seo 内容
- 作者仓库星标 0
- 作者仓库 skills-registry
SEO Content — 2026 Marketing Site Writing
Search Intent Classification — Do This First
Before writing a single word, classify the intent of the target query. Getting this wrong means the page will not rank regardless of quality.
| Intent | What the user wants | Content format |
|---|---|---|
| Informational | Learn something | Long-form guide, tutorial, comparison |
| Navigational | Find a specific site/page | Not a content play — fix your brand SERP |
| Commercial investigation | Evaluate options before buying | Comparison, listicle, review |
| Transactional | Buy or sign up now | Landing page, product page |
To verify intent: Google the keyword. Look at the top 5 results. Match their format and depth. If the SERP is all listicles, write a listicle. If it is all tutorials, write a tutorial.
Title, H1, and Meta Description
Title tag:
- 50–60 characters max (Google truncates at ~580px).
- Primary keyword near the front.
- Differentiation signal: year, number, modifier ("Best", "Complete Guide", "For [ICP]").
- Example:
B2B Cold Email Templates That Actually Book Meetings (2026)
H1:
- Can differ from the title tag — use the keyword naturally, not stuffed.
- One H1 per page only.
Meta description:
- 120–158 characters.
- Not a ranking factor but directly affects CTR.
- Include primary keyword, a benefit, and an implicit CTA.
- Example:
Skip the generic templates. These cold email frameworks generated 3,000+ replies for SaaS teams in 2025. Copy and adapt.
Keyword Research Workflow
Paid tools: Ahrefs Keywords Explorer, Semrush Keyword Magic.
Free stack (in order):
- Google Search Console — find queries where you rank 8–20 and optimize for them (fastest wins).
- Bing Keyword Research Tool — free, underused, shows volume and CPC.
- AlsoAsked.com — maps the "People Also Ask" tree; reveals the sub-questions to answer in the article.
- Google autocomplete + related searches — primary keyword in the search bar, scrape suggestions.
Targeting signal: primary keyword + 3–5 secondary/LSI keywords woven naturally into H2s and body copy. Do not stuff. Google detects unnatural density.
Difficulty vs. opportunity heuristic: target keywords with KD < 30 for new sites, KD 30–60 for established domains. Above 60 requires serious authority and backlinks — a content play alone will not rank.
E-E-A-T Signals (Experience, Expertise, Authoritativeness, Trustworthiness)
Google's quality rater guidelines penalize thin, anonymous, unverifiable content. Countermeasures:
- Author byline with real name, photo, credentials, and LinkedIn link.
- First-person experience. "I tested X" or "Our team ran Y" signals real experience.
- Citations and sources. Link to primary research, studies, and authoritative domains.
- Publication date + last-updated date. Keep evergreen content updated annually at minimum.
- About page and author bio pages. Must exist and be detailed.
- Expert quotes. Pull quotes from named practitioners (with permission).
For YMYL (Your Money or Your Life) topics — finance, health, legal — E-E-A-T requirements are significantly stricter. Anonymous content will not rank.
Internal Linking
Internal links distribute PageRank and signal topical authority. Do both:
- Hub-and-spoke. Create a pillar page for a broad topic; link to it from all related cluster posts. Pillar links back to clusters.
- Contextual links. Link from body copy using descriptive anchor text (not "click here"). Anchor text should describe the destination page's topic.
- Rule of thumb: 3–5 internal links per 1,000 words.
- Audit broken internal links quarterly.
AI Overview / SGE Optimization
Google's AI Overviews (formerly SGE) cite specific sources. To be citation-worthy:
- Answer the question in the first 100 words. The model favors content that answers directly before elaborating.
- Use structured formats. Numbered steps, definition-first paragraphs, comparison tables — all easier for LLMs to parse and cite.
- Claim specificity. "Reduces churn by 18%" is citable. "Reduces churn significantly" is not.
- FAQ schema. Marks up questions and answers explicitly for AI parsing (see snippet below).
- Structured data. Article, BreadcrumbList, and FAQPage schemas increase the signal-to-noise ratio for crawlers.
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "What is a good cold email open rate?",
"acceptedAnswer": {
"@type": "Answer",
"text": "A good cold email open rate for B2B outreach is 40–60%. Below 30% usually signals deliverability problems or weak subject lines."
}
}
]
}
Add this in a <script type="application/ld+json"> block in the page <head>.
On-Page Checklist Before Publishing
- Keyword in title, H1, first 100 words, at least one H2.
- Meta description written, under 158 characters.
- Images have descriptive alt text including keyword where natural.
- Internal links to 3+ relevant pages.
- External links to at least 2 authoritative sources (opens in new tab).
- Author byline with credentials.
- FAQ schema added if the post answers specific questions.
- Core Web Vitals pass (LCP < 2.5s, CLS < 0.1, INP < 200ms) — check via PageSpeed Insights.
- Mobile rendering verified.
What Not to Do
- Do not keyword-stuff. "Best CRM software best CRM best software for CRM" will trigger a manual or algorithmic penalty.
- Do not write for bots. Write for the user who typed that query — what do they need to know?
- Do not publish thin content (< 600 words) unless the intent is genuinely a short answer.
- Do not ignore cannibalization. Two pages targeting the same keyword split authority. Merge or redirect one.
- Do not chase volume alone. A keyword with 200 monthly searches and clear transactional intent often converts better than a 10,000-search informational keyword.
<!-- tomevault:4.0:skill_md:2026-05-23 -->Source: Aldocooek/lukasdlouhy-claude-ecosystem — distributed by TomeVault.
- 流狐分类
- 写作
- 作者声明 Agent
- 未找到明确声明;不据此推断已兼容或已测试
- 静态检查
- 88 / 100 · 启发式扫描,不代表运行安全
- 作者 / 版本 / 许可
- @tomevault-io · 未声明 license
- 流狐 Token 估算
- 低消耗
- 流狐接入估算
- 即装即用
- 是否需要外部 API Key
- 未发现要求
- 检测到的系统要求
- 未声明
- 底层运行要求
- 未声明
- 检测到的文件与系统行为
-
- 只读
- 允许写入 / 修改
- 检测到的网络行为
- 仅限本地
- 安装命令数
- 无(仅作为资料)
档案由构建时根据 SKILL.md 与安装命令自动衍生,可能与作者实际意图存在差异。
需要注意: 未限定 allowed-tools,默认拥有全部工具权限。
作者没有在当前 SKILL.md 中定义固定输出样例。 Before writing a single word, classify the intent of the target query. Getting this wrong means the page will not rank regardless of quality. Intent · What the user wants · Content format Informational · Learn something · Long-form guide, tutorial, comparison
Title tag: 50–60 characters max (Google truncates at ~580px). Primary keyword near the front.
Paid tools: Ahrefs Keywords Explorer, Semrush Keyword Magic. Free stack (in order): Google Search Console — find queries where you rank 8–20 and optimize for them (fastest wins).
Google's quality rater guidelines penalize thin, anonymous, unverifiable content. Countermeasures: Author byline with real name, photo, credentials, and LinkedIn link. First-person experience. "I tested X" or "Our team ran Y" signals real experience.
Internal links distribute PageRank and signal topical authority. Do both: Hub-and-spoke. Create a pillar page for a broad topic; link to it from all related cluster posts. Pillar links back to clusters. Contextual links. Link from body copy using descriptive…
Google's AI Overviews (formerly SGE) cite specific sources. To be citation-worthy: Answer the question in the first 100 words. The model favors content that answers directly before elaborating. Use structured formats. Numbered steps, definition-first…
# SEO Content — 2026 Marketing Site Writing
## Search Intent Classification — Do This First
Before writing a single word, classify the intent of the target query. Getting this wrong means the page will not rank regardless of quality.
| Intent | What the user wants | Content format |
|--------|--------------------|-------------- |
| Informational | Learn something | Long-form guide, tutorial, comparison |
| Navigational | Find a specific site/page | Not a content play — fix your brand SERP |
| Commercial investigation | Evaluate options before buying | Comparison, listicle, review |
| Transactional | Buy or sign up now | Landing page, product page |
To verify intent: Google the keyword. Look at the top 5 results. Match their format and depth. If the SERP is all listicles, write a listicle. If it is all tutorials, write a tutorial.
## Title, H1, and Meta Description
**Title tag:**
- 50–60 characters max (Google truncates at ~580px).
- Primary keyword near the front.
- Differentiation signal: year, number, modifier ("Best", "Complete Guide", "For [ICP]").
- Example: `B2B Cold Email Templates That Actually Book Meetings (2026)`
**H1:**
- Can differ from the title tag — use the keyword naturally, not stuffed.
- One H1 per page only.
**Meta description:**
- 120–158 characters.
- Not a ranking factor but directly affects CTR.
- Include primary keyword, a benefit, and an implicit CTA.
- Example: `Skip the generic templates. These cold email frameworks generated 3,000+ replies for SaaS teams in 2025. Copy and adapt.`
## Keyword Research Workflow
Paid tools: Ahrefs Keywords Explorer, Semrush Keyword Magic.
Free stack (in order):
1. **Google Search Console** — find queries where you rank 8–20 and optimize for them (fastest wins).
… 作者原文负责流程事实;流狐只索引当前章节、要点、文件与命令。
章节 -> Search Intent Classification — Do This First → Title, H1, and Meta Description → Keyword Research Workflow → E-E-A-T Signals (Experience, Expertise, Authoritativeness, Trustworthiness) → Internal Linking → AI Overview / SGE Optimization
要点 -> Title tag · H1 · Meta description · Google Search Console · Bing Keyword Research Tool · AlsoAsked.com · Google autocomplete + related searches · Author byline
文件/命令 -> B2B Cold Email Templates That Actually Book Meetings (2026) · <script type="application/ld+json"> · <head> · site/page · secondary/LSI · application/ld · Aldocooek/lukasdlouhy-claude-ecosystem · github.com/Aldocooek/lukasdlouhy-claude-ecosystem
内容 SHA-256 -> 91fbab40c2f1
方法与流程
适用与边界
原文中的明确线索
B2B Cold Email Templates That Actually Book Meetings (2026)、<script type="application/ld+json">、<head>、site/page、secondary/LSI、application/ld、Aldocooek/lukasdlouhy-claude-ecosystem、github.com/Aldocooek/lukasdlouhy-claude-ecosystem