Skip to content

GA4 cross-domain tracking from a blog to a separate app

Trace a search visit from your blog to an app on another domain. Check GA4 linker parameters, redirects, and the completed action before using attribution.

By Crawlito · 6 min read

In short

  • Two independent domains need an intentional measurement setup to preserve the same GA4 visit.
  • The linker parameter must reach the destination while it is still valid.
  • A completed application action provides a different signal from a click on the blog's call to action.
On this page

If readers arrive from search on your blog but the app credits signups to your own website, test the actual link between the two domains. For sites you administer, configure GA4 cross-domain measurement on the same web data stream and check that the linker parameter reaches the tagged destination through every redirect. Then complete a test action and inspect its event. Google Analytics cross-domain measurement instructions

Draw the visitor's path using actual hostnames

Copy a live article URL and the destination of its call to action. Follow the journey in a browser and record each hostname, including any routing service, authentication host, and final app. A marketing site on example.com handing off to example-app.net crosses a domain boundary. A marketing site and app on sibling subdomains of the same root can present a different cookie configuration problem. Record the exact public arrangement before applying a solution intended for separate domains.

Identify which hosts your organization can tag and configure, and where the relevant action actually completes. A reader can click “Test my page,” open the app, and leave before submitting a URL. Record those steps separately. A marketing team investigating useful traffic needs the successful check or qualified inquiry; a click is evidence that the link was used, even when the next step failed.

Choose a test path that stays within your own product and respects the visitor's analytics choices. If a third-party identity provider sits between the two owned sites, observe its role without sending account details to analytics or copying personal callback URLs into reports. A diagram with hostnames, path categories, and redirect positions is usually enough for the developer to reproduce a broken handoff.

Check the tag configuration and the actual click

Under the Google tag settings for the relevant GA4 web stream, configure the domains involved using Google's documented cross-domain settings. Confirm that the participating pages load the expected tag ID from the same stream. A marketing site running a production property and an app accidentally shipping a staging measurement ID will send observations to different destinations even if the address bar looks fine.

Start from the real article as a new visitor under a test consent state that permits measurement. Open the browser Network panel and enable preservation of the request log. Click the actual call to action. Google's linker uses a _gl URL parameter to pass identifiers for measurement between configured destinations. Check the first outgoing navigation request and then the final landing request. Copying a destination into a fresh tab does not exercise the click logic you intend to test.

A sticky mobile CTA and a desktop text link can share a destination while using different navigation code. Repeat the path from each if customers use both. A scripted button that replaces the navigation, a link shortener that drops parameters, or a destination that normalizes unknown query keys can remove the linker value. Google's instructions explicitly identify redirects and query filtering as troubleshooting points.

Find the hop that loses the measurement

What you see in the network traceWhere to investigate next
No linker on the outgoing cross-domain clickDomain matching, tag ID, consent state, and click handling
Linker present at the first hop and missing after a redirectRedirect rule and forwarded query parameters
Linker reaches the app but events appear in another propertyApp tag configuration and environment
Page events arrive but the business action never appearsEvent trigger, success callback, and measurement definition

Write down just the presence or absence of the parameter on each hop. The actual linker value contains measurement information and is unnecessary in a general ticket. Check the time as well. A copied decorated URL opened much later is a poor test of a parameter intended for an immediate navigation. Repeat with a new click after every change.

A referral exclusion can alter how a referring site is classified. It cannot recreate missing evidence that two visits were one journey. If the problem began after adding a new redirector, fix the point where the parameter disappears and retest. Likewise, tagging your own CTA with fresh campaign parameters can change attribution and make the original acquisition source harder to interpret. The guide to AI citations and referral traffic helps interpret the cases in which an incoming source is observable at all.

Verify a completed app action and its meaning

Use a harmless test input and let the application reach the success state a real user would recognize. Inspect the events and their timestamps in a debugging view available to your team. Compare the initial article view, app landing, submitted action, and completion. A submit event can fire on an invalid form, a network error, or a repeated click. A completion event tied to the actual backend result gives analysts a closer representation of the product outcome.

Run a cancellation path, an invalid submission, and a refresh of the success page. Check which ones increment the completion measure. A repeated render of an already completed report might be useful as a view event, but counting it as a new completed check inflates the result. The conversion measurement guide describes how to define a defensible numerator and denominator before deciding which articles to expand.

Document one technical acceptance test and one reporting comparison. The first uses a newly generated link, the expected tag, and the observed success event. The second examines comparable reporting periods after the change. Reporting identity, attribution rules, consent, and traffic mix can cause aggregate totals to differ even after a single test journey works. An observed improvement in attribution completeness does not establish that the articles suddenly attracted more users.

Trace a redirect that strips unfamiliar query keys

Consider a blog link to an old app path that redirects to the current onboarding page. Your browser's first request may contain a linker while the redirect destination contains only the app's chosen return path. Look at the Location header for that hop in the preserved network log. If the linker disappears there, the redirect owner can reproduce the issue without changing the article or analytics configuration. The HTTP redirect guide explains how to capture the response and its destination.

Test the repaired handoff with a fresh click and a normal target URL containing one legitimate product selector. The selector should still choose the correct app flow after navigation. Keeping needed parameters does not require reflecting arbitrary destinations or sending identifiers to unrelated hosts. Document the intended transfer between the two owned domains and check it again after the next sign-in or routing release.

Frequently asked questions

Does a missing referrer prove the visitor came from an AI answer?

No. Multiple sources can lose referral information. Preserve unknowns in the analysis and use observable session evidence when assigning a source.

Use the supported automatic configuration and test a real click. Hardcoding a transient value in an article can give readers stale or misleading measurement data.

Can I test cross-domain analytics with a crawler alone?

You need a browser journey and analytics observations to test link decoration and the app event. A simple fetch can establish whether the public article and its CTA exist in the delivered HTML.

Check the article end of the journey

Use the free page check to inspect the source article and its link. Then trace that exact link in a browser and verify the completed app action using your own analytics setup.

Filed under Strategy and audits. Topics: analytics, measurement, troubleshooting