1. Crawlability
If Google cannot crawl it, nothing else matters. These checks come first in every audit because they are the most common source of catastrophic, invisible problems.
- robots.txt does not accidentally block CSS, JS or whole sections.
- XML sitemap exists, is referenced in robots.txt, and contains only canonical, indexable, 200-status URLs.
- Sitemap lastmod values are accurate — Google discounts sitemaps with unreliable lastmod.
- No orphan pages: every commercially relevant URL is reachable within three clicks of the homepage.
- Pagination uses crawlable <a href> links, not JavaScript-only controls.
- Infinite scroll has paginated equivalents for crawlers.
- Server response times (TTFB) are under 500ms at the origin.
- CDN is configured with correct cache headers for HTML and static assets.
2. Indexation
Crawl budget is finite. Every URL Google indexes that should not be indexed is capacity taken away from one that should.
- Indexed-to-submitted ratio in Search Console is above 90% for commercial pages.
- Faceted and filtered URLs are either blocked, canonicalised or made noindex — with a deliberate decision per pattern.
- Soft 404s and thin pages are eliminated or consolidated.
- Canonical tags are self-referencing on unique pages and point correctly on duplicates.
- Parameter handling is explicit rather than left to Google.
- Staging and development environments are noindex and password protected.
- Trailing slash and case consistency enforced with 301s.
- Pagination pages are not canonicalised to page one (a common and damaging error).
3. Rendering
JavaScript-rendered content is a frequent cause of partial indexing. We test with rendering disabled, because if the content is not in the initial HTML it is at risk.
- Primary content and internal links are present in the server-rendered HTML.
- Google's URL Inspection tool shows content matching what users see.
- Client-side navigation does not break history or deep linking.
- Dynamic rendering or server-side rendering is in place where a JavaScript framework is used.
4. Core Web Vitals
Field data is what counts. Lab scores are useful for diagnosis but CrUX and Search Console reflect what your users experience.
- LCP under 2.5s at the 75th percentile on mobile.
- INP under 200ms — check third-party scripts first.
- CLS under 0.1 — reserve space for images, ads and embeds.
- LCP image is preloaded and served in AVIF or WebP.
- Fonts are subset, preloaded and use font-display: swap.
- Third-party scripts are audited individually; anything not earning its cost is removed.
- Layout has no unexpected shifts on mobile after data loads.
5. Structured data
Schema does not directly rank pages, but it influences how they display and how confidently AI engines interpret them.
- Organization markup with @id, logo, sameAs and contact details.
- WebSite markup with SearchAction.
- BreadcrumbList on every inner page.
- Product and Offer for e-commerce, Service for service pages.
- FAQPage only where the content is genuinely visible on the page.
- LocalBusiness per location with accurate NAP and geo.
- All markup validated in the Rich Results Test and Schema validator.
6. Speed, security and infrastructure
These sit adjacent to SEO but affect it directly through ranking signals, crawl reliability and trust.
- HTTPS sitewide with no mixed content and HSTS enabled.
- 301 from www to non-www (or the reverse) applied consistently and only once.
- Security headers present: X-Content-Type-Options, Referrer-Policy, Permissions-Policy.
- Image formats optimised and dimension attributes present everywhere.
- Caching strategy defined for HTML, assets and API responses.
- Monitoring in place for uptime, SSL expiry and error rates.
7. Migration safety
If you are replatforming, these five checks prevent the traffic loss that makes migrations infamous.
- Full crawl baseline captured before launch, with ranking and traffic snapshots.
- Complete 1:1 redirect map, validated in staging before DNS switch.
- Metadata and schema parity confirmed between old and new templates.
- Internal link integrity checked after the switch.
- Daily Search Console monitoring for fourteen days post-launch, with a rollback plan ready.
Key takeaways
- Crawlability and indexation problems are invisible in analytics but cost more than any content gap.
- Always evaluate Core Web Vitals on field data, not lab scores.
- Test rendering with JavaScript disabled — that is close to what a crawler may see.
- Every migration needs a pre-launch baseline, a redirect map and post-launch monitoring.
Ayesha Malik
Head of SEO Strategy
Leads the SEO practice across all markets, specialising in technical SEO, international architecture and large-scale migrations.