confidence changes.
  6. Search Console is linked and keyword data becomes available, at which point it should be evaluated as its own source rather than merged into this one.

The field matrix remains the broader source inventory. This document contains mappings that have been evaluated against the canonical model. The canonical metric definitions remain the authoritative definitions of organizational meaning.