Diagnose self-referrals in a cross-domain journey
Trace a two-domain journey through linker parameters, redirects, consent and acquisition reporting before changing self-referral configuration.
Follow the handoff before changing a referral list
A service site and a separate booking domain can form one customer journey while being two technical origins. If the second domain appears as a referral, inspect the handoff rather than immediately excluding the domain from a report. A referral rule can change labeling without fixing continuity.
Google’s cross-domain setup guidance describes the _gl linker parameter and warns that a destination redirect or a page rejecting arbitrary query parameters can remove it. Inspect the actual clicked link and final destination. [Google Analytics: Set up cross-domain measurement]
A synthetic two-domain trace
Use reserved example domains for documentation and an authorized technical journey for the real test. Do not include names, contact values or account recovery data in a campaign URL. The following is a diagnostic sequence, not a ready-to-deploy tagging configuration.
| Stage | Expected evidence | Question if it fails |
|---|---|---|
| Marketing page on site.example | Correct permitted tag configuration | Is this the intended property and stream? |
| Click to booking.example | Expected linker information when applicable | Did the click path actually use the configured link? |
| Redirect to the booking route | Permitted parameters preserved through every hop | Did a redirect or allowlist discard the linker? |
| Booking page loads | Correct destination configuration and consent state | Is collection allowed and initialized on the destination? |
| Processed report | Acquisition continuity under the chosen scope | Does the report reflect the tested journey after processing? |
A JavaScript button, form handoff, short link and ordinary anchor may take different routes. Test the exact entry path used by customers, including a deep link and mobile navigation. Seeing _gl on one copied URL is not proof that every checkout path is configured.
Keep measurement permission in the trace
Record the consent state on each origin and the intended behavior when permission is unavailable. Cross-domain configuration must not be used to bypass a person’s choice. A denied test path can have less measurement continuity by design, which should remain visible in the diagnostic result.
Do not paste real linker values or session identifiers into public reports. Save redacted structural evidence and the expected parameter presence where sufficient. Use protected technical records only when the operating process requires more detail.
Different properties, stream IDs or incompatible configuration can also make a continuous business journey appear as separate measured activity. Confirm the actual architecture before choosing a fix. The right design depends on which domains belong in one measurement scope and what access is authorized.
Verify a narrow correction
- Capture the current configuration and a redacted failing handoff.
- Find the first hop that loses required navigation information or starts the wrong measurement path.
- Correct that hop or configuration under the approved change scope.
- Replay the same route and a denied-permission route.
- Inspect both technical delivery and the later processed acquisition view.
If the platform does not expose enough evidence to confirm continuity, state that limit. A request returning HTTP 200 confirms a response, not the complete attribution relationship. Likewise, a cleaner source label does not prove that every booking is attributed to the correct original campaign.
I would report configuration, tested delivery and business outcomes separately. The handoff can pass technically while no real booking has occurred. Preserve that distinction when deciding whether the acquisition report is ready for budget decisions.
Sources
AI-assisted writing and an AI-generated editorial cover. Source links and checked dates are listed below; illustrative examples are labeled in the text.
- Google Analytics: Set up cross-domain measurement — checked 2026-10-18