|
| 1 | +--- |
| 2 | +title: 'blog-seo-standards' |
| 3 | +description: 'A checklist skill for writing or reviewing a technical blog post that separates editorial readability guidelines from actual 2026 SEO facts — including the current numbers for word count, title length, and meta-description length.' |
| 4 | +tagline: 'Write posts that read well, rank in search, and get quoted by AI.' |
| 5 | +difficulty: Intermediate |
| 6 | +tags: ['SEO', 'AEO', 'Content Creation', 'Technical Writing', 'Claude Code'] |
| 7 | +icon: 'search' |
| 8 | +pubDate: 2026-07-03 |
| 9 | +--- |
| 10 | + |
| 11 | +> [!NOTE] |
| 12 | +> This is the "will search *find* it, and will an AI quote it?" skill. It's the sibling of |
| 13 | +> [`technical-content-structure`](/skill/technical-content-structure) (will a human finish reading |
| 14 | +> it?) and [`tutorial-writing`](/skill/tutorial-writing) (can a |
| 15 | +> reader follow it?). Three focused checklists beat one giant one. |
| 16 | +
|
| 17 | +## What it does |
| 18 | + |
| 19 | +`blog-seo-standards` is a Claude Code [skill](https://docs.claude.com/en/docs/claude-code/skills) — |
| 20 | +a checklist for the *findability* half of a technical post: will search rank it, and can an AI answer |
| 21 | +engine (ChatGPT, Perplexity, Google AI Overviews) quote it? |
| 22 | + |
| 23 | +The thing it does best is **separate two kinds of rules that usually get blended together**: |
| 24 | + |
| 25 | +- **Editorial / readability** — craft choices (length, voice). Great for the reader, but *not* |
| 26 | + search-ranking rules. Break them on purpose when a topic needs it. |
| 27 | +- **SEO facts** — verified against Google Search Central, Ahrefs, Backlinko, Moz. Real numbers, and |
| 28 | + not always the ones a writer memorized years ago. |
| 29 | + |
| 30 | +## When to use it |
| 31 | + |
| 32 | +- You're drafting a new post and want it found by search + AI as you write. |
| 33 | +- You want a quick audit of an existing post — title length, meta description, headings, alt text, |
| 34 | + link text. |
| 35 | +- You're not sure whether a long-held writing rule is an SEO rule or just a habit. |
| 36 | + |
| 37 | +## How it works |
| 38 | + |
| 39 | +It's a plain instruction file — no code. It leads with three numbers worth getting right against |
| 40 | +current guidance: |
| 41 | + |
| 42 | +- **Post length is not an SEO rule.** Google doesn't rank on word count. It's a fine *readability* |
| 43 | + guideline — so the skill keeps it, labeled editorial, not SEO. |
| 44 | +- **Title: ~50-60 characters.** Google truncates the title link in results at ~600 px, so a longer |
| 45 | + title gets cut off. |
| 46 | +- **Meta description: ~155-160 characters** (~120 on mobile). Past that, the snippet is truncated and |
| 47 | + the rest never displays. |
| 48 | + |
| 49 | +A few ideas here generalize to any "standards" skill: say what's a fact vs. a preference (and cite |
| 50 | +sources for the facts), and correct advice that risks a penalty rather than just softening it — the |
| 51 | +old "the meta description doesn't need to be in the article" is a clickbait trap, so the skill says |
| 52 | +plainly to describe what's genuinely on the page. It also folds in what a classic checklist misses: |
| 53 | +E-E-A-T and author bylines, matching search intent, descriptive internal links (never "click here"), |
| 54 | +AEO for AI answer engines, and keeping the URL slug stable. |
| 55 | + |
| 56 | +## The skill definition |
| 57 | + |
| 58 | +Here's the actual `SKILL.md`, verbatim. It cross-references two sibling writing skills; the SEO facts |
| 59 | +and the editorial-vs-SEO split stand on their own for any technical blog. |
| 60 | + |
| 61 | +````markdown title="SKILL.md" |
| 62 | +--- |
| 63 | +name: blog-seo-standards |
| 64 | +description: >- |
| 65 | + SEO + AEO standards for writing or reviewing a technical blog post — title, meta description, |
| 66 | + headings, images, structured data, and the frontmatter that feeds them. Use when the user asks |
| 67 | + whether a post is "SEO-friendly", wants a title/description/heading/alt-text check, or is drafting |
| 68 | + an article to be found by search + AI answer engines. Separates EDITORIAL guidelines from the |
| 69 | + actual 2026 SEO numbers for word count, title length, and meta-description length. This is the SEO |
| 70 | + source of truth; for readability/structure see [[technical-content-structure]], for numbered |
| 71 | + walkthroughs [[tutorial-writing]]. |
| 72 | +--- |
| 73 | + |
| 74 | +# SEO + AEO standards for a technical blog post |
| 75 | + |
| 76 | +The findability half of writing a post: will search rank it, and can an AI answer engine (ChatGPT, |
| 77 | +Perplexity, Google AI Overviews) quote it? Pairs with [[technical-content-structure]] (readability) |
| 78 | +and [[tutorial-writing]] (numbered walkthroughs). |
| 79 | + |
| 80 | +**Two kinds of rules — keep them apart when you report:** |
| 81 | + |
| 82 | +- **Editorial** — craft choices (length, voice). Fine for UX, NOT ranking rules. |
| 83 | +- **SEO facts** — verified against Google Search Central + Ahrefs/Backlinko/Moz (2025-2026). |
| 84 | + |
| 85 | +## The numbers that matter |
| 86 | + |
| 87 | +| Field | The 2026 number | Why | |
| 88 | +| --- | --- | --- | |
| 89 | +| Post length | No SEO target | Google doesn't rank on word count — length is a *readability* choice, not a ranking factor. | |
| 90 | +| Title | ~50-60 characters | Google truncates the title link at ~600 px; longer titles get cut off in results. | |
| 91 | +| Meta description | ~155-160 chars (~120 mobile) | The snippet truncates past that; the rest never shows. | |
| 92 | + |
| 93 | +## Title (the H1 + browser/tab title) |
| 94 | + |
| 95 | +- **~50-60 characters** — the safe zone before Google's ~600 px truncation. |
| 96 | +- **Front-load the primary keyword** — the phrase a reader types. One primary + a natural secondary |
| 97 | + is plenty; **don't stuff 2-5 keywords**. Keyword-in-title is a "ticket to entry," not a boost, and |
| 98 | + stuffing reads as spam. |
| 99 | +- **Match the title to the actual format.** A conceptual headline on a hands-on walkthrough is a |
| 100 | + scope mismatch — searchers and AI won't connect it to the direct query. If the body is a |
| 101 | + walkthrough, say so in the H1; save the clever framing for the lede. |
| 102 | +- **Write for the click**, and keep every title unique. Positive framing out-clicks clickbait. |
| 103 | + |
| 104 | +## Meta description (the search + social snippet) |
| 105 | + |
| 106 | +This ONE field usually feeds the search snippet, the social card, AND the post card — it's what a |
| 107 | +person reads in Google before clicking. |
| 108 | + |
| 109 | +- **~155-160 characters** (~120 mobile). Past that it's truncated. |
| 110 | +- **It must accurately summarize the page.** "It doesn't need to be in the article" is a clickbait |
| 111 | + trap — a mismatch tanks dwell time and trust. Describe what's genuinely there. |
| 112 | +- **Phrase it as the reader's search** — state the concrete outcome, not just the topic. Google bolds |
| 113 | + matching query terms, so natural keywords help CTR. Not a ranking factor; a *click* factor. |
| 114 | +- Keywords that belong to the topic but didn't fit the title can live here — if the post covers them. |
| 115 | + |
| 116 | +## Structure & headings |
| 117 | + |
| 118 | +- **One H1 (the title), then H2 → H3, never skip a level.** Body sections start at H2 (the title is |
| 119 | + the only H1). A H3-led post skips H2 → breaks WCAG 1.3.1 + the table of contents. |
| 120 | +- **Land the primary keyword in at least one H2, not only the H1** — a common gap. Section headers |
| 121 | + double as topic signals for search and AI section-identification. |
| 122 | +- **Descriptive, standalone headings** ("Handling the timeout error"), not bare "Step 1" — readers |
| 123 | + skim headings first. |
| 124 | +- **Lead each section with the answer.** Buried answers don't win featured snippets or get quoted. |
| 125 | + |
| 126 | +## Images & code |
| 127 | + |
| 128 | +- **Explain what an image/code block shows BEFORE you show it** — inverted-pyramid; the reader knows |
| 129 | + what they're looking at and AI reads the lead-in prose for context. |
| 130 | +- **Descriptive alt text on every meaningful image** (WCAG 1.1.1 + image search) — describe what's IN |
| 131 | + it ("the settings page with the API-key field highlighted"), don't keyword-stuff; empty alt for |
| 132 | + purely decorative images. |
| 133 | +- **Optimize the file:** descriptive filename (`user-settings.png`, not `DSC_4536.png`), compressed, |
| 134 | + explicit dimensions (avoids layout shift), lazy-load offscreen images. |
| 135 | + |
| 136 | +## Modern additions (what the classic checklist missed) |
| 137 | + |
| 138 | +- **E-E-A-T — show a real author.** Google asks "is it self-evident who wrote this?" Set a per-post |
| 139 | + author byline + Article JSON-LD. First-hand experience — your own screenshots, "I hit this error |
| 140 | + and fixed it" — is what AI can't fabricate and E-E-A-T rewards. State first-party authority in one |
| 141 | + first-person sentence near the top when you built/operate the thing. |
| 142 | +- **Match search intent.** Check what already ranks for the query — an opinion piece won't win a slot |
| 143 | + the top results fill with tutorials, however clean the on-page SEO. |
| 144 | +- **Internal links, descriptive anchors.** Link related posts; anchor text that names the destination |
| 145 | + ("the setup guide"), **never "click here" / "read more"** (Google flags them; useless to screen |
| 146 | + readers). ~2-4 internal links per 1,000 words. |
| 147 | +- **AEO.** Structure for extraction: question-shaped H2s, a direct answer first, standalone quotable |
| 148 | + sentences, one extractable "X is Y that does Z" definition in the first 100 words for any named |
| 149 | + concept, and a FAQs block. |
| 150 | +- **Structured data.** Emit `Article` + `BreadcrumbList` (and `FAQPage`) JSON-LD — keep `author`, |
| 151 | + `pubDate`, `updatedDate` accurate because they feed it. (Google dropped the FAQ *rich result* in |
| 152 | + 2026, but the visible FAQ text still helps AI + non-Google engines.) |
| 153 | +- **URL slug:** short, human-readable, hyphenated, keyword-bearing. Keep it stable once published — |
| 154 | + changing it 404s the old URL and drops its link equity. |
| 155 | +- **Freshness:** for version-/time-sensitive posts, make *substantive* updates then bump the modified |
| 156 | + date — a real update, not a cosmetic date change. |
| 157 | + |
| 158 | +## How to apply |
| 159 | + |
| 160 | +- **New post:** use this while drafting — title length, description length + accuracy, keyword in an |
| 161 | + H2, explain-before-show. |
| 162 | +- **Reviewing:** read the post, then report findings grouped **Editorial** vs **SEO facts**. Flag: |
| 163 | + title > 60 chars, description > 160 or not reflecting the body, H3-led sections, empty alt, "click |
| 164 | + here" links, a missing author byline, an image/code block with no lead-in. |
| 165 | +- **Verify** any change with a build. |
| 166 | + |
| 167 | +## Sources |
| 168 | + |
| 169 | +Google Search Central (title link, snippet, helpful-content, images, links, URL-structure); Ahrefs & |
| 170 | +Backlinko on-page/title studies; Moz; Nielsen Norman Group. Verified 2026 — SEO numbers drift, so |
| 171 | +re-check before treating any figure as gospel. |
| 172 | +```` |
0 commit comments