Ask five agencies what's in their SEO audit and you'll get five different answers, ranging from a genuinely thorough technical review to a twelve-page PDF generated by a tool nobody has customised in years. Knowing what should actually be checked lets you tell the difference before you pay for either.
This matters beyond just avoiding a wasted invoice. A shallow audit gives false confidence, a business believes its technical foundation is solid because a report said so, then wonders months later why content and link building investments aren't producing the expected results. The missing piece was usually sitting in the audit that never actually happened properly.
The list below is roughly the order a thorough audit should work through, though the specific priorities shift depending on the site's age, platform and history.
Crawlability and indexation
This comes first because nothing else matters if search engines can't find and index your pages. A proper audit checks your robots.txt for accidental blocks, reviews your XML sitemap for accuracy and completeness, confirms canonical tags point where they should, and pulls actual crawl data from Google Search Console to see how the site is really being indexed rather than assuming it matches what's intended.
AI crawler access
This is now a standard line item, not an optional extra. Check whether GPTBot, Google-Extended, PerplexityBot and similar crawlers are blocked in robots.txt. A site can rank perfectly well in classic Google search while being invisible to every AI answer engine simply because of one overlooked directive.
Site performance and Core Web Vitals
Largest Contentful Paint, Interaction to Next Paint and Cumulative Layout Shift are measured against real user data, not just a lab test score. A page that scores well in a synthetic test but loads slowly for actual visitors on mobile connections is still a performance problem. The audit should identify specific causes, oversized images, render-blocking scripts, unoptimised fonts, rather than a generic "improve page speed" recommendation that gives no one anything to action.
Site architecture and internal linking
How pages connect to each other tells search engines and AI models what matters most on the site. An audit should map the click depth from the homepage to every important page, flag orphaned pages with no internal links pointing to them, and check whether the site's structure reflects a genuine content hierarchy or has grown ad hoc over several years of adding pages wherever seemed convenient at the time.
Structured data and schema markup
Schema markup helps both classic search results and AI tools understand exactly what a page is about, whether that's a service, an article, an FAQ or a local business listing. A proper audit checks which schema types are already implemented, validates them for errors, and identifies where missing schema is a genuine opportunity rather than recommending markup for its own sake.
Content quality and duplication
This covers thin or duplicate content, missing or poorly written title tags and meta descriptions, and pages that target the same keyword as another page on the same site. It's also where cannibalisation gets caught, two pages competing for the same search rather than each covering distinct ground, which is one of the more common and most overlooked issues on established websites.
Backlink profile health
A full audit reviews the quality and relevance of existing backlinks, flags anything that looks spammy or manipulative, and benchmarks the overall link profile against genuine competitors rather than presenting a raw number as though more links always means better rankings.
What a checklist should look like in practice
Pulling this together, a genuinely useful audit checklist covers:
- Robots.txt and AI crawler access, checked directly rather than assumed
- XML sitemap accuracy and Search Console crawl and index data
- Core Web Vitals from real user data, with specific causes identified
- Site architecture, click depth and orphaned pages
- Structured data coverage and validation errors
- Duplicate content, thin pages and keyword cannibalisation
- Backlink profile quality, relevance and any manipulative patterns
What this looks like on a real website
Take a mid-sized B2B site with around three hundred pages. An audit might reveal that forty pages are accidentally blocked from AI crawlers through an overly broad robots.txt rule added during a security review two years ago, that the sitemap still references fifteen URLs deleted during a rebrand, and that twelve service pages have near-identical title tags because they were built from the same template without being customised. None of these are individually catastrophic, but together they explain a plateau in both classic rankings and AI visibility that no amount of new content would have fixed on its own. This is exactly why an audit needs to look across every category rather than assuming the problem sits wherever it's easiest to check first.
Why order matters
Fixing content quality before crawlability is fixed is wasted effort, since new content can't be indexed if the crawl setup is broken. Working through issues in the order above, from foundational access issues through to backlink quality, means each fix actually has a chance to show its full impact rather than being masked by a more fundamental problem further up the list.
Common objections, answered
Our website was audited last year, do we need another one?
Yes, periodically, and sooner if anything significant has changed: a redesign, a platform migration, a large content push, or simply the passage of twelve months. Technical issues accumulate quietly, a plugin update breaks a redirect, a new content type ships without proper schema, AI crawler directives get added during an unrelated security review. An audit that's a year old is a snapshot of a website that no longer fully exists.
Can't we just run an automated tool and get the same result?
Automated tools are useful for surfacing obvious issues quickly, but they can't judge whether your site architecture genuinely reflects your business priorities, whether two pages are quietly cannibalising each other, or whether your content actually demonstrates the expertise a specific audience is looking for. A tool produces a list. A proper audit produces judgement about what that list actually means for your business.
How do we know which issues to fix first?
Prioritise by a combination of impact and effort, not just severity. A moderate issue affecting your highest-traffic page usually deserves attention before a severe issue on a page nobody visits. A properly run audit should rank findings this way explicitly, rather than leaving you to guess which of forty flagged issues actually matters most for your specific business.
What happens after the audit
A checklist is only useful if it turns into prioritised, actioned work. The strongest audits rank findings by both impact and effort, so a business can see clearly what to fix first rather than facing a flat list of forty issues with no sense of where to start. A proper SEO audit should end with exactly that: a prioritised plan, not just a diagnosis.
Want a technical SEO audit that ends in a clear, prioritised plan rather than a long list of issues?
Book a call