A useful technical SEO audit checklist is ordered by consequence. The first item should be the one most likely to be costing you traffic right now, and the last should be the one you would only get to in a quiet week. Most published checklists are ordered by topic instead, which is why they produce 140 findings and zero fixes.
This one runs in five passes, roughly in the order that problems cost money. Work top to bottom and stop when you run out of time — you will have covered the expensive things first, which is the whole trick.
Pass One: Can Google Crawl and Index the Site
Nothing else matters if this is broken, and it is broken more often than anyone expects.
- robots.txt — read the whole file. Confirm no Disallow rule blocks a directory that contains money pages, and confirm the sitemap directive is present and correct.
- Sitemap reality check — fetch it directly. Does it return 200? Does the URL count roughly match your published page count? Are the URLs canonical, indexable, and 200-returning? Check sitemap.xml, sitemap_index.xml, and sitemap-index.xml rather than assuming a link exists on the homepage.
- Index coverage in Search Console — look at the excluded reasons. “Crawled — currently not indexed” at scale is a quality signal, not a technical one. “Discovered — currently not indexed” is usually a crawl budget or internal linking problem.
- Noindex audit — check rendered HTML and HTTP headers on your top 50 pages. An X-Robots-Tag header set at the server level is invisible in the CMS and catches people out regularly.
- Status codes — count 404s with inbound internal links, redirect chains over one hop, redirect loops, and any 5xx.
- Canonical consistency — self-referencing on canonical pages, pointing correctly on duplicates, and never pointing to a redirecting or noindexed URL.
Fix everything in this pass before moving on. A single misplaced Disallow line has cost more organic revenue than every unoptimized image on the internet combined.
Pass Two: Can Google Render What Users See
Rendering is where modern sites break in ways that traditional crawls miss entirely. Your crawler fetched the HTML and found the content; the question is whether the content exists before JavaScript runs.
Test it directly. Disable JavaScript in your browser and load a category page, a product page, and a blog post. If the main content, the internal links, or the title tag disappear, you have a rendering dependency. Google will usually resolve it eventually, but “eventually” can mean weeks, and other crawlers — including the ones feeding AI answer engines — frequently will not resolve it at all.
Also check that internal navigation uses real anchor tags with href attributes. Links implemented as JavaScript click handlers on div elements are not links as far as a crawler is concerned, and this pattern silently orphans entire sections of large sites.
Pass Three: Site Speed With Real-User Data
Pull field data rather than lab scores, and report by template. Core Web Vitals thresholds worth measuring against: Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1.
Be honest about the weight of this. Core Web Vitals is a genuine but modest ranking input, and a site failing all three while ranking well will not double its traffic by passing them. The stronger business case is conversion, which is easier to fund and easier to prove.
One exception deserves emergency treatment. If server response time averages over two seconds, Googlebot fetches fewer pages per crawl session. On a large site that means new pages take weeks longer to be discovered and changes take weeks longer to register. That is a crawl budget problem wearing a performance costume, and it belongs back in pass one.
Pass Four: Ecommerce Traps
Any technical SEO audit for an ecommerce site needs a dedicated pass, because the platform generates problems that content sites never encounter. These five account for most of it:
- Faceted navigation. Filters generating crawlable URLs multiply combinatorially — color plus size plus price plus sort order becomes thousands of near-identical pages. Decide which facet combinations deserve indexing, canonical the rest to the clean category URL, and block the parameter patterns you never want crawled.
- Out-of-stock and discontinued products. Returning 404 destroys accumulated links. Redirecting everything to the homepage creates soft-404 signals at scale. Keep the page live with clear availability status if the product returns; redirect to the closest relevant category or replacement if it does not.
- Product variants. Every size and color as a separate indexable URL splits ranking signals across a dozen pages that should be one. Canonicalize to the parent product.
- Pagination. Category page five needs a self-referencing canonical and real crawlable links to adjacent pages. Canonicalizing every paginated page to page one hides the products on pages two through twenty from the index.
- Internal search results pages. These should never be indexable. They generate infinite low-quality URLs and are a classic source of index bloat.
Audit ecommerce sites with a full crawl rather than a sample. A 25-page quick scan catches template issues; only a deep crawl running into the hundreds or thousands of pages reveals that a filter combination has generated 6,000 URLs nobody knew existed. SEO Rocket runs both — an instant quick scan of roughly 25 pages and deep full-site crawls verified past 900 pages, with actual titles, URLs, and H1 text attached to each issue rather than a bare count.
Pass Five: Structure, Markup, and Housekeeping
Lower urgency, still worth doing. Check click depth from the homepage and flag any revenue page more than three clicks deep. Count internal links per page — under three inbound links and a page will struggle to rank for anything competitive. Review internal anchor text, because “learn more” wastes the clearest relevance signal you fully control.
On markup: validate Product, Article, Organization, and Breadcrumb schema against Google’s Rich Results Test, and confirm the structured data matches what is visible on the page. Mismatched prices or fabricated review markup risk a manual action, which is a much worse outcome than having no schema at all.
Then the small stuff. HTTPS everywhere with no mixed content. One canonical hostname, with www and non-www resolving to it. Hreflang tags reciprocal and pointing at 200-returning URLs if you run multiple markets. Trailing slash handling consistent site-wide.
Turning the Checklist Into a Report
Findings without evidence get deferred forever. Every line in the report needs the actual URL, the actual offending value, and the count of affected pages. “Duplicate title tags on 34 product pages, all reading ‘Shop Online | Brand’ — examples: [three URLs]” gets fixed this sprint. “Title tag issues detected” gets read and forgotten.
Group findings by who fixes them rather than by issue type. Developers get the crawl, render, redirect, and performance items. Content gets thin pages, duplicate titles, and cannibalization. SEO keeps architecture and internal linking. A report grouped by topic gets forwarded to three people who each skim the wrong sections.
Cap the action list at fifteen items. Everything else goes in an appendix. A technical audit surfacing ninety issues is accurate and useless — the team will do none of them, and you will run the same audit again next year with the same results. Fifteen prioritized items with named owners produces a measurably better site in a quarter.
Rerun It on a Schedule
Technical debt accumulates every time someone ships. A quarterly full audit plus continuous monitoring on the handful of signals with catastrophic failure modes — indexed page count, robots.txt file contents, sitemap URL count, non-200 status codes, canonical consistency — catches regressions in days rather than quarters.
That monitoring layer is the highest-return automation in most SEO programs. A staging robots.txt deployed to production is detectable within minutes by a script watching one file, and undetectable for months by a team relying on annual audits.
