Semrush Site Audit: Setup, Priorities and Retesting

Define the website and the questions to answer

Choose a site you own or are authorized to audit. Record the preferred hostname, important sections, approximate size and intended outcome. For a store, start with the product pages, categories, buying guides and their links rather than treating every authenticated utility endpoint as a public landing page.

Check the account’s available crawl budget and project permission before starting. If the quota covers only part of the site, label the result as a sample. A tool stopping at its page limit does not mean the remaining pages are error-free.

Check crawl settings before launching

The official configuration guide describes settings such as scope, page limit, crawler choice and source. Match the scope to the question: a whole domain, a subdomain or a particular section. Keep a record of settings so the next run is comparable.

Avoid sending the crawler into checkout, order records or other private areas. When a crawl fails, inspect the reported response and provider troubleshooting guidance before changing anything. A blocked research crawler and a blocked verified Googlebot are different tests; neither should be inferred from a score alone.

Turn a finding into an action, not just a count

Prioritize problems that prevent the intended page from being reached or understood, then improve the experience and secondary quality issues. The following is an editorial triage worksheet; investigate each case before applying a bulk change.

Example: investigate a canonical mismatch

Suppose a product declares www while Google’s stored result selects an older non-www URL. Check the live non-www response, the final www page, internal links and sitemap. If the old host already redirects permanently and every current signal agrees, a historical Google-selected canonical may need recrawl rather than another host migration.

Keep the evidence with dates. Do not change a stable preferred hostname repeatedly in response to an old inspection record. A syntactically valid canonical is useful, but Google decides whether pages belong in the same duplicate group.

Verify the repair in the real browser and search tools

After a change, test the actual URL and affected navigation or purchase flow. Confirm headings, canonical, robots directives and visible content. For a shared template change, check other page types as well so a local improvement does not break the catalog or cart.

Re-crawl the affected section using comparable settings. Use Search Console URL Inspection for Google’s stored indexing/canonical state and its live test where available. Lab speed tests and real-user Core Web Vitals measure different things; record which evidence supports your claim.

Keep an issue log that leads to ownership and retests

Use columns for affected URL, issue, evidence, priority, owner, deployed date and retest result. Mark unresolved or inaccessible checks clearly. Track improvements over comparable periods rather than repeatedly running a full audit without acting on it.

Use traffic evidence to prioritize important pages and map relevant query intent before rewriting content. If you need paid access for a larger crawl, confirm the exact package’s Site Audit and export allowances rather than assuming the provider’s marketing tier applies.

Frequently Asked Questions