What most SEO checklists miss
TL;DR
Most SEO checklists cover the same categories. What gets left out is a small set of items that change how the rest of the list reads: field data over lab data on Core Web Vitals, bidirectional hreflang, the cleanup layer, and the items that have shifted since 2022.
Most popular SEO checklists are built around the same categories. Setup, technical SEO, on-page, links, content, performance. Compare the Backlinko SEO checklist and the Ahrefs SEO checklist against any other and the categories repeat. The basics are well covered.
What gets left out, on each one, is a small set of items that change what the rest of the checklist means. Whether your Largest Contentful Paint matters depends on which dataset Google is actually reading. Hreflang annotations only do something if the page on the other end of the reference points back. The "advanced" optimisation work that fills most case studies tends to sit on cleanup items that don't make it into the bullet list at all.
Here are the four most common gaps.
Field data is what Google reads
PageSpeed Insights opens with a Lighthouse score and a coloured ring around it. That's lab data: a single controlled run in a defined environment so the result is reproducible. The number Google ranks on comes from elsewhere. Core Web Vitals are assessed from real Chrome users in the Chrome User Experience Report (CrUX), and the web.dev field-versus-lab guidance is explicit that field measurement reflects the actual devices, networks, and conditions of your visitors.
The Chrome UX Report documentation states that CrUX data is "used by Google Search to inform the page experience ranking factor". Search Engine Journal's summary of the same guidance reaches the same conclusion: focus on field data.
The practical effect is that the Lighthouse number near the top of PageSpeed Insights can pass while the field section below it fails, and vice versa. The field section is empty on sites without enough traffic to populate CrUX, which is why low-volume pages show no field assessment at all. The CrUX Dashboard in Looker Studio reports the field data over time and is the version that maps to what Google reads.
Most checklists say "pass Core Web Vitals" without naming which dataset that means. The field section is the one to focus on.
Hreflang is a relationship between pages
The hreflang annotation isn't a single line you drop on a page. It's a set of bidirectional references. Google's localised-versions reference puts the rule directly: if two pages don't both point to each other, the tags will be ignored. Each language version must list itself as well as every other language version.
The most common failure mode on multilingual sites isn't the absence of the tag. It's a cluster where the UK page references the US page but the US page references only itself, or where a recent locale was added without back-references from the existing set. The annotation collapses silently and Search Console reports nothing useful unless the property has language targeting set per directory.
A site with twenty language and region variants has roughly four hundred bidirectional reference relationships to keep in sync, plus the self-reference on each page. A checklist that says "add hreflang tags" without saying "validate that every reference is reciprocal" is listing the easy half.
Cleanup is the work
Most checklists organise themselves around adding things. Add a sitemap. Add structured data. Add a meta description on every page. The set of items that gets left out is the cleanup half. Untangling old redirects, removing duplicate canonical signals, fixing conversion tracking that's been miscounting for the past six months. None of it looks like "doing SEO" in the way the checklist genre frames the work, and all of it is where the gains tend to sit.
Google's redirect guidance recommends keeping any chain to fewer than five hops, ideally no more than three, and notes that Googlebot stops following after ten. Three or four-step chains accumulate across years of redesigns and platform migrations without anyone noticing, and the cost is in user-perceived latency and crawl efficiency rather than the PageRank dilution older guides used to warn about (Google has been clear that 301 redirects don't lose PageRank).
Conversion tracking is the sharpest example, because audit data covers it directly. GrowthSpree's audit of B2B SaaS Google Ads accounts found that 70 to 80 per cent had at least one significant tracking issue, with about 60 per cent firing multiple conversion events under a single primary goal. Sarah Stemen's practitioner observations from her own audit work put a similar pattern at around 80 per cent of accounts she sees. Why measurement layered on broken tracking doesn't rescue the data is the subject of a separate post on enhanced conversions.
The reason this matters for an SEO checklist is that everything else on it rests on the cleanup. Schema markup on a page blocked by the wrong canonical tag does nothing. A meta description on a URL that 404s isn't read by anyone. The rest of the list assumes a foundation that's actually there.
The items that have shifted since 2022
The shape of most public SEO checklists hasn't changed much in three years. The shape of search has. Three items in particular don't tend to appear in lists written before late 2024, and where they do appear, they show up as one-line bullets.
IndexNow. An open protocol that pushes content changes to participating search engines instead of waiting for the next crawl. It has been live since 2021. Wix, WordPress.com, and Duda include it natively. Cloudflare exposes it as a single Crawler Hints toggle. Google doesn't currently support it; Bing, Yandex, Naver, Seznam, and Yep do, and that covers the surfaces feeding most non-Google answer engines. The IndexNow documentation lists the current participants.
Agent-readable interfaces. Autonomous browsing agents (ChatGPT Operator, Gemini Agents, Claude Computer Use) parse pages through the rendered DOM and the accessibility tree. Buttons built from styled div elements, transparent overlays sitting on top of controls, and layouts that shift between similar pages all interfere with agent workflows. The same patterns help screen readers and traditional crawlers. We covered the read-side foundation and what WebMCP changes in a separate post on agent-readiness.
Generative engine surfaces. AI Overviews, ChatGPT search, Perplexity. The optimisation work is closer to thorough quality-focused SEO than most GEO content suggests, but it isn't a zero on the checklist either. The post on generative engine optimisation covers the overlap with traditional SEO and where it diverges.
The four items above all sit in the free SEO checklist under the technical, performance, and international sections. The categories there are the same as every other list. The interpretation around each item is what differs.
What the pattern is
The items most checklists leave out aren't exotic. They're the items that change how the rest of the list reads.