name: seo-optimization description: Keyword research, on-page SEO, technical audits, programmatic SEO, content optimization. Triggers on "SEO", "keywords", "ranking", "search traffic", "từ khóa", "tối ưu tìm kiếm".
SEO Optimization
Research keywords, optimize content for search engines, run technical audits, and build programmatic SEO systems. Covers on-page optimization, keyword strategy, site structure, schema markup, and Vietnam-specific search behavior.
Initial Assessment
Before any SEO task, determine:
- Task type? Keyword research, on-page optimization, technical audit, content brief, or programmatic SEO?
- Target market? Vietnam (google.com.vn), global (google.com), or both?
- Current state? New site (no authority) or existing site (has indexed pages)?
- Content language? Vietnamese, English, or bilingual? Vietnamese SEO has unique keyword patterns.
- Competitor context? Identify 3-5 competitors ranking for target terms.
- Brand context loaded? Read
docs/brand-guidelines.mdfor product, audience, positioning.
Core Methodology
Keyword Research Framework
Step 1 — Seed keywords. List 10-15 core terms from product features, customer pain points, and competitor pages.
Step 2 — Expand via intent mapping. For each seed keyword, generate variations across all four intent types:
| Intent | Signal Words | Example (EN) | Example (VI) |
|---|---|---|---|
| Informational | how, what, why, guide | "how to track expenses" | "cách theo dõi chi tiêu" |
| Navigational | brand, login, app | "[brand] pricing" | "[brand] bảng giá" |
| Commercial | best, review, compare, top | "best expense tracker app" | "app quản lý chi tiêu tốt nhất" |
| Transactional | buy, sign up, download, price | "expense tracker free download" | "tải app quản lý chi tiêu miễn phí" |
Step 3 — Prioritize. Score each keyword on a 3-point scale for relevance (to product), volume (search demand), and difficulty (competition). Prioritize: high relevance + medium volume + low difficulty.
Step 4 — Cluster. Group keywords by topic into clusters. Each cluster becomes a content piece or page. Map clusters to funnel stages: TOFU (informational), MOFU (commercial), BOFU (transactional).
On-Page Optimization Checklist
For every page or article being optimized:
Title tag: Primary keyword in first 60 chars. Benefit or number hook.
Meta description: Under 160 chars. Keyword + CTA. Vietnamese: include both có dấu and không dấu.
URL slug: Short, keyword-rich, hyphenated. Vietnamese: use không dấu (e.g., /quan-ly-chi-tieu).
H1: One per page. Contains primary keyword. Not identical to title tag.
H2-H3: Secondary keywords. Each H2 is a standalone topic.
First 100 words: Primary keyword appears naturally.
Image alt text: Descriptive, keyword-included where natural.
Internal links: 2-5 links to related content within the site.
Schema markup: Article, FAQ, Product, or HowTo — whichever matches content type.
Technical SEO Audit Framework
Run through these categories in order of impact:
1. Crawlability:
- robots.txt not blocking important pages
- XML sitemap submitted and updated
- No orphan pages (every page reachable within 3 clicks)
2. Indexability:
- No accidental noindex tags on important pages
- Canonical tags correct (self-referencing or pointing to primary)
- Duplicate content resolved via canonical or 301
3. Performance:
- Core Web Vitals: LCP < 2.5s, FID < 100ms, CLS < 0.1
- Page size under 3MB, images compressed (WebP preferred)
- Mobile-first: test at 360px width
4. Structure:
- Logical URL hierarchy (/blog/category/post-slug)
- Breadcrumbs with schema markup
- Internal linking distributes authority to priority pages
Workflow
Scenario 1: Keyword Research for New Content
- Load brand context from
docs/brand-guidelines.md - Identify the topic area and target audience
- Generate 10-15 seed keywords from product features and customer problems
- Expand using the intent mapping table — produce 40-60 keyword variations
- Score each on relevance (0-3), estimated volume (0-3), difficulty (0-3)
- Filter to top 15-20 keywords scoring 6+ combined
- Cluster into topic groups — each cluster = one content piece
- Map clusters to funnel: informational → TOFU, commercial → MOFU, transactional → BOFU
- Output keyword map as markdown table
- Save to
assets/reports/keyword-research-[topic]-YYYY-MM-DD.md
Scenario 2: Optimize Existing Content
- Identify the target page and its current primary keyword
- Run the On-Page Optimization Checklist against the page
- List all missing or suboptimal elements
- Rewrite: title tag, meta description, H1, and first paragraph with keyword integration
- Add internal links to 2-5 related pages
- Suggest schema markup type and provide JSON-LD template
- Save optimization recommendations to
assets/reports/seo-optimize-[slug].md
Scenario 3: Programmatic SEO (pSEO)
- Identify a repeatable template: one page structure × many data variations
- Examples: "[City] + [Service]", "[Tool] alternatives", "[Topic] statistics [Year]"
- Define the template structure:
- H1 pattern:
{{variable_1}} {{static_text}} {{variable_2}} - Section pattern: consistent H2s with variable data injected
- H1 pattern:
- List all variable combinations (cities, tools, topics)
- Generate 3-5 sample pages for quality review before bulk generation
- Ensure each generated page has unique value (not just variable swapping)
- Save template and variable list to
assets/reports/pseo-template-[name].md
Vietnamese SEO Specifics
Diacritics Strategy (Có Dấu vs Không Dấu)
Vietnamese users search both with and without diacritics. Google treats them as related but not identical:
Best practice:
- Primary content: always write with full diacritics (có dấu) — this is correct Vietnamese
- URL slugs: use không dấu (no accents) for cleaner URLs:
/quan-ly-chi-tieu/not/quản-lý-chi-tiêu/ - Meta description: include primary keyword có dấu, add không dấu variant if space allows
- Alt text: có dấu (Google increasingly understands Vietnamese diacritics)
Keyword research implication: Always check volume for both variants. "Quản lý chi tiêu" and "quan ly chi tieu" may show different volumes. Target the có dấu version in content, Google will match both.
Vietnamese Search Behavior
- Google dominates at 95%+ market share — Bing/Yahoo negligible in Vietnam
- Mobile-first: 70%+ of Vietnamese searches are mobile
- Voice search growing: Vietnamese voice queries tend to be longer and more conversational
- YouTube is the #2 search engine in Vietnam — video content ranks in Google VN
- Local search is critical: "gần đây" (near me), "[service] [city]" patterns are high-intent
Vietnamese Keyword Patterns
| Pattern | Example | Intent |
|---|---|---|
| "[service] là gì" | "SEO là gì" | Informational — educational |
| "[product] tốt nhất" | "phần mềm CRM tốt nhất" | Commercial — comparison |
| "[service] giá bao nhiêu" | "thiết kế website giá bao nhiêu" | Transactional — pricing |
| "[product] có tốt không" | "[brand] có tốt không" | Commercial — review |
| "cách [action]" | "cách tăng traffic website" | Informational — how-to |
| "[service] ở [city]" | "dịch vụ SEO ở HCM" | Local — service |
| "so sánh [A] và [B]" | "so sánh Shopify và Haravan" | Commercial — comparison |
Local SEO for Vietnam
- Google Business Profile: claim and optimize with Vietnamese content
- Consistent NAP (Name, Address, Phone) across all listings
- Include district and city in content: "Quận 1, TP.HCM" not just "Ho Chi Minh City"
- Register on local directories: Foody.vn (F&B), Chotot.com (marketplace), Batdongsan.com.vn (real estate)
- Reviews in Vietnamese carry more weight for local rankings
Output Specification
- Format: Markdown tables for keyword research, checklists for audits, JSON-LD for schema
- Location:
assets/reports/for research and audits - Naming:
keyword-research-[topic]-YYYY-MM-DD.md,seo-audit-YYYY-MM-DD.md,seo-optimize-[slug].md
Quality Checklist
- Keywords scored on relevance + volume + difficulty before prioritizing
- Intent mapped for every keyword (informational/commercial/transactional/navigational)
- On-page elements complete: title, meta, H1, URL, schema
- Vietnamese content has both có dấu content and không dấu URLs
- Internal linking recommendations included
- Mobile-first considerations addressed
- Brand context from docs/brand-guidelines.md applied
- Output saved to assets/reports/ with correct naming
Related Skills
content-marketing— Content creation driven by keyword researchcopywriting— Headline and meta description optimizationanalytics— Tracking keyword rankings and organic trafficcompetitor-analysis— Competitive keyword gap analysis