Short answer: a technical SEO checklist has one job, which is to make sure a search engine can reach your pages, read them, and store them. Everything else is decoration until those three are true. Work through crawlability, indexation, speed, structure and duplication in that order, because each one gates the next.
Last updated 20 August 2026. Every threshold quoted here is Google’s own published figure, not a rule of thumb. The scores we mention for our own site are ones you can measure yourself.
Most technical SEO checklists you find are a list of two hundred items with no order to them. That is not useful, because the items are not equal. If a page cannot be crawled, its heading structure does not matter. If it is crawled but not indexed, its speed does not matter either. So this checklist runs in dependency order: each section only becomes worth doing once the one above it is true. Once the crawl and index layers pass, the head of each page is next — our guide to meta tags for SEO covers that layer tag by tag.
Technical SEO checklist, step 1: can a crawler reach the page?
Start with robots.txt. One careless Disallow line can remove a whole section from search, and it will not announce itself. Open yours at /robots.txt and read every line out loud.
Then check the sitemap. It should list the pages you want found, and only those. Archives, tag pages and thin taxonomy pages do not belong there. Finally, make sure the pages are actually linked from somewhere. A page reachable only by typing its URL is a page nobody crawls often. After a redesign the sitemap also has to list the new addresses instead of the old ones, one of the launch-day checks in our guide to protecting your rankings during a site redesign.
Increasingly this section is about more than Google. AI assistants use their own crawlers, and blocking them is a decision rather than an accident. We cover which ones to allow in how to get found by AI, and the Bing side in how to get found on Bing.
Step 2: is the page actually indexed?
Reachable and indexed are not the same thing. Google can crawl a page and decide not to store it, which shows up in Search Console as “crawled, currently not indexed”. That message means the page was judged and set aside.
Check three things when you see it. Does the page carry a noindex tag it should not? Is the canonical pointing somewhere unexpected? And is the content thin enough that Google reasonably passed on it? A page of twenty words usually deserves the verdict it got.
| What Search Console says | What it usually means | What to do |
|---|---|---|
| Discovered, not indexed | Queued, not yet crawled | Wait, and link to it from a page that is crawled often |
| Crawled, not indexed | Judged and set aside | Improve the page, then request indexing |
| Excluded by noindex | You told it to stay out | Check whether you meant to |
| Duplicate, no canonical | Two pages saying the same thing | Merge them, or set the canonical |
One warning from our own work: the indexing report lags by days. Always confirm a specific URL with the inspection tool rather than trusting the summary, because pages listed as not indexed are sometimes indexed already.
Step 3: is it fast enough on a phone?
Google publishes the thresholds, so there is nothing to guess. According to Core Web Vitals, the largest element should paint within 2.5 seconds, interaction to next paint should stay under 200 milliseconds, and layout shift under 0.1. Those are measured on real visits, not in a lab.
Fix them in order of return: compress and resize images, remove plugins you do not use, enable caching, then drop any third-party script you cannot justify. Most sites gain more from those four than from anything else. Our own site scores 99 to 100 on mobile while running a background video, and the full method is in how to make your website faster.
Step 4: can a machine read what it says?
- One H1 per page, saying what the page is about in the words people search.
- Headings in order, describing sections rather than decorating them.
- Alt text on every meaningful image, describing the image rather than repeating a keyword.
- Structured data where it fits: LocalBusiness, Service, FAQPage. Only mark up what is visible on the page.
- Internal links with descriptive anchors, placed mid-sentence rather than as a “click here” at the end.
That last one does more than most people expect, because internal links are how you tell a search engine which of your pages matters. We measured our own and published the numbers in internal links: why they matter for SEO.
Step 5: is anything competing with itself?
Two pages targeting the same phrase do not double your chances. They split them, and Google picks one, often not the one you wanted. This is the failure mode nobody notices, because both pages look fine on their own.
Find it by searching your own site for a phrase and seeing how many pages come back. If two do, decide which one owns it, and rewrite the other to cover a different angle. Location pages are where this bites hardest: five near-identical neighborhood pages are five pages competing, not five chances to win. Build one strong page per genuinely different area instead, which is the reasoning behind the local SEO New York playbook and our local SEO service page.
Once all five steps are true, technical SEO stops being the bottleneck and content starts being the work. That is the point of running them in order: you find out where you actually are, and what a good SEO audit tells you is the longer version of the same idea.
Common questions about a technical SEO checklist
How often should I run a technical SEO checklist?
Fully, twice a year. Briefly, after any change to the theme, the hosting or the plugin list, because those are what break crawling and speed without warning. A five-minute check after a change beats a full audit six months later.
Do I need paid tools for this?
No. Search Console and PageSpeed Insights are free and come from Google itself, which makes them the most reliable sources you have. Paid crawlers save time on large sites, but on a site of thirty pages you can do every step above by hand.
Will fixing technical SEO make me rank?
It removes the reasons you cannot. That is not the same thing. Technical work is a gate, not an engine: it lets the content compete, and then the content and your reputation decide the outcome.
What does a technical SEO checklist miss?
Everything off your site. Reviews, mentions, listings and links are what Google calls prominence, and no amount of technical work substitutes for them. A perfectly built site with no reputation still loses to an ordinary one with a strong local profile.
Is any of this different for AI search?
Less than people claim. Google states that no special file or markup is required for its AI features. What helps is the same clarity: reachable pages, a question as the heading, and the answer directly underneath in a sentence that stands on its own.
