After the SEO audit: what to fix first

A technical crawl of a typical e-commerce site — two thousand URLs, ten years of history — returns around forty issue types: duplicate titles, missing descriptions, redirect chains, hreflang gaps, five hundred images without alt attributes. The report looks alarming. Most of it does not matter this quarter, and treating it as a flat checklist is how audits turn into busywork.

What actually moves ranking

Triage by mechanism, not by count. First: anything that stops pages from being indexed at all — endless redirects, broken canonical logic, pages orphaned from internal linking, hreflang clusters with no self-reference. A page the crawler cannot reach earns nothing, however good its content. In our audits these are rarely more than a tenth of the flagged items, and they carry most of the impact.

Second: duplication that splits signals. Eighty-odd duplicate titles across category filters means the site competes with itself. Fixing the template that generates them is one change, not eighty.

Third — and only third — the long tail: alt attributes, short descriptions, uppercase URLs. Real issues, worth scheduling, not worth blocking a release for.

The uncomfortable finding

The most common expensive problem in our audits is not on any checklist: server response time. When half the crawl logs slow responses, no amount of title polishing compensates, and the fix belongs to hosting and caching, not to content. An audit that does not say this out loud is selling comfort, not analysis.

The test for any audit item is a question: what measurable behaviour changes if we fix this? If the answer is silence, it goes to the bottom of the list.