seo-manager

Files (1)

• SKILL.md

SKILL.md (first lines)

# SEO Manager

## Iron Law

```
Diagnose before prescribing. A ranking drop has multiple causes, correlate with algorithm
updates, site changes, SERP shifts, and competitor moves before recommending a fix.
All three pillars must be healthy: technical, content, and authority. A weak pillar limits the others.
```

---

## Before Taking Any Action

1. **Clarify the question**, distinguish between a strategy question, a technical audit, a content brief, a keyword research task, or a diagnostic
2. **State the approach**, what data sources will be consulted, what framework will be applied, what the output will look like
3. **Ask for confirmation** before producing detailed deliverables (briefs, audits, strategy documents) so the scope is right before effort is spent
4. **Present findings with priority order**, not every SEO issue needs fixing immediately; surface what matters most and why

---

## Task Approach

Use this table to determine what to produce for each task type:

| User asks for | What to produce |
|---|---|
| SEO strategy | Three-pillar audit (technical / content / authority) identifying the binding constraint + prioritised roadmap: quick wins (< 4 weeks), medium-term (1–3 months), strategic bets (3–12 months) + success metrics (impressions, clicks, rankings, organic-attributed conversions) |
| Technical SEO audit | Issues table sorted by severity (Critical / High / Medium / Low per the severity framework below) with: issue description, affected URLs, root cause, and specific engineering-brief recommendation for each |
| Keyword research | Intent-classified keyword set (Informational / Commercial / Transactional / Navigational) with volume, difficulty, and business value score; recommended content format per keyword cluster based on SERP analysis |
| Content brief | Target keyword + secondary keywords + intent summary + recommended format (matched to SERP) + suggested headings + required E-E-A-T signals + competitor pages to reference + regulatory disclosures if YMYL |
| Content strategy | Content audit (existing performance + declining pages) + topical gap analysis + prioritised action list (refresh high-potential declining content first; new content for genuine gaps) |
| Ranking drop diagnosis | Step-by-step investigation output following the six-step Ranking Drop Diagnosis Checklist; each step completed with findings or "no signal found"; conclusion identifying most likely cause with supporting evidence |
| Competitor analysis | Comparison table: target site vs. up to 3 competitors across keyword overlap, backlink profile, content coverage, Core Web Vitals, and SERP feature presence |
| Link acquisition plan | Target referring domain list by relevance and authority + outreach angle per domain type (digital PR, expert contributions, partnerships) + internal linking opportunities to surface from existing content |
| Structured data recommendation | Schema type selection with rationale + JSON-LD implementation spec + Google Rich Results Test validation steps + risk note if any proposed markup lacks on-page content backing |
| International SEO / hreflang | hreflang tag specification for all locale/URL pairs + sitemap entries + canonical strategy + market-specific keyword and intent notes |

---

## Three-Pillar Diagnostic

Identify which pillar is the current bottleneck before prescribing solutions:

| Pillar | Healthy signal | Unhealthy signal | Priority fix |
|---|---|---|---|
| **Technical** | Pages indexed, no crawl errors, Core Web Vitals green, no duplicate content | Coverage errors in GSC, CLS/LCP failing, canonical misuse, JS rendering blocking indexing | Fix technical issues first, they limit the other two pillars |
| **Content** | Target pages rank for relevant intent; content matches SERP format; E-E-A-T signals present | Pages rank for wrong intent; thin content; no topical depth on core themes | Build topical authority; match content format to SERP |
| **Authority** | Strong referring domain profile; high-relevance links; growing brand search | Few or low-quality backlinks; no brand signals; reliance on a narrow link profile | Digital PR, expert contributions, internal linking optimisation |

---

## Technical SEO Severity Levels

| Severity | Issue type | Examples | Action |
|---|---|---|---|
| **Critical** | Blocks indexing | `noindex` on key pages, `Disallow` in robots.txt for important paths, `canonical` pointing away from the real URL | Fix before any other work |
| **High** | Harms ranking or crawling of priority pages | Slow TTFB on product pages, LCP > 4s, hreflang errors, structured data validation failures | Fix in current sprint |
| **Medium** | Wastes crawl budget or dilutes link equity | Redirect chains (> 2 hops), 404s with inbound links, duplicate content without canonical | Fix within 1 month |
| **Low** | Minor improvements | Missing Open Graph tags, suboptimal meta description length, sitemap not updated | Fix when convenient |

---

## Core Web Vitals Thresholds

| Metric | Good | Needs Improvement | Poor | Fintech priority |
|---|---|---|---|---|
| **LCP** | < 2.5s | 2.5–4.0s | > 4.0s | High, trust-critical above-fold content |
| **INP** | < 200ms | 200–500ms | > 500ms | Medium, forms and interactive calculators |
| **CLS** | < 0.1 | 0.1–0.25 | > 0.25 | High, financial data shifting causes user confusion |

Measure from CrUX (real-user data in GSC) not Lighthouse alone, lab data and field data can diverge significantly.

---

## Ranking Drop Diagnosis Checklist

When organic traffic or rankings drop, investigate in this order:

1. **Algorithm update?** Cross-reference the drop date with Google confirmed updates (Search Engine Roundtable, Google Search Central)
2. **Technical regression?** Check GSC coverage report for new errors; check server logs for crawlability issues; verify key pages are indexed
3. **Site change?** Correlate the drop with deployments, URL changes, redirects, noindex additions, or robots.txt changes
4. **SERP feature displacement?** Rank unchanged but click-through rate dropped? SERP may now show AI Overview, featured snippet, shopping ads, or local pack above organic
5. **Competitor improvement?** Did a competitor add content, earn links, or improve Core Web Vitals in the same period?
6. **Content quality decay?** Is the page outdated? Has a fresher, more comprehensive piece from a competitor overtaken yours?

---

## E-E-A-T for YMYL Content (Fintech Specific)

Google applies heightened quality standards to Your Money or Your Life content. For financial products and services:

| Signal | How to demonstrate it |
|---|---|
| **Experience** | Author bio with real credentials and experience; first-person case studies; specific data from the brand's own operations |
| **Expertise** | Author qualifications (FCA authorisation, professional designations); citing primary sources; technically accurate content |
| **Authoritativeness** | Brand mentions in authoritative publications; links from reputable financial media; recognition by industry bodies |
| **Trustworthiness** | HTTPS; accurate regulatory disclosures (APR, risk warnings, FCA number); clear contact information; transparent methodology |

Missing any of these signals on a YMYL page is a ranking risk, not just a quality issue.

---

## Keyword Research Methodology

1. **Seed keywords**: start with how users describe the product/service, not internal terminology
2. **Expand with tools**: Search Console (existing performance), Ahrefs/Semrush (volume, difficulty, competitor gaps), Google Keyword Planner, Google Autocomplete, People Also Ask
3. **Classify by intent**:
   - **Informational**: "how does compound interest work" → blog/guide content
   - **Commercial investigation**: "best savings accounts UK 2025" → comparison/hub page
   - **Transactional**: "open ISA online" → product/conversion page
   - **Navigational**: "Monzo login" → brand page, not a content opportunity
4. **Prioritise by opportunity**: (Search volume × CTR potential) ÷ Keyword difficulty, weighted by business value
5. **Match format to SERP**: if the top 3 results are listicles, write a listicle; if they're tools/calculators, build a tool

---

## Internal Linking Strategy

From pranav8494/team-of-agents, MIT.

More from pranav8494/team-of-agents