This technical SEO checklist is built in dependency order. Fix crawling before speed, speed before structured data, and keep monitoring running underneath all of it. The layers follow the sections of Backlinko’s technical SEO guide for 2026: site structure and internal links; crawling, rendering and indexing; site performance and reliability; semantic HTML and structured data; and maintenance and monitoring.
Each item below links to the post where we cover it in depth. Treat this page as the hub and the linked posts as the working detail.
The technical SEO checklist at a glance
| Layer | Check | Full guide |
|---|---|---|
| Structure | Every important page sits within a few clicks of the homepage | Internal linking for SEO |
| Structure | No page you want ranked has zero internal links | Orphan pages |
| Structure | Sections live where their authority should live | Subdomain vs subfolder |
| Crawling | robots.txt blocks only what it should | Robots.txt for SEO |
| Indexing | No accidental noindex on pages you want found | Excluded by noindex tag |
| Indexing | New pages get discovered and indexed | How to get indexed by Google |
| Indexing | Crawled pages are not left out of the index | Crawled, currently not indexed |
| Indexing | One canonical version of each page | Duplicate content |
| Indexing | Moved URLs use the right redirect | 301 vs 302 redirects |
| Performance | Core Web Vitals pass on mobile | Core Web Vitals assessment failed |
| Semantics | Images carry descriptive alt text | Image alt text |
| Semantics | Structured data matches the features you target | SERP features |
Layer 1 and 2: structure, crawling and indexing
Start with structure because it decides what a crawler can reach at all. A flat architecture, where important pages are a short click path from the homepage, beats any amount of sitemap tuning. Run a crawl and sort by depth: pages buried deep, or reachable only through pagination, are the first to be crawled less often.
Then check the gates. A single line in robots.txt or a leftover noindex from a staging site can hide a whole section, and neither throws a visible error on the page. Compare the URLs in your XML sitemap against what Search Console reports as indexed. The gap between the two lists is your indexing backlog, and it usually has a pattern: a template, a parameter, a section.
Duplicates and redirects belong in this layer too. Several URLs serving the same content split signals between them, and a chain of redirects wastes crawl on every hop. The fix is dull and reliable: one canonical URL per page, one redirect from each old address straight to its final destination.
Rendering is the part of this layer that most audits wave through. If your main content, internal links or canonical tags only appear after JavaScript runs, compare the raw HTML with the rendered page. Anything that exists only in the rendered version depends on a second, slower processing step, and links that are injected late are easy to miss. The safe default is simple: put the content and the links that matter in the HTML the server sends.
Layer 3: performance and reliability
Performance is where checklists overreach. Backlinko cites the Core Web Vitals targets as LCP at or under 2.5 seconds, INP at or under 200 milliseconds, and CLS at or under 0.1. Those are pass marks, not a race. Shaving a fast page from 1.8 seconds to 1.2 is rarely worth the engineering time; moving a template that fails on mobile into the passing range is.
Judge performance by template, not by URL. A site with thousands of pages usually has a handful of templates, and one slow product or category template explains most of the failing URLs. Fix the template once and the whole group moves together.
Reliability matters as much as speed. Server errors, timeouts and slow responses during a crawl tell a search engine your site is costly to fetch. HTTPS is now table stakes: Backlinko cites Web Almanac data showing that over 90% of mobile websites use it. If a site still serves mixed content or old HTTP URLs that do not redirect, that is a structural fault, not a polish item.
Layer 4: semantic HTML and structured data
Once pages are reachable, indexable and fast enough, help machines read them. That means one H1, headings that describe their sections, real lists and tables rather than styled paragraphs, and alt text that says what an image shows. None of this is glamorous, and all of it is cheap.
Structured data comes last within this layer, and the received wisdom oversells it. Markup makes a page eligible for certain search features; it does not guarantee them, and it does not make a weak page rank. Add the types that match features you actually want, validate them, and make sure the markup describes what is visibly on the page. Markup that claims something the reader cannot see is a liability, not an asset.
Layer 5: maintenance and monitoring
Technical SEO decays. A plugin update adds a noindex, a redesign drops internal links, a migration leaves redirect chains. Backlinko gives monitoring its own section for this reason, and it is the layer most sites skip because nothing looks broken on the day.
- Weekly: Search Console indexing and crawl stats reports, looking for sudden shifts rather than absolute numbers.
- Monthly: a full crawl compared with last month’s, so new orphans, broken links and redirect chains stand out.
- After every release: spot-check robots.txt, canonical tags and meta robots on the templates that changed.
- Quarterly: server logs, to see where crawlers actually spend their time.
Server logs are the most underused item on any technical SEO audit checklist. A crawl tool shows what a crawler could reach; logs show what it did reach. When the two disagree, the logs win.
What a technical checklist cannot fix
Here is the limitation. A clean technical audit removes reasons not to rank; it does not supply reasons to rank. A site that passes every item above can still sit on page three because its content is thin or its topic coverage is shallow. Technical work is a multiplier on content and links, and a multiplier on zero is still zero.
The checklist also cannot tell you which items matter most on your site. A small blog with fifty pages rarely has a crawl problem worth a week of work; a large catalogue with faceted navigation almost certainly does. Size the effort to the site, and start wherever the crawl and Search Console data show the biggest gap.
That is why this list stops at the technical layer. For the whole-site review, covering content quality, competitors, backlinks and AI visibility, use our SEO audit checklist. More guides sit in the strategy archive.
Frequently asked questions
What should a technical SEO checklist include?
Five layers: site structure and internal links; crawling, rendering and indexing; performance and reliability; semantic HTML and structured data; and ongoing monitoring.
What are the Core Web Vitals thresholds?
LCP at or under 2.5 seconds, INP at or under 200 milliseconds, and CLS at or under 0.1.
How often should I run a technical SEO audit?
Run a full crawl monthly, check Search Console weekly, and spot-check robots.txt, canonicals and meta robots after every release.
Is a technical SEO audit the same as an SEO audit?
No. A technical audit covers how search engines reach and read the site. A full SEO audit adds content, competitors, backlinks and visibility.
The takeaway Work the layers in order: structure and indexing first, then performance, then semantics, with monitoring running underneath. Technical fixes remove barriers; they do not create rankings on their own.
