Technical SEO Checklist
Technical SEO has one job: remove every reason a search engine cannot crawl, index, understand or trust your pages. Work down this list in order - the first section is binary, and nothing below it matters until it passes.
Indexability: can the page be in the index at all?
A page can only rank if it returns 200, carries no noindex directive, is not blocked in robots.txt, and canonicalises to itself. Any one of these failing makes everything else on the page irrelevant.
The most common serious fault is not exotic. It is a staging noindex shipped to production, or a template change that points every canonical at the homepage. Both are invisible in the browser and both remove a site from search. This is the argument for re-crawling your own site on a schedule rather than only when something looks wrong.
- 1Every URL you want ranked returns HTTP 200, not 3xx, 4xx or a soft 404.
- 2No noindex in the robots meta tag or the X-Robots-Tag response header.
- 3robots.txt does not disallow the path - and does not disallow the CSS or JS needed to render it.
- 4Each page has a self-referencing rel=canonical. Cross-canonicals merge pages you meant to keep.
- 5One version of the site is canonical: HTTPS, one of www or non-www, with the others redirecting once.
- 6Pagination, filters and session parameters do not generate infinite crawlable variants.
Crawlability: can the crawler reach and afford your pages?
Search engines allocate a finite crawl budget per site. Reachability through internal links, a clean sitemap and a fast server determine how much of your site gets seen and how quickly changes are noticed.
- XML sitemap
- Lists your canonical, indexable URLs only. A sitemap containing redirects, 404s or noindexed pages teaches the crawler to trust it less.
- Internal links
- Every important page reachable within three clicks of the homepage, with descriptive anchor text. Orphan pages are usually the ones that never rank.
- Server response time
- Slow responses reduce how much a crawler fetches per visit. On a large site this is a direct ceiling on how much of it gets indexed.
- Rendering
- Critical content in the HTML rather than assembled by client-side JavaScript. What is not in the response may not be indexed and will not be quoted.
On-page: does the page say what it is about?
Every indexable page needs a unique title tag containing its target query, a unique meta description, exactly one H1, a logical heading hierarchy and descriptive alt text on meaningful images.
Duplicated titles and descriptions across a template are the most common on-page fault at scale, and the easiest to fix. If two pages carry the same title, you have told the search engine they are the same page, and it will pick one.
- 1Unique title tag per page, target query near the front, roughly 50-60 characters.
- 2Unique meta description per page. Not a ranking factor, but it decides the click.
- 3Exactly one H1, stating what the page is about. Multiple H1s dilute the signal.
- 4H2s and H3s in a real hierarchy, phrased as the questions the section answers.
- 5Descriptive alt text on images that carry meaning; empty alt on decorative ones.
- 6Enough body copy to answer the query properly - thin pages rank thinly.
Structured data: is it accurate?
Add Organization, WebSite and BreadcrumbList sitewide, plus the type that matches each page - Article, FAQPage, HowTo or Product. Accuracy matters more than coverage: markup describing something not visible on the page is a manual-action risk.
Structured data does not directly raise rankings. It makes a page eligible for richer results and makes its content easier to parse, which matters increasingly for AI Overview citation. Both benefits depend on the markup honestly describing the page.
The failure mode to avoid is marking up FAQs that do not appear on the page, or review scores no user ever left. It is one of the few SEO mistakes that can result in a manual penalty rather than merely wasted effort.
Performance and security
Serve everything over HTTPS, pass all three Core Web Vitals, and make sure the mobile rendering is the full experience. Mobile-first indexing means the mobile version is the version being judged.
- 1HTTPS everywhere, with HTTP redirecting once to the canonical HTTPS URL.
- 2A viewport meta tag, and a layout that works at 360px wide without horizontal scroll.
- 3LCP under 2.5s, INP under 200ms, CLS under 0.1 - all three, on mobile.
- 4Images sized and compressed for the viewport that requests them.
- 5No content hidden on mobile that exists on desktop - the mobile version is what gets indexed.
Frequently asked questions
How often should I run a technical SEO audit?
Monthly for a stable site, weekly if you deploy frequently. Most serious technical faults are introduced by a deployment rather than appearing on their own, so audit frequency should track release frequency.
Should I fix every issue an audit tool reports?
No. Fix everything in the indexability section, because those are binary. Below that, judge by consequence - a missing alt attribute on a decorative icon is not worth an engineering ticket.
Does a page need a canonical tag if it is the only version?
Yes. A self-referencing canonical costs nothing and protects you from duplicate URLs created later by tracking parameters, trailing slashes or pagination.