Two websites can carry the same quality of writing, yet only one of them ranks. The difference often comes down to what search engines can actually see behind the page: whether it gets crawled, rendered, and indexed correctly in the first place. That underlying work is technical SEO, and in 2026 it also decides whether AI systems such as ChatGPT and Perplexity can read a site at all.
What is technical SEO?
Technical SEO is the work done to a website's infrastructure so that search engines, and increasingly AI crawlers, can find, read, and interpret its pages correctly. It sits underneath content and link building, closer to how a site is built than to what it says.
A site can carry strong writing and a healthy backlink profile and still fail to rank if its technical foundation blocks discovery. Broken crawl paths, slow load times, or missing structured data can quietly undo the rest of an SEO strategy without ever showing up in the content itself.
What does a technical SEO checklist cover?
A technical SEO checklist covers crawlability, indexing, site speed, mobile usability, security, structured data, and AI crawler access, working through each area systematically rather than guessing at fixes.
| Area | What it covers |
|---|---|
| Crawlability | Whether Googlebot and other crawlers can reach and read key pages |
| Indexing | Whether crawled pages are actually stored and eligible to rank |
| Site speed and Core Web Vitals | Loading, responsiveness, and visual stability, measured against Google's thresholds |
| Mobile usability | Whether the site works cleanly across screen sizes |
| Security | HTTPS, valid certificates, and no mixed-content warnings |
| Structured data | Schema markup that tells machines what a page represents |
| AI crawler access | Whether GPTBot, PerplexityBot, ClaudeBot, and similar bots can reach the site |
Crawlability and indexing
Crawlability comes first because nothing else matters if search engines cannot reach a page. A misconfigured robots.txt file, an incomplete XML sitemap, or a long chain of redirects can hide entire sections of a site without any visible warning on the page itself.
Indexing is a separate check that often gets missed. A page can be crawled successfully and still fail to appear in search results if it carries a stray noindex tag, a canonical pointing to the wrong URL, or duplicate content competing against itself.
Site speed and Core Web Vitals
Site speed is measured through three metrics: Largest Contentful Paint, Interaction to Next Paint, and Cumulative Layout Shift, collectively known as Core Web Vitals. Google's own thresholds classify a page as good when LCP stays under 2.5 seconds, INP stays under 200 milliseconds, and CLS stays under 0.1, measured at the 75th percentile of real visits.
| Metric | Measures | Good threshold |
|---|---|---|
| LCP (Largest Contentful Paint) | Loading speed | Under 2.5 seconds |
| INP (Interaction to Next Paint) | Responsiveness | Under 200 milliseconds |
| CLS (Cumulative Layout Shift) | Visual stability | Under 0.1 |
Most sites still fall short of these numbers. The HTTP Archive's 2025 Web Almanac found that only 48% of mobile pages and 56% of desktop pages pass all three Core Web Vitals at once, which means more than half the web is still working through this exact list.
Mobile usability and security
Mobile usability matters because Google indexes the mobile version of a site by default, not the desktop one. Text that overlaps, buttons placed too close together, or content that only loads on larger screens can all undermine rankings even when the desktop experience looks fine.
Security is a shorter check but not an optional one. HTTPS should be active sitewide, certificates need to stay valid, and no page should trigger a mixed-content warning by loading scripts or images over an unencrypted connection.
Structured data and AI readiness
Structured data, delivered through schema markup, tells search engines and AI systems what a page actually represents rather than leaving them to infer it from surrounding text. Article, Organization, FAQ, and Breadcrumb schema are among the most commonly used types, and each one adds a layer of machine-readable context to a page.
AI readiness is the newest addition to this list. Checking robots.txt for accidental blocks on GPTBot, PerplexityBot, or ClaudeBot has become a standard step, since a site can be perfectly optimised for Google and still be invisible to AI-generated answers if these crawlers are blocked.
What is a simple example of technical SEO?
A simple example of technical SEO is adding an XML sitemap so search engines can find every page on a site without relying on internal links alone. Enabling HTTPS is another common example, and so is compressing an oversized image or fixing a broken redirect chain. Even a canonical tag, which tells search engines which version of a duplicate page to treat as the original, counts as technical SEO in its simplest form.
None of these fixes involve writing new content. They adjust how a site is built and delivered, which is exactly what separates technical SEO from the content and off-page work covered elsewhere in this series.
How do you run a technical SEO audit checklist?
A technical SEO audit checklist starts with a full crawl of the site, usually with a tool such as Screaming Frog or Ahrefs, to surface broken links, redirect chains, and duplicate content before anything else is reviewed. From there, indexing status gets checked against Google Search Console to confirm that important pages are actually eligible to rank.
Core Web Vitals come next, tested through PageSpeed Insights or a similar tool against real user data rather than a single lab test. Structured data is then validated, followed by a check of mobile rendering and a final pass through robots.txt to confirm both traditional and AI crawlers have the access they need.
How often should a technical SEO audit be carried out?
Audit frequency depends mainly on site size and how often the site changes. Large or frequently updated sites, such as e-commerce or news platforms, benefit from a monthly check, while smaller business sites can typically manage on a quarterly cycle.
An audit outside that regular schedule is also worth running after a redesign, a platform migration, or a sudden, unexplained drop in traffic. Waiting for a scheduled review in these situations usually means losing weeks of visibility that a same-week check could have caught.
Keeping a website technically sound is an ongoing commitment rather than a single project, and it works best alongside the content and off-page efforts covered elsewhere in this series. Margo Digitech runs this kind of audit and monitoring on an ongoing basis for businesses that would rather not manage it in-house.
Frequently Asked Questions
Crawlability and indexing come first, since nothing else on the checklist matters if search engines cannot reach or store a page in the first place.
A small site can usually be audited within a day or two. Larger sites with thousands of URLs often take one to two weeks, particularly when log-file analysis is involved.
Yes. AI systems rely on the same crawling and structured data foundations as traditional search, so a technically weak site is likely to be missed by both.
It can, particularly after large changes such as a site migration or a major redirect cleanup. Most fluctuations settle within a few weeks as search engines recrawl and reindex the affected pages.
Screaming Frog and Ahrefs handle the crawling, Google Search Console tracks indexing data, and PageSpeed Insights covers Core Web Vitals. Together, they account for most of what a standard audit needs.