Context: A four-week technical SEO project on a React single-page app. The first crawl on 29 August covered 26 internal URLs and found 23 errors: 23 pages with no outgoing links and 23 duplicate pages without a canonical URL. A second crawl the same day found more pages and exposed multiple title and meta description tags, pushing errors to 26. Because the site is a JavaScript app, most of the problems were in the code that builds routes, links and the document head, not in the page content. I used Ahrefs as a feedback loop inside development: audit the live site, find the root cause in the app, fix it in code, recrawl and compare, then catch regressions as the site grows.