Google now runs on a rendering and evaluation pipeline that treats speed, crawl efficiency, and machine-readability as first-class ranking inputs, not nice-to-haves. A site can have flawless keyword targeting and backlinks and still get outranked because Googlebot cannot efficiently crawl it, key pages are excluded from the index, or Core Web Vitals fail on real user devices. That is the gap a technical SEO audit exists to close.
The stakes are higher in 2026 than they were even two years ago. Search Console data now shows a measurable CTR gap between pages in the top LCP quartile and pages that fail Core Web Vitals thresholds, INP has fully replaced FID as the interactivity metric, and a growing share of discovery traffic comes from AI crawlers (GPTBot, ClaudeBot, PerplexityBot) that behave nothing like Googlebot, most of them do not execute JavaScript at all. An audit that only checks “does it rank” and ignores “can machines actually read it” is incomplete.
This checklist walks through every major area a thorough technical audit should cover, in the order most practitioners actually work through them: crawlability first, then architecture, then performance, then the structured-data and duplication issues that quietly cap growth. Budget at least two to three days for a full audit on a site of any real size, and plan to repeat the core checks every month between full audits: regressions creep in fast after a CMS update, a plugin change, or a migration.
Crawlability and Indexability
If Google can’t crawl a page, nothing else on this list matters. Start here every time.
Robots.txt and Crawl Directives
Pull the live robots.txt file and read it line by line, don’t trust what you think is deployed. Look for accidental Disallow rules left over from staging environments, blocked CSS/JS paths that prevent proper rendering, and conflicting rules between user-agent blocks. Test key URL patterns using the URL Inspection tool, which now surfaces robots directives directly, rather than assuming.
XML Sitemaps
Confirm the sitemap only includes canonical, indexable, 200-status URLs (a sitemap padded with redirects, noindexed pages, or 404s signals low data quality to Google and can suppress crawl frequency for the whole site. Split sitemaps by content type (posts, products, categories) for sites over a few thousand URLs, and verify the sitemap index is actually referenced in robots.txt and submitted in Search Console. Check the “Discovered but not indexed” and “Crawled but not indexed” buckets in the Pages report) a large volume here on a mature site usually points to thin or duplicate content rather than a crawling problem.
Index Coverage in Search Console
Segment the Pages report by directory or template type. A spike in “Not Found (404)” often traces back to a bad migration or an internal linking bug; a spike in soft 404s (pages returning 200 with essentially no unique content) is common on faceted e-commerce and thin tag or category pages. Cross-reference against server logs to see whether Googlebot is actually requesting these URLs, GSC data is sampled and lags real crawl activity by a few days.
Site Architecture and Internal Linking
Flat, logical architecture keeps important pages within two or three clicks of the homepage and passes link equity efficiently. Run a full crawl with Screaming Frog or Sitebulb and check click depth for every URL (pages buried five-plus clicks deep rarely get crawled frequently or rank well, even with strong content. Look specifically for orphan pages (no internal links pointing to them at all) cross-reference the crawl against your XML sitemap and analytics data to find pages that get traffic but no internal links) and for internal link equity being wasted on low-value pages like tag archives, author pages, or paginated series.
Anchor text matters more than most audits give it credit for: descriptive, varied internal anchor text helps Google understand topical relationships between pages, which directly supports topical authority for competitive terms.
Core Web Vitals and Page Speed
Core Web Vitals remain a confirmed ranking signal, and the field-data gap between good and poor performers is significant, pages in the top LCP quartile see meaningfully higher organic CTR than pages that fail thresholds. The three metrics to check:
- Largest Contentful Paint (LCP): under 2.5 seconds. Usually a server response time, render-blocking resource, or unoptimized hero image problem.
- Interaction to Next Paint (INP): under 200 milliseconds. This fully replaced First Input Delay in the Core Web Vitals report and is far more sensitive to bloated JavaScript execution than FID ever was: heavy third-party scripts (chat widgets, ad tags, tag managers) are the usual culprit.
- Cumulative Layout Shift (CLS): under 0.1. Almost always missing width and height attributes on images and embeds, or web fonts causing layout reflow.
Tools and Workflow
Use PageSpeed Insights for a quick lab-plus-field snapshot of any single URL, the Chrome UX Report (CrUX) dashboard for site-wide field data trends over time, and Lighthouse (built into Chrome DevTools) for detailed lab diagnostics on individual pages during development. For ongoing monitoring, WebPageTest and DebugBear give more granular waterfall and filmstrip views than Search Console’s Core Web Vitals report, which only shows aggregated field data with roughly a 28-day lag. Fix the biggest offenders first: usually unoptimized images (serve WebP/AVIF, use responsive srcset), unused JavaScript and CSS (audit with Lighthouse’s “reduce unused code” flags), and third-party script bloat (audit with the PageSpeed Insights “reduce the impact of third-party code” section).
Mobile-Friendliness and Mobile-First Indexing
Google has indexed mobile-first for years now, meaning the mobile version of a page is what gets crawled, indexed, and ranked, not the desktop version. Verify parity: check that mobile pages carry the same content, structured data, headings, and internal links as desktop. It’s still common to find mobile templates that strip out content sections, hide internal links behind unexpanded accordions that don’t get crawled with the same weight, or serve smaller and cropped images. Test with the URL Inspection tool’s mobile rendering view, and manually check tap-target sizing and viewport configuration on templates that haven’t been touched recently.
Structured Data and Schema Markup
Structured data doesn’t directly boost rankings but it controls whether a page is eligible for rich results, and increasingly it’s how AI systems and answer engines parse page meaning. Audit for:
- Valid syntax, use the Rich Results Test and Schema Markup Validator; a single malformed property can disqualify an entire page from rich result eligibility.
- Coverage gaps: Article or BlogPosting schema on blog content, Product/Offer/AggregateRating on e-commerce, FAQPage and HowTo where genuinely applicable, BreadcrumbList site-wide, and Organization/WebSite schema at the domain level.
- Consistency with visible content, schema claiming a rating or price that doesn’t match what’s rendered on the page violates Google’s structured data guidelines and can trigger manual actions.
Check the Search Console “Enhancements” reports monthly for new errors, these regress silently after theme or plugin updates far more often than people expect.
HTTPS and Security
Confirm the entire site (not just the homepage) serves over HTTPS with a valid, non-expiring-soon certificate, and that HTTP requests 301-redirect to HTTPS rather than returning mixed content. Check for mixed-content warnings (HTTP images, scripts, or stylesheets loaded on HTTPS pages), verify HSTS is configured, and confirm there’s exactly one canonical version of the domain being served (www vs non-www, http vs https variants should all resolve, via redirect, to a single canonical host).
Duplicate Content and Canonicalization
Duplicate content dilutes ranking signals across multiple URLs instead of consolidating them onto one. Common sources: URL parameters (session IDs, sort and filter params on e-commerce category pages), print-friendly page versions, http/https and www/non-www variants, and staging or dev subdomains that got indexed by accident. Audit every template’s canonical tag with a crawler: check that self-referencing canonicals are present and correct, that paginated series use appropriate handling, and that faceted navigation either canonicalizes to the parent category or is blocked from indexing via noindex rather than a robots.txt disallow, which would prevent Google from even seeing the canonical signal.
Redirect Chains and Loops
Crawl the full site looking specifically for redirect chains (page A redirecting to B, which redirects to C, instead of A going directly to C) and loops (A redirecting back to itself through B). Each hop in a chain adds latency, dilutes a small amount of link equity, and multiplies the risk that Googlebot gives up before reaching the final destination. This is especially common after multiple migrations layered on top of each other over the years. Screaming Frog’s redirect chain report and Ahrefs’ Site Audit both map these clearly, fix by pointing every redirect directly at its final destination.
Log File Analysis and Crawl Budget
Server log files show exactly what Googlebot requested, when, and what status code it got back, ground truth that Search Console’s sampled data can’t fully replace. On sites with 10,000+ URLs, it’s common to see up to 30% of crawl activity wasted on duplicate, low-value, or blocked URLs, meaning important pages get crawled less often than they should. Pull at least 30 days of logs (Screaming Frog Log File Analyser or a direct server export) and check: which sections get crawled most and least frequently, whether new or updated pages get picked up within a reasonable window, and whether Googlebot is spending disproportionate time on faceted URLs, old paginated archives, or redirect chains instead of money pages.
International SEO and Hreflang
For sites targeting multiple countries or languages, hreflang errors are one of the most common technical issues found in audits: largely because the tags require every page to reference every other language and region variant, including itself, and one broken reciprocal link invalidates the whole cluster for that page. Validate with Search Console’s International Targeting report or Ahrefs’ hreflang audit, checking for missing return tags, incorrect language or region codes, and hreflang pointing to non-canonical or redirecting URLs.
JavaScript Rendering and AI Crawler Accessibility
This is the area that’s changed the most heading into 2026. Googlebot can render JavaScript via headless Chromium, but rendering isn’t guaranteed and adds a second wave of processing that can delay indexing by days. Test with the URL Inspection tool’s rendered HTML and screenshot view to confirm content and internal links appear post-render, not just in the initial HTML response.
More importantly: AI crawlers largely do not render JavaScript at all. GPTBot, ClaudeBot, and PerplexityBot fetch raw HTML and move on: content that only appears after client-side rendering is effectively invisible to them, which matters increasingly for visibility in AI-generated answers and citations. Practical steps: confirm critical content (headings, body copy, pricing, FAQs) is present in the initial server-rendered HTML rather than injected entirely by JavaScript, consider server-side rendering or static generation for key templates, and add an llms.txt file at the root pointing AI crawlers to canonical high-value pages such as the homepage, core service or product pages, documentation, and FAQs if the site depends on AI-driven discovery.
Essential Tools for a 2026 Audit
- Google Search Console: index coverage, Core Web Vitals field data, manual actions, structured data enhancements, International Targeting.
- Screaming Frog SEO Spider: full-site crawling, redirect chains, canonical and hreflang auditing, click-depth mapping.
- Ahrefs Site Audit: scheduled crawls, historical health-score tracking, hreflang and duplicate content detection at scale.
- Semrush Site Audit, similar crawl-based diagnostics with strong integration into keyword and traffic data for prioritization.
- PageSpeed Insights, Chrome UX Report, and Lighthouse, Core Web Vitals lab and field data.
- Screaming Frog Log File Analyser, crawl budget and log file analysis.
- Rich Results Test and Schema Markup Validator, structured data validation.
A 90-Day Technical SEO Action Plan
Weeks 1-2: Run a full crawl with Screaming Frog or Sitebulb, pull 30 days of log files, and export the full Search Console Pages and Core Web Vitals reports. Triage findings into critical (indexing blockers, security issues), high (Core Web Vitals failures on high-traffic templates, broken canonicals), and medium (redirect chains, schema gaps) priority buckets.
Weeks 3-4: Fix critical issues first, robots.txt errors, noindex tags left on live pages, broken canonical tags, mixed content. These are usually fast to fix and highest impact.
Month 2: Work through Core Web Vitals fixes template by template, starting with the highest-traffic page types. Clean up redirect chains and duplicate content sources found in the crawl.
Month 3: Roll out or repair structured data, validate hreflang if applicable, check JavaScript rendering and AI-crawler accessibility, and set up recurring monthly monitoring of crawl stats, Core Web Vitals, and index coverage so regressions get caught within weeks, not months.
Treat this checklist as a cycle, not a one-time project. Technical SEO health degrades quietly: a plugin update breaks structured data, a developer ships an unoptimized hero image, a migration introduces a redirect chain nobody notices for months. The sites that consistently win the technical fight are the ones auditing every month between full audits, not the ones that only look when rankings have already dropped.
Related
Want this checked on your own site?
A 20 minute review, no obligation. I will tell you honestly which of this actually applies to you and which does not.
Written by Shahzaib Ul Hassan, senior AI SEO consultant and founder of ShazzSEO. Ranking sites since 2009.
Working through a checklist yourself is the right call if you have the time. If you would rather have the two things that actually matter named for you, that is what an SEO audit is for, at a fixed price with the fee credited against the fix.