What are GA4's three reporting identity modes?
GA4's reporting identity setting controls how users are identified and counted across sessions and devices. The three modes are: Blended (default — uses User-ID where available, falls back to Google signals, then device ID), Observed (uses only User-ID and device ID — no Google signals modelling), and Device-based (uses only the device's GA4 cookie, closest to UA behaviour). The mode you choose directly affects your user counts, session counts, and conversion attribution. Blended typically shows the lowest user count (de-duplicates cross-device users most aggressively), Device-based shows the highest (counts each device separately).
The difference can be 15–40% depending on how multi-device your audience is. This setting is in Admin → Property Settings → Reporting Identity.
The three modes in detail
Blended (default)
What it uses for identification (priority order):
- User-ID (if implemented and user is logged in)
- Google signals (if user has personalised ads enabled in their Google account)
- Device ID (GA4's first-party cookie,
_ga)
What this means: GA4 uses Google's cross-device graph to attempt to stitch together sessions from the same person across multiple devices, even if they're not logged in to your site. A user who browses on mobile and converts on desktop may be counted as one user (if Google can link the devices via Google signals) rather than two.
Effect on metrics:
- Lower user counts than Device-based (cross-device stitching de-duplicates)
- More accurate cross-device conversion attribution
- Modelled data is incorporated (users without User-ID or Google signals are partially estimated)
When to use: Most businesses — particularly e-commerce and content sites where users browse across multiple devices. The cross-device deduplication generally produces more accurate user counts.
When NOT to use: When your legal team requires no use of Google's cross-device signals (some privacy-strict interpretations), or when you need raw, unmodelled data for compliance reporting.
Observed
What it uses for identification:
- User-ID (if implemented and user is logged in)
- Device ID (GA4's
_gacookie)
What this excludes: Google signals modelling. No cross-device stitching beyond what User-ID provides.
Effect on metrics:
- Higher user counts than Blended (no Google signals deduplication)
- Lower user counts than Device-based (User-ID deduplication still applies)
- No modelling — only what GA4 directly observed
When to use: Properties with strong User-ID implementation (most users are logged in), privacy-focused businesses that don't want Google's cross-device graph involved in attribution, or properties where you've audited your User-ID implementation and trust it.
Device-based
What it uses for identification:
Want to see which hidden implementation gaps are affecting your GA4 data quality?
- Device ID only (GA4's first-party
_gacookie)
What this excludes: User-ID, Google signals, cross-device stitching.
Effect on metrics:
- Highest user counts (each device = a separate user)
- Closest behaviour to Universal Analytics (which was also primarily device-based)
- No cross-device attribution — mobile browse and desktop convert = two separate "users"
When to use:
- When you're comparing to historical UA data and need the most comparable metric
- When User-ID is not implemented and you don't want modelled data in your reports
- When your stakeholders are confused by "lower user counts since switching to GA4" — Device-based gives the most familiar UA-comparable numbers
The User-ID requirement for Observed and Blended
Observed and Blended modes are only meaningfully different from Device-based if you've implemented User-ID. Without User-ID:
- Blended relies entirely on Google signals (which you can't control or verify)
- Observed falls back entirely to device ID (same as Device-based)
Implementing User-ID:
Or via GTM: GA4 Configuration tag → Fields to Set → user_id → value from dataLayer variable containing the logged-in user's internal ID.
User-ID rules:
- Must be a consistent, stable identifier across all user sessions
- Must NOT contain PII (no emails, phone numbers, names)
- Must only be set when the user is authenticated (not set for anonymous visitors)
- The same user must get the same ID every session, on every device
How identity mode affects your numbers: real-world impact
For a property with 100,000 monthly sessions and a moderately multi-device audience:
| Mode | Typical user count | Sessions/user | Notes |
|---|---|---|---|
| Device-based | 65,000 | 1.54 | Each device counted separately |
| Observed (no User-ID) | 65,000 | 1.54 | Same as device-based without User-ID |
| Observed (with User-ID) | 55,000 | 1.82 | Logged-in cross-device users de-duplicated |
| Blended | 52,000 | 1.92 | Additional de-duplication via Google signals |
The difference between Device-based and Blended can be 10,000+ users on a 65,000-user property — a 15% discrepancy. If stakeholders compare GA4 user counts across identity modes without knowing the setting changed, the numbers look incompatible.
Reporting identity and conversion attribution
Identity mode also affects conversion attribution in GA4's attribution reports. In Blended mode:
- A user who views an ad on mobile (no conversion) then converts on desktop may be attributed as a single conversion journey (ad → conversion, cross-device)
- In Device-based, this same journey appears as two separate users — mobile with no conversion, desktop with direct conversion
For businesses with significant mobile browse / desktop convert patterns (common in fashion, considered purchases, B2B), Blended mode produces substantially more accurate attribution. The attribution reports in Blended mode surface cross-device touchpoints that Device-based completely misses.
FAQ: GA4 Reporting Identity: Blended vs Observed vs Device-Based
How close should ga4 reporting identity: blended vs observed vs device-based numbers be before I worry?
What should I validate first when ga4 reporting identity: blended vs observed vs device-based numbers disagree?
When is a discrepancy a tracking bug instead of a reporting difference?
Related guides for GA4 Reporting Identity: Blended vs Observed vs Device-Based
Server-Side GTM Hosting Cost Benchmarks: Cloud Run vs Stape vs Self-Hosted (2026)
Server-side GTM hosting falls into three pricing models in 2026: Google Cloud Run (variable, starts at ~€120/month for 3 minimal servers, scales to €240–€300 for higher traffic, plus optional logging fees), managed providers like Stape (fixed monthly: €20/month for 500k requests…
Server-Side GTM vs Client-Side GTM: A Decision Matrix (2026)
Move to server-side GTM if you (1) need Conversions API integrations with Meta, Google Ads, TikTok, or LinkedIn for offline-conversion match quality (the strongest single justification — typical 9–24% conversion lift)…
Run a GA4 audit before ga4 reporting identity: blended vs observed vs device-based spreads into reporting decisions
Use GA4 Audits to surface implementation gaps, broken signals, and the next fixes to prioritize before the issue becomes harder to trust or explain.