Technical SEO Checklist: What Actually Ranks You

Checklists rot. Google’s own numbers don’t.

A technical SEO checklist written in 2021 still tells you HTTPS is a ranking signal on its own. It isn’t, and Google’s own documentation says so directly. Another one still repeats a mobile-abandonment statistic that stopped holding up under scrutiny years ago. It gets copied from post to post anyway, long after anyone checked where it came from.

Checklists rot for a boring reason. A threshold gets published and gets copied forty times. Nobody goes back to the primary source once Google updates the number. The list keeps its structure. The plumbing it describes moves on without it.

This one works differently in two ways. Every threshold below traces to a live Google or web.dev page, dated, with the source attached. If Google changes a number, this page is wrong until it gets fixed. The checklist genre at large stays quietly wrong forever. The checks are also ranked by what actually blocks rankings, not padded out to a round number of tips. The close explains where the job stops entirely: technical SEO removes blockers. It doesn’t build authority. Fixing the plumbing is necessary. It was never sufficient.

Tier 1: crawl and index

If Google can’t crawl a page or won’t index it, nothing else on this list matters yet. Start here.

• Robots.txt allows what should rank. Check the file for an accidental Disallow on a production path. Good looks like only admin panels, staging environments, and internal search results blocked. Nothing a buyer would ever need to find. A blocked page never gets crawled, so it never ranks, no matter how strong the content underneath it is.

• The XML sitemap matches the live site. Check it against the actual published URL set, not the one from launch day. Good looks like no 404s or redirected URLs sitting inside it. For most sites, this single habit is the whole crawl-budget conversation. Google’s own guidance says budget only becomes a real concern above 1 million pages updating weekly. The same holds above 10,000 pages updating daily. Below that, keeping the sitemap current and checking index coverage regularly is adequate, in Google’s own words.

• Canonical tags point to one URL, not several. Check for a self-referencing canonical on unique pages and a single clear canonical on any duplicate or parameter-driven variant. Good looks like one canonical destination per piece of content, no competing canonical loops. Two pages competing for the same query split the same authority two ways.

• Redirects resolve in one hop. Check for chains three, four, five hops deep. They accumulate quietly over years of URL changes. Good looks like a straight 301 from the old URL directly to the current one. Chains slow crawlers down and leak signal at every hop.

• The index coverage report gets checked on a cadence, not just after a drop. Check Search Console’s Page Indexing report for a rising “Crawled, currently not indexed” count. Good looks like a stable pattern month over month, reviewed on a schedule. Not opened for the first time during a panic.

• Internal links reach every page worth finding. Check for orphan pages with zero internal inbound links and for links pointing at dead URLs. Good looks like every published page reachable within a few clicks of the homepage. No live link aimed at a 404. Strong content with nothing linking to it is invisible to a crawler and to a reader.

Tier 2: speed and stability

Google measures three specific things about how a page feels to use, each with a published threshold. Each one is measured at the 75th percentile of real page loads, not a lab test.

• Largest Contentful Paint (LCP) is good at 2.5 seconds or under. Needs Improvement runs 2.5 to 4.0 seconds. Poor is anything over 4.0 seconds. LCP measures how long the biggest visible element takes to load, the thing a visitor is actually waiting on.

• Interaction to Next Paint (INP) is good at 200 milliseconds or under. Needs Improvement runs 200 to 500 milliseconds. Poor is over 500. INP captures the full click-to-response time across a visit, not just the first tap.

• Cumulative Layout Shift (CLS) is good at 0.1 or under. Needs Improvement runs 0.1 to 0.25. Poor is over 0.25. CLS catches the page that jumps while someone is reading it. An ad loads late. A button moves right as a thumb goes to tap it.

Speed doesn’t just move a Core Web Vitals report. It moved revenue directly in the one enterprise-scale study anyone has run at this. Google and Deloitte measured 37 brand sites and more than 30 million sessions in 2019. A 0.1-second mobile speed improvement lifted retail conversion by 8.4% and travel conversion by 10.1%.

Speed and conversion, 2019 enterprise benchmark

Google and Deloitte, Milliseconds Make Millions, 2019

Read that chart for what it is. It’s 2019 data from large consumer brands, and it measures conversion, not search rankings directly. Treat it as a floor under why speed is worth fixing before Google’s algorithm ever enters the conversation.

Tier 3: structure and meaning

• HTTPS is served everywhere, with no mixed content. Check every template, not just the homepage, for an HTTP resource loading inside an HTTPS page. Google’s own page-experience documentation asks whether pages are “served in a secure fashion” as one input among several, and says plainly that beyond Core Web Vitals, other page experience aspects don’t directly help a site rank higher. HTTPS is table stakes now, not a lever.

• The site is built for the mobile version Google actually indexes. Check that content, structured data, and internal links on mobile match desktop exactly. Mobile-first indexing means Google uses the mobile version, crawled with a smartphone user agent. That’s the version used for indexing and ranking, per Google’s current documentation. Full stop. A feature that only exists on desktop is invisible to Google if it never made it to mobile.

• Structured data validates clean, with zero errors. Check every template type in a rich-results testing tool: article, FAQ, local business, product, wherever each applies. Structured data doesn’t guarantee a ranking, but it can move click-through rate hard. Google’s own documentation cites Rotten Tomatoes adding structured data to 100,000 unique pages and measuring a 25% higher click-through rate.

• Content gets reviewed after every core update, not mid-panic. Check top pages against the intent and quality bar of the query once a named update finishes rolling out. The March 2026 core update ran from March 27 to April 8, about 12 days, and Google describes these as regular updates “designed to better surface relevant, satisfying content for searchers from all types of sites.” Reacting mid-rollout to a metric that hasn’t settled yet usually means fixing the wrong thing.

Site migration: the highest-risk technical SEO event

A domain change, a URL restructure, a platform move. Any of these breaks more of the checklist above at once than anything else does. This is the compact version.

• Every old URL maps to its closest live equivalent, one to one. Check the redirect map before launch, not after. Good looks like no shortcut redirecting every old URL to the homepage. Google’s guidance names a specific hold time: keep redirects “for as long as possible, generally at least 1 year.”

• Ranking movement during the transition gets read as expected, not as failure. Google says plainly that “with any significant change to a site, you may experience ranking fluctuations while Google recrawls and reindexes your site.” Reverting three days in usually costs more than the fluctuation would have.

• The recovery timeline gets set before launch, not guessed at afterward. In Google’s own words, “a medium-sized website can take a few weeks for most pages to move in our index, larger sites can take longer.” No credible source publishes a specific traffic-loss percentage for a migration. Any checklist quoting one invented it.

Local technical SEO, the compact version

Local ranking runs on inputs a national checklist doesn’t touch. Compact version now, more depth is coming as its own guide.

• Google Business Profile inputs are set for what Google actually weighs. Google names three factors directly: relevance (“how well a Business Profile matches what someone is searching for”), distance, and prominence (“how well-known a business is”). Check that category, services list, and description language match the terms a buyer actually searches. Citations and reviews should stay current too.

• LocalBusiness structured data lives on every location page. Check it validates the same way the sitewide structured data check above does. This is the local variant of that Tier 3 item, not a separate discipline.

• Name, address, and phone match across every listing, checked on a schedule. These details drift quietly across directories over time. It’s a technical check, not a content one: a mismatch confuses the same systems that read the structured data above.

What this checklist cannot do for you

Every item above removes a blocker. None of them make a firm the one an answer engine reaches for.

In our own study of AI answers to legal queries, a firm’s own site was the cited source 67% of the time, ahead of directories and news mentions combined. That’s 236 citations across 26 queries in 8 metros. That number rewards firms that publish something worth citing. A perfect Core Web Vitals score doesn’t put an original finding on the page. Clean structured data doesn’t get a firm cited when an AI system needs a number nobody else has.

Technical health and being the cited brand are two different jobs, related but not interchangeable. GEO vs SEO covers how the second job actually works. What it takes to get cited by AI covers what happens once the plumbing is clear.

Run this list against your own domain and something on it is wrong. Almost every site has at least one real defect sitting on this list unaddressed. A checklist tells you what to check. It doesn’t check it. Get a graded audit of your own site, ranked worst-first, before the next fix.

References

1. web.dev. "Largest Contentful Paint (LCP)." Updated September 2025.
2. web.dev. "Interaction to Next Paint (INP)." Updated September 2025.
3. web.dev. "Cumulative Layout Shift (CLS)."
4. web.dev (Google and Deloitte). "Milliseconds Make Millions." Case study, 2019 data across 37 brand sites and 30M+ sessions.
5. Google Search Central. "Site moves and migrations." Updated June 2026.
6. Google Search Central. "Large site owner's guide to managing your crawl budget." Updated December 2025.
7. Google Search Central. "Mobile-first indexing." Updated December 2025.
8. Google Search Central. "Understanding page experience in Google Search results."
9. Google Business Profile Help. "Tips to improve your local ranking on Google."
10. Google Search Central. "Intro to how structured data markup works." Updated December 2025.
11. Google Search Status Dashboard. "March 2026 core update."