Nine SEO checks to run before you publish a page

The problem with SEO checklists Most of them list things you can see. The real problems are the ones you can't. A missing canonical, a duplicated h1, a meta description that gets truncated — none of these break anything. The page renders, the content is good, and you find out weeks later in a report nobody reads closely. So this is a list of checks for invisible failures. The order is deliberate: structural checks first, because fixing structure usually changes the text, and re-running the content checks after a rewrite wastes the first pass. 1. Does the page have exactly one h1? Not zero, not two. Zero usually means the title was styled with a class instead of a heading element. Two usually means a template printed the site name as an h1 and the page title as another one. Both are invisible in a browser, and both make the page's main subject ambiguous. Read the heading outline rather than an issue count. If the h2 list reads like a table of contents for the page, the structure is fine. 2. Do heading levels descend without skipping? An h2 followed by an h4 tells a reader a level is missing. Sometimes a level genuinely is missing and the outline is still correct — but more often the h4 was chosen for its smaller font size. Choose levels for meaning and style them in CSS. Nothing in HTML requires an h2 to be visually larger than an h3. 3. Is the title tag specific to this page? Open ten tabs of your site and read the titles. If several are identical, those pages compete for the same query and none of them wins clearly. A title should describe what this page covers, not the category it belongs to. 4. Will the description survive truncation? The widely quoted 155-character limit is an average, not a rule. The real constraint is rendered pixel width, and mobile is meaningfully narrower than desktop. Put the sentence that must survive in the first ~120 characters and treat the rest as elaboration. Preview it rather than counting — measure in pixels, not characters. 5. Does the canonical point at itself? A page's canonical should be its own URL, in the same form the page is served at. Trailing slashes are the common case. If /tools redirects to /tools/, the canonical must say /tools/ — otherwise you're telling Google the real page is a URL that redirects. Check index pages specifically. They're the ones most often generated with a canonical that omits the trailing slash. 6. Is the page in the sitemap, in the right URL form? Same trailing-slash question. A sitemap entry that redirects is a wasted crawl. If your sitemap is generated, generate it from the same value as the canonical. Two independent string constructions will eventually disagree — and nothing will tell you when they do. // Wrong: two places that must agree, with nothing enforcing it const canonical = `${SITE}/blog/${slug}`; sitemap.push(`${SITE}/blog/${slug}/`); // ← drifts // Right: one value, used twice const url = `${SITE}/blog/${slug}/`; Enter fullscreen mode Exit fullscreen mode 7. Are the internal links real and useful? Every new page should link to at least two existing pages, and at least one existing page should link to it. A page nothing links to is a page that's hard to find. Internal links are the only navigation signal you fully control. Use descriptive anchor text. "Read the guide" tells a reader nothing about where it goes. 8. Do images have dimensions and alt text? Set explicit width and height on every image. Without them the browser can't reserve space, and the page shifts as images load — the single most common cause of a bad layout-shift score. Enter fullscreen mode Exit fullscreen mode Alt text describes the image for someone who can't see it. If the image is purely decorative, alt="" is correct — and better than a description nobody needs. 9. Does the structured data describe this page accurately? If you mark up an Article, the headline should match the visible title and the dates should be real. If you mark up an FAQPage, those questions must be visible on the page. Markup describing content that isn't there is a violation, not an optimisation. What is deliberately not on this list Keyword density. There's no target to hit, and writing to one produces worse text. Run a density check only to see whether the page is actually about the topic you intended — if the top phrases surprise you, that's a content problem, not a density problem. Word count. Length is a consequence of covering the subject, not a goal. A page that answers the question in 400 words is finished. Padding it to 1200 makes it worse. Meta keywords. Ignored for over a decade. The full version of this, with the reasoning behind each item, is on SerpPrism. The tools mentioned above — heading analyzer, SERP preview, robots.txt tester — all run client-side, nothing is uploaded. Curious whether anyone has a check I've missed. The one that keeps biting me is #5, because the failure only shows up on index pages.