SEO Audit Checklist: Check website and correctly prioritize actions

The checklist creates a common evaluation framework, makes gaps visible, and prevents individual tool scores from determining the entire audit. It can be used completely for small websites and based on representative templates, directories, and URL samples for larger websites.

Your boundaries should be documented beforehand:

• A crawl does not automatically identify which pages support revenue, leads, or key user journeys.
• Search Console shows Google data, but not every known query or individual URL is fully displayed in all reports.
• Labortests do not replace real user data; average values can mask problems in individual templates.
• Without CMS, analysis, server, or log access, some causes remain unexplained.
• An audit reveals technical conditions and data patterns. It cannot explain every ranking change in a single cause.
• Neither correct technology nor structured data guarantees crawling, indexing, rankings, rich results, or visibility in generative answers.

Suspicious indicators are therefore marked as a hypothesis. A verifiable finding refers to data source, affected URLs, reproduction path, and expected desired behavior.

Before the actual SEO audit, business objectives, markets, languages, domains, subdomains, and key conversion paths are identified. Then, page types such as services, categories, products, guides, locations, and PDFs are inventoried.

Starting point should include as follows:

• complete or meaningful segmented crawl
• XML Sitemaps and CMS Export
• Google Search Console and Web Analysis
• Ranking, backlink, and conversion data, if available
• Server logs for crawl or infrastructure issues
• Releases, relaunches, campaigns, and well-known outages as event timeline

Comparison periods must account for seasonality and changes. Only once it is clear which pages and results are important can a finding be weighted.

Now the desired URL inventory is being matched against the actual state:

• Is the URL internally linked and crawlable?
• Allow robots.txt, Meta Robots, and X-Robots tag for the intended access?
• Does the page return status 200 and a complete, renderable content?
• Does the Canonical point consistently to the preferred URL?
• Does the sitemap contain only intended, canonical, and indexable URLs?
• Create filters, parameters, searches, calendars, or session IDs unnecessary variants?
• Are remote content professionally relevantly forwarded or do they return a true 404-/410-status?
• Do internal links, redirects, canonicals, hreflang, and sitemaps all point to the same target?

robots.txt primarily controls crawling and is not a reliable substitute for noindex. To see a noindex signal, the crawler must be allowed to retrieve the URL. A sitemap supports the discovery of preferred URLs, but it guarantees neither crawling nor indexing. Anomalies from the indexing report are therefore verified at specific URL samples using URL checking.

Every prioritized page requires a visible task. Check:

• Does the page type and content align with the user’s search intent and target audience?
• Does the page fully and without unnecessary detours answer the central question?
• Does she provide her own experience, data, examples, tools, or a verifiable context?
• Are statements accurate, up-to-date, and supported by primary sources when needed?
• Are author, company, responsibility, and update status transparent?
• Are there outdated, redundant, very similar, or conflicting pages?
• Does the content support a meaningful next action?

Google recommends helpful, reliable, and human-created content. Google does not specify a minimum word count. Content should therefore not be simply extended, compiled, or deleted merely to appear to comply with supposed SEO rules. Decisions regarding updates, consolidation, or removal require URL performance data and a dedicated subject matter page. Deep dive: Content SEO.

On-page checks are not only evaluated page by page, but also over templates and directories:

• clear, descriptive page title and a clear main heading
• logical subheadings and understandable content structure
• Useful meta description without unverified promises
• readable, stable URLs and consistent Breadcrumbs
• compelling image alternatives, when images convey information
• appropriate side roller within Hub, Category, Service, and Guide
• no unintentional cannibalization of multiple URLs with the same task
• structured data only for visible and actually relevant content

SEO titles and H1 tags may differ in phrasing but should express the same page objective. Automatically generated metadata is verified against real examples; a simple completeness rate says little about its quality.

Conducting an SEO audit also means checking whether the measurement is robust:

• Search Console property, sitemap, and relevant directories correctly detected
• Analytics delivery in conjunction with consent tested
• Forms, phone, email, appointment, and purchase events reproducible
• organic landing pages and conversion paths meaningfully segmented
• considers internal accesses, referral exclusions, and cross-domain paths
• Documented changes and outages in an event timeline
• KPI definitions, time zone, filters and comparison periods clearly defined

Search Console and web analysis measure different aspects and can display divergent values. Instead of adjusting numbers to fit, definitions and data gaps are documented. Regular monitoring thereafter controls implementation, crawling, indexing, search performance, and business signals collectively.

For Google’s AI Overviews and AI Mode, Google states the same fundamental SEO practices apply. A page must be indexed and included in search snippets; Google does not specify any additional technical requirements.

A meaningful GEO audit therefore considers:

• clear entities for companies, individuals, offerings, and topics
• consistent names, profiles, and verifiable company information
• precise answers, definitions, and comparisons in visible content
• original sources, primary data and clearly attributable expertise
• structured data in alignment with visible statements
• External mentions and source context
• reusable topic and prompt sets instead of individual screenshots

GEO is not a schema trick. The inclusion or positioning within a generated response cannot be guaranteed. Moreover, other response systems may employ their own access, source, and measurement logics; these must be verified separately.

A prioritized action list answers more than „errors present.“ Evaluate each finding according to:

• Evidence: verified, reproducible or still hypothesis?
• Scope: single URL, template, directory, or entire domain?
• Effect: Does it block crawling, indexing, usage, or conversion?
• Business relevance: Which query and which user journey are affected?
• Effort and risk: How complex is the change and what can it inadvertently alter?
• Dependencies: Which teams, systems, or preliminary work is required?

Critical blockers like accidental noindex, widespread 5xx errors or failed conversions come first. Then follow verified actions with high impact. Hygiene tasks with no discernible effect are grouped or deferred.

Every ticket should include problem, evidence, example URLs, desired state, ownership, and acceptance test. This turns a long SEO checklist into a actionable plan.

• Treat a tool-health score as a business objective
• only the homepage or only URLs from the sitemap crawl
• classify every non-indexed URL as an issue
• Evaluate canonicals, redirects, and robots rules in isolation
• Content evaluated based on word count rather than user task
• JavaScript content to be evaluated only in the source code or only in the browser
• Mobile, page types, and real user data not segmented
• Rankings without seasonality, releases and search intent comparison
• hundreds of recommendations without accountable parties and acceptance testing handed over
• after implementation, no control crawl and no monitoring to be conducted