SEO Checklist: The Order to Work In

An SEO checklist is a fixed list of checks you run against a site or a page so that nothing important is decided by memory. A useful one is ordered by dependency rather than by topic: confirm the page can be crawled and indexed, then that it answers the query it targets, then that the site's structure supports it, then that other sites vouch for it. Work in that order, because each layer is wasted effort if the one above it has failed — the best-written page on the internet earns nothing while a noindex tag sits in its head. Most published lists are flat runs of forty or fifty items with no ordering at all, which is why people finish one and still cannot say what to do first. This page gives the order, points each block of checks at the specialist list that covers it in full, and ends with six items Google's own documentation says you can strike off entirely.

What the list is actually for

A checklist is a memory aid for work you already know how to do. It is not a curriculum, and it is not a substitute for judgement about your own site. Its value is narrow and real: it stops a competent person skipping a step under time pressure, and it makes the same standard repeatable across pages, across people, and across months.

That narrowness matters because of how these lists are usually sold. Searchers phrase the need a dozen ways — “basic seo checklist”, “seo checklist for beginners”, “SEO to-do list” — and nearly every one of those queries returns a numbered list of tips. A list of tips is not a checklist. A checklist has a pass condition on every line: you can say, without argument, whether the item is done. If a line reads “improve your content”, it is a tip. If it reads “every indexable URL returns HTTP 200 and does not appear under Not indexed in Search Console”, it is a check. The difference is not pedantry — it is the difference between a document you can hand to someone else and one only you can interpret.

The second thing a good list does is tell you when not to run an item. Skipping steps that do not apply to your site is correct behaviour, not corner-cutting. A single-location service business does not need faceted-navigation checks; a documentation site does not need review schema; a brand-new domain has no historical performance to diagnose. A list you are not allowed to skip parts of is a list that gets abandoned in week three, because every run of it produces a pile of items marked “not applicable” and the person running it stops trusting the document.

Third, and least often stated: the list should say who owns each line. Some checks are a developer's, some are an editor's, and some are nobody's until someone is named. Items with no owner are the ones that stay unticked across every audit, quarter after quarter, and they are usually the same three items.

The four layers, and the order to work in

Every general list worth running sorts into four layers. The order below is a dependency order, not a priority order — a failure high up does not merely outrank the layers beneath it, it makes them unmeasurable. You cannot tell whether your title is working on a page Google has never fetched.

Layer 1 — Can it be crawled and indexed?

Nothing else can be evaluated until this passes, and the checks are binary and fast. The URL returns 200. robots.txt does not disallow it. There is no noindex robots meta tag and no X-Robots-Tag header. The canonical tag points at the URL you actually want indexed, not at a parameterised or paginated cousin. The URL appears in an XML sitemap that Search Console has read. Server-rendered content is present without JavaScript execution, or you have satisfied yourself that rendering completes.

Google's own starter guidance opens with essentially this instruction: before doing anything else, run a site: search for your domain and see whether your pages are there at all. If they are not, fix access before touching a word of copy. It is the cheapest check on the entire list and the one most often skipped, because it feels too simple to be the problem. It frequently is the problem.

Full detail lives in the technical SEO checklist, and the diagnostic path for a single page that refuses to index is in Google indexing issues.

Layer 2 — Does the page answer the query?

Now the content checks earn their place. One page owns one primary query. The title element and the H1 agree with each other and with that query. The opening passage answers the question without requiring a scroll, because that passage is what gets quoted into featured snippets and AI answers. The headings describe what sits under them rather than teasing it. Internal terminology matches the vocabulary the searcher used, not the vocabulary your industry uses internally.

The check that catches the most failures is the one people run last: does the page match the intent of the query, as evidenced by what currently ranks for it? If the first page of results is all comparison tables and you have written an essay, no amount of on-page work will close that gap. That is a research question rather than a writing question — see keyword search intent. The per-page version of these checks, with pass conditions attached, is the on-page content optimization checklist.

Layer 3 — Does the structure support the page?

A correct page inside an incoherent site underperforms, and the failure is invisible from the page itself. The checks here are all about relationships. The page is reachable in a small number of clicks from the homepage. It is linked from the hub that owns its subject, and it links back. No other page on the site targets the same query — competing with yourself is a structural defect that no amount of on-page work repairs. Anchor text describes the destination instead of saying “click here”. The URL sits in a directory that matches its topic, which Google notes also helps it learn how often that directory changes.

Two guides carry this layer: topic clusters for how pages should be grouped, and internal linking strategy for how they should be wired.

Layer 4 — Off page seo checklist: does anyone vouch for it?

The off-page layer is the shortest of the four and the slowest to move. It asks whether the page and the site have earned citation from elsewhere: editorial links from sites with a genuine audience; brand mentions where the subject is actually discussed; consistent business listings if location matters; and authorship attributable to a real, identifiable person with a traceable record. Nothing here can be completed in an afternoon, which is precisely why it belongs last rather than first. Teams that start here spend months acquiring links to pages that were never indexable. Start instead with links in SEO, and use the E-E-A-T checklist for the authorship and trust items.

Auditing a site that is already live

A website seo checklist run against an existing site is a diagnostic pass, not a build. Work top-down and stop at the first layer that fails, because everything below it is noise until it is repaired.

1. Coverage. Pull Search Console's Pages report. Every URL you care about should be indexed, and every indexed URL should be one you meant to publish. Both directions matter; a site with 400 indexed URLs and 120 intentional pages has a generation problem, not a content problem.
2. Duplication. One piece of content, one URL. Where that is not true, redirect or canonicalise. Note that Google describes duplicate content as inefficient rather than as something that triggers a manual action — so treat it as crawl waste to clean up, not as a penalty to panic about.
3. Ownership. Every target query belongs to exactly one page. Two pages chasing the same query split their own signals and swap positions unpredictably.
4. Answers. Each page's opening passage answers its own query within the first screen, in plain language, without a preamble about why the topic matters.
5. Links in. Every page has at least one internal link pointing at it from somewhere relevant. Pages with none are orphans, and orphans behave as though they do not exist regardless of how good they are.
6. Freshness. Perishable facts — prices, dates, product names, statistics, screenshots of interfaces that have since changed — are still true. Update them or delete the page. Changing the date without changing the content is not a refresh.

For a longer walkthrough of the diagnostic itself, see how to do an SEO audit, or work through our free SEO audit guide.

Launching something new

An seo checklist for new website builds inverts the priorities, because at launch there is no history to diagnose and no rankings to protect. The work is almost entirely Layer 1 and Layer 3.

Decide the URL structure before anything is published, because it is expensive to change afterwards and cheap to get right now. Group topically similar pages into directories rather than dumping everything at the root. Verify the property in Search Console on day one so that you have a baseline to compare against later. Generate and submit a sitemap. And — the single most common launch failure of all — make sure the staging site's noindex directives and Disallow: / rules do not travel to production. A site can sit invisible for weeks on the strength of one line in a file nobody thought to re-check after go-live.

If the launch is a replatform or a redesign rather than a genuinely new domain, the risk profile changes completely and the redirect map becomes the whole job. That work has its own list in the SEO migration checklist. For the pre-release pass on any single deployment, large or small, see the SEO QA checklist.

What runs once, what runs every publish, what recurs

Most published lists conflate three different kinds of item, which is exactly why they feel impossible to finish. Separate them and the workload becomes obvious.

Once, at setup. Search Console and analytics verified. HTTPS enforced. A canonical domain chosen and the alternatives redirected to it. URL and directory structure decided. Sitemap generated and submitted. These tick once and stay ticked, and re-auditing them quarterly is wasted motion unless something structural changed.

Every time you publish. Title and H1 agree with the target query. The answer sits in the first screen. Internal links in and out are placed. Schema validates. The URL is in the sitemap. This is the per-page list, it is the one that actually compounds, and it is the one worth automating rather than remembering.

Recurring. A monthly seo checklist should be short by design: check Search Console for new coverage errors, look for pages that lost impressions or average position against the previous period, re-verify perishable facts on your highest-value pages, and confirm no new duplicate or orphan URLs have appeared. Four items. If the recurring pass takes a full day, it is not a checklist, it is a project, and it will quietly stop happening by month three — which is worse than never having scheduled it, because you will believe it is still running.

Documented, versus widely believed

The most valuable edit you can make to an inherited list is deletion. Google's SEO Starter Guide carries a section titled “Things we believe you shouldn't focus on”, and six of the items on it still appear as line items on lists being published today. Each of these is documented — taken from Google's own wording, not inferred.

• Meta keywords tag. Google states plainly that Search “doesn't use the keywords meta tag”. That has been public since 2009 and it is still on beginner lists.
• Minimum word count. Google's guide says the length of the content alone does not matter for ranking, and that there is no magical word-count target. “Write 1,500+ words” is not a check; it is a habit.
• Heading order. Google says semantic heading order is excellent for screen readers but does not matter from a Search perspective. Keep your headings in order for accessibility — a real and sufficient reason — not for rankings.
• E-E-A-T as a ranking factor. Asked directly whether E-E-A-T is a ranking factor, Google's documentation answers no. The underlying qualities still shape how content is assessed; the acronym itself is not a dial you turn.
• Duplicate content “penalty”. Content reachable at several URLs is described as inefficient, not as something that causes a manual action. Copying someone else's content is a separate matter with separate consequences.
• Keywords in the domain name. Google says keywords in the domain or URL path alone have hardly any ranking effect beyond appearing in breadcrumbs, and that the TLD matters mainly for country targeting.

Two further claims, tagged honestly because the documentation does not exist for them. Consensus, not documented: the widely repeated idea that there is an official Google checklist to follow. Google publishes a Starter Guide and the Search Essentials; neither is structured as a checklist, and the guide opens by saying there are no secrets in it that will automatically rank a site first. Inference, not measurement: that a shorter ordered list gets completed more reliably than a long unordered one. That matches how checklists behave in other fields, but we have no published study of it for this one and are not presenting one as if we did.

Frequently asked questions

What are the five main components of SEO?

There is no Google-defined set of five. The grouping most commonly taught is technical SEO, on-page SEO, content, off-page SEO, and measurement — and that is a teaching convention rather than a documented structure, which is why different sources confidently give four, five, or six. Treat the count as a way to organise work and assign owners, not as a fact about how search engines are built.

Is there an 80/20 shortcut through this list?

Sort of, and it is not the one usually sold. The cheap binary checks in Layer 1 decide more outcomes than the entire long tail of micro-optimisations, so if you only ever run six lines, run those six. The Pareto framing itself is folklore rather than anything Google documents — where it came from and what it is actually worth is covered in SEO for beginners.

What are the top 10 SEO checklists?

Ranking them against one another is the wrong question, because they differ in scope rather than in quality: some are technical-only, some are launch-only, some are per-page, and a few are marketing collateral for a tool. The useful test is whether every line has a pass condition you can evaluate without debate, and whether the document tells you what order to work in. A list that fails both tests is not improved by being longer, and most of the popular ones are long precisely because length is easier to produce than ordering.

Can an AI assistant run these checks for you?

It can run parts of them. A language model will review copy against a target query, draft titles and meta descriptions, explain what a Search Console report means, and spot obvious intent mismatches. It cannot crawl your site, read your server responses, or see your performance data unless you hand them over — so it cannot verify Layer 1, which is the layer that decides whether the rest of the pass means anything. Treat it as a reviewer of evidence you gather, not the instrument that gathers it. Opportunity Telemetry covers where automated auditing holds up and where it does not.

Where to go next

Pick the layer that is failing rather than starting at the top of a list. If pages are missing from the index, the technical checks in full are the place to run first. If they are indexed but not ranking, the on-page and structural layers are where the work sits. If the vocabulary on this page is new to you, SEO for beginners covers what it assumes. Local businesses should add the location-specific checks in the local SEO guide; online stores should add the catalogue, variant and faceting checks in the ecommerce SEO guide, and stores on Shopify can work this same order platform-by-platform in the Shopify SEO checklist. Stores selling an imported catalogue they do not hold have a further set of checks in SEO for dropshipping.

If you would rather the per-page layer ran on every page automatically instead of by hand, that is what Precision Page Deployment does.