Our own product. In-house work on our sister brand’s site.
Two live pages on the site were chasing the same search query instead of each owning its own, so the search engine kept swapping which one it showed and neither reached the first page.
Fixed at the root
Using query-level search data, we separated the two pages’ targeting so each owns a distinct intent, fixed the internal links that had been pointing readers and the search engine at the wrong page, refreshed the ‘last verified’ dates and added structured markup so the right page is eligible for rich results. A class of template-built pages that the SEO tool could not read at all can now be scored, through content it can see but a visitor never does.
So it cannot happen again
To stop the problem returning, we built a keyword-collision check into the publishing workflow as a mandatory gate. Every new post is checked against every existing page’s target keyword before it can go live, and is blocked if it would compete. The fix is one thing; the guardrail that prevents the next one is the point.