Identify where the path stops
When a cited source will not open, test the interaction and the destination separately. The failure could be in a platform control, an intermediate feed, a redirect or the publisher's page. A visible citation alone does not tell you which part stopped the visit.
Search Engine Journal reported on 27 October that Google disabled a Discover experiment after a source-tap bug. The report does not establish its duration or reach. I could not independently retrieve the underlying X statement. The workflow below is my diagnostic recommendation, not a reproduction of that incident.
Record enough context to repeat the observation
Capture the surface, device, app or browser version, date and exact action taken. Note whether a tap does nothing, opens another feed, redirects or reaches an error. Preserve an approved screenshot or recording if it helps show the boundary.
Avoid collecting personal account details or unrelated feed contents. If you cannot reproduce the issue consistently, record both successful and failed attempts. A useful incident account explains what was observed and what was not tested. It should not turn an intermittent symptom into a platform-wide assertion.
Test the destination through an independent route
If the cited URL is available, open it directly in a normal browser session. Check the final destination and whether the content the citation implies is still present. If direct access fails, inspect the publisher-side response and redirects before rewriting the article.
For an illustrative case, a source card might fail to open its intermediate list while the article itself loads normally. That would justify investigating the platform interaction. A different case might reach an old URL returning an error. Similar visible frustration can therefore require different repair work.
Distinguish incident evidence from performance reporting
Write the observation in a small incident log: affected path, reproduction steps, direct-destination result, likely boundary and unresolved checks. Where the boundary is uncertain, say so. Impression counts do not prove that the source-tap path worked.
Our answer-to-landing guide concerns continuity once a visitor reaches the site. This diagnostic precedes that work. Keep the two investigations separate so a platform interaction problem does not trigger unrelated content changes, or a broken destination get dismissed as an AI interface experiment.
Retest and close the specific incident
Use the same reproduction steps after a change or a reasonable review interval. Record whether the observed path works, remains broken or can no longer be found. An experiment disappearing is different from confirming that every affected interaction was repaired.
JQ's SEO Engine illustrates the broader role of dated operational checks. Add this kind of incident evidence to an existing review before buying another visibility dashboard. The objective is a smaller, better-supported repair decision and a clear account of what the next check needs to resolve.
Link Map
Sources and editorial review
Assigned to the 27 October–2 November 2026 editorial series. This edition was prepared and source-checked on 2 November 2026; assigned dates do not imply prior public deployment. Proposed workflows and illustrative examples are JQ AI SYSTEMS editorial analysis, not completed client experiments.
- Matt G. Southern, SEJ: Discover source-click experiment disabled (27 October 2026; secondary reporting; underlying X statement returned 403).