SEO
A structured walkthrough of what a real technical SEO audit checks, why each piece matters, and how to prioritize what you find.
Technical SEO audits get treated as a checklist exercise — run a crawler, export a spreadsheet, fix everything flagged red. That approach produces a lot of activity and not much improvement, because it skips the step that actually matters: prioritization based on real impact.
Before anything else, confirm search engines can actually find and index your important pages. Check robots.txt for accidental blocks, review the XML sitemap for accuracy, and audit canonical tags for conflicts that split ranking signals across duplicate URLs.
A surprising number of "ranking problems" are actually indexing problems — a page simply isn't in the index at all, which no amount of on-page optimization will fix.
Search engines use internal linking as a signal of importance. A flat, poorly linked site structure tells search engines that everything is equally (un)important. Audit how many clicks it takes to reach your priority pages from the homepage, and whether your most valuable content is actually linked to from elsewhere on the site.
LCP, INP and CLS each fail for different technical reasons — render-blocking resources, heavy JavaScript execution, or unstable layout during load. Diagnose each independently using field data (from real users), not just lab data from a single test run.
Schema markup doesn't guarantee rich results, but it does help search engines understand page content unambiguously. Audit for both presence and validity — broken schema is worse than none, since it can trigger manual review flags.
Duplicate title tags, missing meta descriptions and thin content are lower-severity than the technical issues above, but still worth fixing systematically rather than page-by-page.
Not every flagged issue deserves the same urgency. Rank findings by estimated traffic impact and implementation effort, and fix indexing and architecture issues before cosmetic metadata problems — the former can be blocking rankings entirely, while the latter is incremental.
Automated crawlers (Screaming Frog, Sitebulb and similar) are excellent at surfacing issues but terrible at prioritizing them. Treat their output as a starting list, not a finished audit.