GA4 to Adobe Analytics metric translation reference

What this is

A lookup card that maps Adobe Analytics dimensions, metrics, and session rules to their nearest GA4 equivalents, with each platform’s scope, limit, and persistence behavior stated side by side. Use it when building an Adobe-to-GA4 migration crosswalk or reconciling a dual-tagging period; every value was read from the official page listed under Sources on the date in the header.

Sessions and visits

Adobe closes and splits visits on several hard rules; GA4 has an inactivity timeout only and reports sessions as a statistical estimate. They do not reconcile exactly.

GA4 auto-collects a session_start event when a session begins, generating ga_session_id (timestamp in seconds at session start) and ga_session_number (count of that user’s total sessions).

What starts a new Adobe visit versus what does not (when hits stay under the 30-minute gap):

Users and visitors

A user becomes active via an engaged session, or collection of first_visit / engagement_time_msec (web), first_open / engagement_time_msec (Android), or first_open / user_engagement (iOS). GA4 emphasizes Active Users as its primary metric where Universal Analytics emphasized Total Users, which makes the default GA4 “Users” number lower than a raw visitor count.

Variable crosswalk: eVars, props, events to GA4

Notes:

• Props support pathing and can be enabled as list props (multiple values in one hit). Enabling pathing makes Entry and Exit dimensions available.
• Props appear in real-time reporting; eVars have a roughly 30-minute reporting delay.
• GA4 user properties persist but are not retroactive and have no eVar-style allocation or expiry model. Reproduce eVar persistence at collection time (see Quick recipes).
• Adobe event serialization enables deduplication via eventx.id in XDM, or a serialization text field in the Adobe Analytics extension.
• Prebuilt Adobe events include purchase (Orders), prodView (Product Views), and cart events scOpen / scAdd / scRemove / scView / scCheckout.

GA4 custom definitions and collection limits

GA4’s registration slots are far smaller than Adobe’s variable inventory, so not every eVar, prop, or event can become a registered custom definition.

• Scopes: event-scoped reports on event parameters; user-scoped reports on user properties; item-scoped covers ecommerce product-array data such as color or size. Custom metrics analyze numerical event-parameter values.
• After deleting custom dimensions you must wait 48 hours before adding new ones.

Collection limits (enforced at collection, not by the interface):

Events, key events, and conversions

GA4’s “Is key event” dimension populates the text true when an event was marked as a key event; “Event name” populates with the event’s name (for example purchase).

Engagement and bounce

GA4 bounce is the inverse of engagement, which is directionally opposite to the Adobe/UA single-hit definition.

GA4 excludes first_visit, first_open, and session_start from engaged-session calculations even if they are marked as key events. The CJA default bounce counts sessions with only a single event/hit recorded (see the CJA note below).

Adobe counting metrics: instances, occurrences, page views

Example: an internal search that sets eVar1 on the search page and persists across 5 subsequent pages records 1 Instance versus 6 Occurrences.

Segments and audiences

• Adobe sequential segmentation uses the THEN operator to enforce navigation order.
• GA4 audiences are forward-looking, not retroactive: they accumulate users from creation onward. Exception: Google Ads audiences backfill the last 30 days.
• GA4 audience maximum membership duration is 540 days / 18 months (14 months if “Reset on new activity” is enabled).
• Adobe segments apply retroactively to historical data; GA4 audiences do not.

Attribution and analysis tooling

CJA note: the “no eVars/props/success events”, “no attribution by default”, and single-event bounce behaviors describe Adobe Customer Journey Analytics (CJA), not classic Adobe Analytics. Classic Adobe Analytics behavior can differ; the classic rows above were confirmed from the component pages.

Gotchas

1. Adobe Visits and GA4 Sessions never reconcile exactly. Cause: Adobe splits a visit at 12 hours, 2,500 hits, or 100 hits in the first 100 seconds, while GA4 has no max duration and reports sessions as an HLL++ estimate. Fix: treat parity as directional only; for exact GA4 session counts use the BigQuery export.
2. GA4 “Users” comes in lower than Adobe Unique Visitors. Cause: the default GA4 “Users” is Active Users (engagement-gated), whereas Adobe counts every distinct visitor ID. Fix: compare Adobe Unique Visitors against GA4 Total Users explicitly, and still expect identity/cookie gaps.
3. Migrated eVar values only show on the event where they were sent. Cause: eVars persist with allocation and expiry; GA4 event parameters are hit-scoped with no persistence. Fix: reproduce persistence at collection time by populating the parameter on every relevant event.
4. Bounce rate looks inverted across platforms. Cause: GA4 bounce is 1 − engagement rate, while Adobe/UA bounce counts single-hit visits. Fix: never compare bounce at face value; rebuild the target platform’s definition as a custom segment.
5. Itemized Adobe values collapse into (other) in GA4. Cause: GA4 groups low-frequency values into (other) once a report exceeds its row limit, and treats dimensions over 500 values as high-cardinality. Fix: use the BigQuery export or bucket values to reduce cardinality.
6. Stakeholders cannot find “Conversions” in GA4. Cause: the March 2024 rename shows “Key events” in reports; “Conversions” now means Google Ads actions only. Fix: map Adobe success events to GA4 key events and reserve “conversions” for the Google Ads context.
7. Not every Adobe variable fits GA4’s registration slots. Cause: GA4 standard caps at 50 event-scoped + 25 user-scoped + 10 item-scoped dimensions and 50 custom metrics, well below Adobe’s 250 eVars / 75 props / 1000 events. Fix: register business-critical variables; leave the rest as raw parameters, and batch-plan around the 48-hour re-add wait.
8. eVar/prop values truncate or fail to register in GA4. Cause: GA4 caps parameter values at 100 characters and names at 40 (names over 40 do not report as key events), against Adobe’s 255-byte eVar payloads. Fix: audit long payloads before migration; shorten names to <= 40 chars and split oversized values.

Quick recipes

Exact GA4 session count from the BigQuery export, bypassing the UI’s HLL++ estimate:

-- One session = one distinct (user, ga_session_id) pair.
SELECT COUNT(DISTINCT CONCAT(
  user_pseudo_id, '-',
  CAST((SELECT value.int_value FROM UNNEST(event_params)
        WHERE key = 'ga_session_id') AS STRING)
)) AS sessions
FROM `project.analytics_123456.events_*`
WHERE _TABLE_SUFFIX BETWEEN '20260701' AND '20260731';

Compare Adobe Unique Visitors against GA4 Total Users (not the default Active Users):

-- Total Users = every user who triggered any event in the range.
SELECT COUNT(DISTINCT user_pseudo_id) AS total_users
FROM `project.analytics_123456.events_*`
WHERE _TABLE_SUFFIX BETWEEN '20260701' AND '20260731';

Reproduce eVar last-touch persistence at collection time, because GA4 parameters are hit-scoped:

// Carry the value on every relevant event, not just where it was first set.
let lastCampaign = null;
function trackWithPersisted(name, params = {}) {
  if (params.campaign) lastCampaign = params.campaign; // "set" like an eVar
  gtag('event', name, {
    ...params,
    campaign: params.campaign ?? lastCampaign,
  });
}

Rebuild GA4’s engaged-session rule for cross-platform bounce parity:

-- Engaged if the session runs > 10s OR has 2+ page/screen views.
-- (The key-event branch is omitted here for brevity; add it if used.)
WITH per_session AS (
  SELECT
    CONCAT(user_pseudo_id, CAST((SELECT value.int_value FROM UNNEST(event_params)
      WHERE key = 'ga_session_id') AS STRING)) AS sess,
    COUNTIF(event_name IN ('page_view', 'screen_view')) AS views,
    SUM((SELECT value.int_value FROM UNNEST(event_params)
      WHERE key = 'engagement_time_msec')) AS engaged_ms
  FROM `project.analytics_123456.events_*`
  WHERE _TABLE_SUFFIX BETWEEN '20260701' AND '20260731'
  GROUP BY sess
)
SELECT
  1 - SAFE_DIVIDE(COUNTIF(engaged_ms > 10000 OR views >= 2), COUNT(*))
    AS bounce_rate
FROM per_session;

Clamp migrated Adobe payloads to GA4’s collection limits before they truncate:

// Adobe eVars allow 255 bytes; GA4 values cap at 100 chars, names at 40.
// Names over 40 chars will not report as key events.
const GA4_VALUE_MAX = 100;
const GA4_NAME_MAX = 40;
function toGa4Param(name, value) {
  const key = name.slice(0, GA4_NAME_MAX);
  const val = String(value).slice(0, GA4_VALUE_MAX);
  return { [key]: val };
}

Related

• Hub: GA4 guides
• GA4 limits reference card
• BigQuery SQL cookbook

Sources

Official pages these values were read from (read 2026-07-20):

• Visits (metric) — Adobe Analytics (read 2026-07-20)
• Unique Visitors (metric) — Adobe Analytics (read 2026-07-20)
• Instances (metric) — Adobe Analytics (read 2026-07-20)
• Page Views (metric) — Adobe Analytics (read 2026-07-20)
• eVar (dimension) — Adobe Analytics (read 2026-07-20)
• Prop (dimension) — Adobe Analytics (read 2026-07-20)
• events (variable) — Adobe Analytics Implementation (read 2026-07-20)
• Segmentation overview — Adobe Analytics (read 2026-07-20)
• Processing and architecture differences (GA to AA/CJA) — Adobe (read 2026-07-20)
• Why GA4 and Customer Journey Analytics data differs — Adobe (read 2026-07-20)
• GA4 Sessions (About sessions) — Google (read 2026-07-20)
• GA4 Total users vs. Active users — Google (read 2026-07-20)
• GA4 Key events (About key events) — Google (read 2026-07-20)
• GA4 Conversions vs. key events — Google (read 2026-07-20)
• GA4 Custom dimensions and metrics — Google (read 2026-07-20)
• GA4 Engaged sessions, engagement rate, bounce rate — Google (read 2026-07-20)
• GA4 Audiences — Google (read 2026-07-20)
• GA4 Explorations (Introduction) — Google (read 2026-07-20)
• GA4 High-cardinality dimensions and the (other) row — Google (read 2026-07-20)
• UA to GA4 Comparing metrics: sessions and users — Google (read 2026-07-20)
• GA4 Events / collection limits — Google (read 2026-07-20)
• GA4 Analytics dimensions and metrics — Google (read 2026-07-20)

Scope notes: the compare-data and processing-differences pages describe GA4 versus Customer Journey Analytics (CJA), so the CJA-specific rows are flagged as such rather than attributed to classic Adobe Analytics. GA4 standard-report table row limits and Adobe unique-value processing ceilings are left unstated where the fetched pages did not give a hard number.