Root Cause Analysis Report: Production-Grade SEO & Authority Building

Jekyll + GitHub Pages Personal Site Optimization

Date: 2026-01-04 Site: https://camilo-cf.github.io/ Engineer: Staff+ Jekyll/SEO Specialist Status: INCIDENT-STYLE ROOT CAUSE ANALYSIS - Required before implementation

Executive Summary

The site has solid i18n infrastructure (EN/ES-419/PT-BR), basic SEO plugins (jekyll-seo-tag, jekyll-sitemap), and spam-safe contact patterns. However, seven critical issues prevent it from being production-grade for positioning Camilo as a reference in ML/AI/DS/MLOps/GenAI:

CRITICAL ISSUES (P0)

SEO Snippet Mismatch — Google shows outdated “Mechatronic Engineer… PhD student” instead of “Staff ML Engineer & AI Architect”
Redirect Stubs Unpolished — Legacy /cv/, /about_me/ show bare redirect pages (bad UX, questionable SEO)
Missing Authority Features — No Speaking/Media Kit, no Impact page, no lead magnets
Outdated Talks Archive — Only 2012-2017 talks visible; no “book me” CTA or current speaking credentials
Template Incoherence — Publication pages show generic academic template; doesn’t match modern AI brand
No Persona Routing — Welcome page lacks hiring/engineering/speaking persona CTAs
Meta Description Staleness — Some pages lack locale-specific descriptions; OG images missing

ISSUE 1: SEO Snippet Mismatch (Critical)

What Google Shows vs What Site Says

Current State (as of 2026-01-04):

Google SERP for homepage and /cv/: “Mechatronic Engineer focused on mechatronics, AI, and assistive robotics. PhD student of Mechatronics from UNICAMP”

Actual site description (_config.yml line 15): “Staff ML Engineer & AI Architect (Data Science Expert) leading GenAI, MLOps, and applied data science at scale

Ph.D. in Mechanical Engineering (Mechatronics) from UNICAMP

EMBA (Strategic Leadership)”

Why the Mismatch Exists:

Google Cache Staleness: The site description was recently updated (likely within the last 30-90 days based on commit history), but Google hasn’t re-crawled and re-indexed the pages yet.
Meta Description Priority: Google’s snippet selection hierarchy:
- Meta description tag (if present and relevant)
- OpenGraph description
- First paragraph of content
- Schema.org description
Current Meta Implementation:
- File: _includes/head.html line 5 uses `

Root Cause Analysis Report: Production-Grade SEO & Authority Building | Camilo Andrés Cáceres Flórez

` tag from jekyll-seo-tag plugin

File: _includes/seo.html exists but appears to be legacy (custom implementation)
File: _includes/seo_schema.html has JSON-LD Person schema with correct description
Plugin Behavior: jekyll-seo-tag generates meta description from:
- page.description (frontmatter)
- OR page.excerpt (frontmatter)
- OR site.description (_config.yml line 15)

Per-Page Frontmatter Audit:
- /en/index.md line 9: excerpt: "Production ML, causal measurement for ads, MLOps, and GenAI in one place." ✓ Good
- /en/cv/ (_pages/cv.md): No explicit description or custom excerpt; relies on site.description fallback
- /about_me/ (_pages/about_me.md): No frontmatter description; uses old narrative in content

Root Causes:

RC1: Old content indexed before recent site description update
RC2: /about_me/ page content starts with old narrative (“Mechatronic Engineer… PhD student”) which Google may prefer over site.description
RC3: No explicit per-page meta descriptions in frontmatter for critical pages (CV, About)
RC4: No active sitemap resubmission or search console ping after description update

Evidence Files

_config.yml line 15 (current description: ✓ correct)
_pages/about_me.md line 10 (starts with outdated narrative: ✗ problem)
_pages/cv.md (no explicit description: ⚠ relying on fallback)
_includes/seo_schema.html line 12 (schema description: ✓ correct, uses site.description)

Fix Approach (Priority 1)

Immediate: Add explicit description frontmatter to critical pages:
- /en/index.md (already has good excerpt ✓)
- _pages/cv.md → Add: description: "Staff ML Engineer & AI Architect with production GenAI, MLOps, and causal ads experience"
- _pages/about_me.md → Add description + rewrite first paragraph
Force Re-Crawl: Submit sitemap to Google Search Console after deploy
Add OpenGraph: Ensure jekyll-seo-tag generates og:description (verify in HTML output)
Monitor: Use Google Search Console to track snippet updates (typically 2-7 days after re-crawl)

Risk: Very low - additive only, no breaking changes Impact: High - fixes primary brand positioning issue

ISSUE 2: Redirect Stub Pages Unpolished

Current State

Redirect Mechanism:

Plugin: jekyll-redirect-from (GitHub Pages compatible, line 354 in _config.yml)
How it works: Generates HTML pages at old URLs with <meta http-equiv="refresh"> tags pointing to new URLs

Current Redirect Mappings:

/cv/ → /en/cv/ (defined in _pages/cv.md line 10)
/about_me/ → (no explicit redirect defined; page exists at /about_me/ with old content)
/contact/ → /en/contact/ (defined in _pages/contact.md line 10)
/resume/ → /en/cv/ (defined in _pages/cv.md line 11)

What Users See: When visiting /cv/ or /contact/, the jekyll-redirect-from plugin generates a bare HTML page:

<!DOCTYPE html>
<html>
<head>
  <meta http-equiv="refresh" content="0; url=/en/cv/">
  <link rel="canonical" href="/en/cv/">
</head>
<body>
  <p>Redirecting...</p>
</body>
</html>

Problems:

Bare UX: No site header, footer, or branding during redirect
SEO Ambiguity: While <link rel="canonical"> is present, the redirect page itself could be indexed if Google is slow to process the redirect
No Language Preservation: /cv/ redirects to /en/cv/ always, even if user came from Spanish or Portuguese context

Why This Exists

Historical: Site migrated from single-language structure to locale-based URLs
Good Practice: Using redirects to preserve old URLs (prevents 404s)
Plugin Limitation: jekyll-redirect-from generates minimal HTML by default; doesn’t use site layout

Root Causes

RC1: Plugin default template is intentionally minimal (not a bug, design choice)
RC2: No custom redirect template implemented to brand the interim page
RC3: No locale detection/preservation logic in redirect mappings

Evidence Files

_pages/cv.md line 9-11 (redirect_from definitions)
_pages/contact.md line 9-10 (redirect_from definitions)
Plugin docs: https://github.com/jekyll/jekyll-redirect-from (default template is minimal)

Fix Approach (Priority 2)

Option A: Custom Redirect Template (Recommended)

Create _layouts/redirect.html with:
- Site header (simplified, logo + tagline)
- “Taking you to the new location…” message
- Meta refresh + canonical (preserve existing)
- Optional: JavaScript redirect with locale detection
- Site footer (minimal)
Configure jekyll-redirect-from to use custom layout (if plugin supports, else use Option B)

Option B: Client-Side Smart Redirect (Fallback)

Create custom redirect pages manually (e.g., /cv/index.html)
Use JavaScript to:
- Detect browser language
- Redirect to matching locale version (/en/cv/, /es-419/cv/, /pt-br/cv/)
- Fallback to /en/ if no match
Show branded loading state during redirect

Option C: Noindex Redirect Pages

Keep current minimal redirect pages
Add <meta name="robots" content="noindex,follow"> to redirect pages
Ensures Google doesn’t index the redirect stubs

Recommended: Option B (smart redirect) + Option C (noindex) for best UX and SEO

Risk: Low - redirect pages aren’t critical content, failures gracefully degrade to working redirects Impact: Medium - improves polish and brand consistency

ISSUE 3: Where Meta Description / OG / Schema Are Defined

Current Implementation Architecture

Layer 1: Jekyll SEO Tag Plugin (Primary)

File: _includes/head.html line 5: `

Root Cause Analysis Report: Production-Grade SEO & Authority Building | Camilo Andrés Cáceres Flórez

Plugin: jekyll-seo-tag (line 355 in _config.yml)
What it generates:
- <title> tag
- <meta name="description">
- OpenGraph tags (og:title, og:description, og:url, og:type)
- Twitter Card tags (twitter:card, twitter:title, twitter:description)
- Canonical link (<link rel="canonical">)
Data Sources (priority order):
1. page.description or page.excerpt (frontmatter)
2. Falls back to site.description (_config.yml line 15)
3. For og:image: page.header.image or site.og_image (_config.yml line 80, currently empty)

Layer 2: Custom SEO Schema (JSON-LD)

File: _includes/seo_schema.html (included in head.html line 7)
What it generates: Person schema with:
- name, jobTitle, description (from site.author and site.description)
- sameAs links (LinkedIn, GitHub, Google Scholar)
- alumniOf, knowsAbout fields
Status: ✓ Correctly implemented, uses updated site.description

Layer 3: Legacy Custom SEO Include (Unused?)

File: _includes/seo.html (exists but NOT included in head.html)
Status: ⚠ Potentially obsolete; verify if any layouts include it directly
Contents: Custom meta tag generation (likely redundant with jekyll-seo-tag)

Layer 4: Hreflang Links

File: _includes/hreflang_links.html (included in head.html line 6)
What it generates: <link rel="alternate" hreflang="..."> tags for language variants
Status: ✓ Correctly implemented, uses translation_targets helper

Layer 5: Config Defaults

File: _config.yml
- Line 15: description (site-wide default) ✓ Updated correctly
- Line 80: og_image (currently empty) ✗ Missing
- Lines 83-90: social (Person schema links) ✓ Correct

Why It’s Stale

Problem 1: Missing Per-Page Descriptions

Most pages rely on site.description fallback
Generic site description isn’t optimized for specific page types (CV vs About vs Home)
Result: Google may ignore generic descriptions and extract from content instead

Problem 2: No OG Images

site.og_image is empty (line 80)
No per-page og:image definitions
Result: Social media shares show no preview image; Twitter cards are “summary” type instead of “summary_large_image”

Problem 3: Outdated Content Indexed

Old pages (/about_me/) have outdated narrative in content
Even with correct meta tags, Google may prefer content snippets over meta descriptions
/about_me/ page starts with “Mechatronic Engineer… PhD student” (line 10) which Google indexes

Problem 4: No Description in Locale Variants

/es-419/index.md and /pt-br/index.md may not have locale-specific descriptions
Falls back to English site.description

Evidence Files

_includes/head.html lines 5-8 (SEO stack)
_includes/seo.html (legacy, not in use)
_includes/seo_schema.html (JSON-LD, correct)
_config.yml line 80 (og_image: empty)
_pages/about_me.md line 10 (outdated content)

Fix Approach (Priority 1)

Add per-page descriptions to all critical pages:
- CV, About, Home (all locales)
- Pillar hubs, Case study index
- Format: description: "40-160 chars, focus on value + keywords"
Add OG images:
- Create default: /images/og-default.jpg (1200x630px, branded)
- Add to _config.yml: og_image: "/images/og-default.jpg"
- Optional: Per-page og:image via header.image frontmatter
Rewrite /about_me/ first paragraph:
- Current: “I’m a Staff ML Engineer… [old narrative]”
- New: Lead with current role, then progression
Audit locale-specific descriptions:
- Verify /es-419/index.md has Spanish description
- Verify /pt-br/index.md has Portuguese description

Risk: Very low - additive frontmatter + config changes Impact: High - fixes SEO snippet mismatch and social sharing

ISSUE 4: What Pages Are Being Indexed Unintentionally

Current Indexing State (Hypotheses)

Potentially Over-Indexed:

Legacy redirect stubs: /cv/, /contact/, /resume/ (if Google hasn’t processed redirects)
Old talks archive: /talks/ shows only 2012-2017 conference papers (outdated authority signal)
Old about_me narrative: /about_me/ with mechatronics-focused content (dilutes modern AI brand)
Publication pages: Individual /publications/YYYY-MM-DD-title/ pages (academic template, may not match brand)
Teaching pages: /teaching/ (if exists) with old content
Language selector pages: Root / (x-default welcome page) - should be indexed with correct hreflang
Blog filter states: /en/blog/?category=foo (if using query params, not an issue with current implementation)

Correctly Indexed (Should Stay):

/en/, /es-419/, /pt-br/ (locale homepages)
/en/pillars/, /en/case-studies/ (modern content)
/en/cv/, /en/about/ (canonical locale versions)
Blog post permalinks

How to Verify

Check Current Index:

site:camilo-cf.github.io
site:camilo-cf.github.io/talks/
site:camilo-cf.github.io/publications/
site:camilo-cf.github.io/cv/
site:camilo-cf.github.io/about_me/

Expected Results:

/cv/ should redirect to /en/cv/ or show canonical tag (may still be in index if redirect is recent)
/talks/ should be indexed BUT needs update to modern speaking page
/about_me/ should redirect to /en/about/ (verify if redirect is configured)

Root Causes

RC1: Legacy Content Not Migrated

Old talks (2012-2017) are in _talks/ collection
No modern speaking/media kit page exists
Navigation links to /talks/ (line 4 in _pages/talks.html)

RC2: No Noindex on Redirect Stubs

Redirect pages generated by jekyll-redirect-from don’t have <meta name="robots" content="noindex,follow">
Google may index them before processing redirect

RC3: About_me Not Redirecting

_pages/about_me.md has permalink: /about_me/ (line 2)
Has redirect_from: /about_me.html (line 5) but doesn’t redirect TO anywhere
Page exists with old content, no redirect to /en/about/

RC4: Publications Use Generic Template

_publications/ collection uses single.html layout
Shows academic template with comments, share buttons
Doesn’t match modern AI/ML brand narrative

Evidence Files

_pages/talks.html (shows old talks only)
_talks/ collection (2012, 2013, 2017 papers)
_pages/about_me.md (exists, no redirect, old content)
_config.yml lines 239-252 (collections config)

Fix Approach (Priority 2)

Action 1: Fix /about_me/ Redirect

# _pages/about_me.md frontmatter
redirect_to: /en/about/  # Add this
# OR delete the file and add redirect_from to _pages/about.md

Action 2: Create Modern Speaking Page

Create /en/speaking/index.md (new)
Create /es-419/charlas/index.md (new)
Create /pt-br/palestras/index.md (new)
Content: Talk proposals, bio, headshot, “book me” CTA
Update /talks/ to redirect to /en/speaking/ OR keep as archive with link to new speaking page

Action 3: Noindex Legacy Pages (Option A)

# _pages/talks.html frontmatter
title: "Talks Archive (2012-2017)"
robots: noindex,follow  # Add this

Action 3: Redirect Legacy Pages (Option B - Recommended)

# _pages/talks.html
redirect_to: /en/speaking/

Action 4: Update Navigation

Replace “Talks” with “Speaking” in navigation.yml
Point to /en/speaking/ instead of /talks/

Risk: Low - redirect/noindex changes, no content deletion Impact: Medium - improves brand coherence, removes outdated authority signals

ISSUE 5: Missing Authority-Building Features

Current State vs Required

MISSING (Must Implement):

1. Speaking / Media Kit Page

Status: ✗ Does not exist
Current: Only /talks/ archive with 2012-2017 academic papers
Required: /en/speaking/, /es-419/charlas/, /pt-br/palestras/ with:
- 4-6 talk proposals (title, abstract, audience, duration)
- Bio (short/medium/long)
- Headshot assets
- “Book me” CTA
- Recent appearances section

2. Impact / Proof of Outcomes Page

Status: ✗ Does not exist
Required: /en/impact/ (+ locales) with:
- 5 quantified outcomes (sanitized)
- Links to 2-3 case studies
- CTA: speaking, hiring, contact

3. Lead Magnet / Resources

Status: ✗ No lead magnets
Current: Blog exists but no gated content
Required: At least 1 resource page:
- Example: “Ads ML Systems at Scale – Checklist”
- Format: Ungated (trust-building) or email-gated (list-building)

4. Newsletter Page (Locale-Specific)

Status: ⚠ Partial - /newsletter/ exists but no locale versions
Files: /newsletter/index.md exists (line 266 in file listing)
Missing: /en/newsletter/, /es-419/newsletter/, /pt-br/newsletter/
Required: Locale-specific pages with:
- Value proposition above fold
- RSS link prominent
- Email signup (Substack embed OR link)
- Localized copy

5. Persona Routing on Welcome Page

Status: ✗ Not implemented
Current: _pages/home.md (root /) shows language selection only
Required: Add persona routing AFTER language selection on locale homepages:
- “Hiring” → /en/cv/ + top case studies
- “Engineering” → /en/pillars/
- “Speaking” → /en/speaking/

Why This Matters

For Hiring Managers:

Need proof of impact (quantified outcomes)
Need case studies (already exist ✓)
Need clear CV (exists ✓)

For Peers:

Need technical depth (pillars exist ✓)
Need credibility signals (publications exist ✓, but need modern framing)
Need contact/collaboration path (exists ✓)

For Conference Organizers:

CRITICAL GAP: No speaking page, no talk proposals, no bio kit
Current /talks/ shows 2012-2017 academic papers (outdated, wrong signal)
No “book me” CTA or speaking topics listed

Root Causes

RC1: Site Evolved from Academic Template

Originally based on AcademicPages Jekyll theme
Focused on publications/teaching, not modern ML leadership
Speaking = past conference papers, not forward-looking talk proposals

RC2: No Conversion Funnel Strategy

Blog exists but no lead magnets
Newsletter exists but not localized or prominently featured
No persona-based routing (hiring vs speaking vs engineering)

RC3: Privacy-First Approach (Good) Without Conversion Optimization

Contact page has spam-safe email (✓ correct)
But no scheduling link, no “book me” for speaking
No frictionless paths for different personas

Evidence Files

_pages/home.md (welcome page, no persona routing)
_pages/talks.html (archive only, no media kit)
/newsletter/index.md (exists but not localized)
Navigation.yml (no speaking link)

Fix Approach (Priority 1 - Required for Spec)

Must Implement:

Create Speaking Pages (EN/ES/PT) - See detailed spec in deliverables
Create Impact Page (EN/ES/PT) - See detailed spec in deliverables
Add Persona Routing to Locale Homepages - Not root /, but /en/, /es-419/, /pt-br/
Localize Newsletter Pages - Create /en/newsletter/, etc.
Create 1 Lead Magnet Page - Example: “Ads ML Systems Checklist”

Implementation Priority:

Week 1: Speaking pages (highest gap for conference organizers)
Week 1: Impact page (hiring managers)
Week 2: Persona routing (UX improvement)
Week 2: Newsletter localization (conversion)
Week 3: Lead magnet (optional, nice-to-have)

Risk: Low - new pages, no breaking changes Impact: Critical - unlocks conference/speaking opportunities, improves hiring positioning

ISSUE 6: Template Incoherence (Publications)

Current State

Publications Collection:

Files: _publications/*.md (academic papers)
Layout: Uses single.html layout (line 295-302 in _config.yml defaults)
Template Features:
- Author profile sidebar ✓
- Share buttons (Twitter, Facebook, LinkedIn)
- Comments (Giscus enabled, line 26-38 in _config.yml)
- Generic CTA block (“Continue the conversation”)
- Related posts (if enabled)

Problem:

Academic publications are framed with blog-style social sharing and comments
CTA block is generic (“Subscribe to newsletter, see case study”) - not contextual
Doesn’t match modern AI/ML brand narrative
Feels like academic blog posts, not professional portfolio pieces

Why This Exists

RC1: Inherited from AcademicPages Theme

Default publication layout was designed for academic blogs
Assumes readers want to comment/discuss papers
Share buttons optimized for academic Twitter

RC2: No Custom Publication Layout

Currently uses generic single.html (same as blog posts)
No publication-specific layout exists

RC3: Generic CTA on Every Page

_layouts/single.html line 76-77: `

Continue the conversation

Need a sounding board for ML, GenAI, or measurement decisions? Reach out or follow along with new playbooks.

Contact Subscribe via RSS or email See a case study

Applies to publications, blog posts, pages, case studies
Not contextually relevant

Evidence Files

_config.yml line 295-302 (publications defaults)
_layouts/single.html line 76-77 (CTA inclusion)
_includes/post_cta.html (generic CTA content)

Fix Approach (Priority 3 - Polish)

Option A: Custom Publication Layout

Create _layouts/publication.html (extends single.html)
Override:
- Remove comments section (publications don’t need discussion)
- Custom CTA: “See how this research applies to production ML” → link to related pillar
- Muted share buttons (keep LinkedIn, remove Facebook)
- Add “Citation” box (BibTeX, APA)
Update _config.yml line 296: layout: publication

Option B: Contextual CTA System

Create _includes/cta_publication.html
Update single.html to detect page.collection: ```liquid

Continue the conversation

Need a sounding board for ML, GenAI, or measurement decisions? Reach out or follow along with new playbooks.

Contact Subscribe via RSS or email See a case study

**Recommended:** Option B (contextual CTA) - more flexible, applies to talks/teaching too

**Risk:** Low - layout changes, no data migration
**Impact:** Medium - improves brand coherence, reduces CTA fatigue

---

## ISSUE 7: Cookie Banner UX (Minor)

### Current State

**Cookie Consent Implementation:**
- **File:** `_includes/cookie_banner.html`
- **Config:** `_config.yml` lines 133-141
- **Message:** "This site uses cookies… Ads and analytics stay off until you accept."
- **Current Choice Display:** Shows "Current choice: denied" (debug-like text)

**Problem:**
- "Current choice: denied" reads like debug output
- Creates compliance noise (correct behavior, but poor UX)
- Banner is functional but not polished

### Why This Exists

**RC1: Debug Text Left in Production**
- Likely added during development to verify consent state
- Forgot to remove or polish before deploy

**RC2: Compliance-First, UX-Second**
- Correctly implements GDPR/privacy requirements ✓
- But UX polish is lacking

### Evidence Files
- `_includes/cookie_banner.html` (check for "Current choice" text)
- `assets/js/consent.js` (if exists, verify it's updating banner state)

### Fix Approach (Priority 4 - Nice-to-Have)

**Quick Fix:**
1. Remove "Current choice: denied" text from banner
2. Show banner only if consent not yet given (hide after accept/reject)
3. Optional: Add small status indicator in footer ("Privacy settings") instead of inline banner text

**Better Fix:**
1. Move consent status to footer: "Privacy settings: Minimal cookies | Change"
2. Hide banner entirely after user makes choice
3. Allow re-opening banner via footer link

**Risk:** Very low - UX polish only
**Impact:** Low - minor polish, doesn't affect functionality

---

## Summary: SEO & Metadata Architecture

### Where Things Are Defined (Map)

| Component | File | Status | Priority to Fix |
|-----------|------|--------|-----------------|
| Site-wide description | `_config.yml` line 15 | ✓ Correct | - |
| Per-page description | Page frontmatter `description:` | ⚠ Missing on most pages | P1 |
| OG image (default) | `_config.yml` line 80 | ✗ Empty | P1 |
| OG image (per-page) | Page frontmatter `header.image:` | ⚠ Not used | P2 |
| Meta tags (auto) | `<!-- Begin Jekyll SEO tag v2.8.0 -->
<title>Root Cause Analysis Report: Production-Grade SEO &amp; Authority Building | Camilo Andrés Cáceres Flórez</title>
<meta name="generator" content="Jekyll v3.10.0" />
<meta property="og:title" content="Root Cause Analysis Report: Production-Grade SEO &amp; Authority Building" />
<meta name="author" content="Camilo Andrés Cáceres Flórez" />
<meta property="og:locale" content="en_US" />
<meta name="description" content="Staff ML Engineer &amp; AI Architect (Data Science Expert) leading GenAI, MLOps, and applied data science at scale Ph.D. in Mechanical Engineering (Mechatronics) from UNICAMP EMBA (Strategic Leadership)" />
<meta property="og:description" content="Staff ML Engineer &amp; AI Architect (Data Science Expert) leading GenAI, MLOps, and applied data science at scale Ph.D. in Mechanical Engineering (Mechatronics) from UNICAMP EMBA (Strategic Leadership)" />
<link rel="canonical" href="https://camilo-cf.github.io/docs/root-cause-report/" />
<meta property="og:url" content="https://camilo-cf.github.io/docs/root-cause-report/" />
<meta property="og:site_name" content="Camilo Andrés Cáceres Flórez" />
<meta property="og:type" content="website" />
<meta name="twitter:card" content="summary" />
<meta property="twitter:title" content="Root Cause Analysis Report: Production-Grade SEO &amp; Authority Building" />
<meta name="twitter:site" content="@camilocaceresf" />
<meta name="twitter:creator" content="@Camilo Andrés Cáceres Flórez" />
<script type="application/ld+json">
{"@context":"https://schema.org","@type":"WebPage","author":{"@type":"Person","name":"Camilo Andrés Cáceres Flórez"},"description":"Staff ML Engineer &amp; AI Architect (Data Science Expert) leading GenAI, MLOps, and applied data science at scale Ph.D. in Mechanical Engineering (Mechatronics) from UNICAMP EMBA (Strategic Leadership)","headline":"Root Cause Analysis Report: Production-Grade SEO &amp; Authority Building","publisher":{"@type":"Organization","logo":{"@type":"ImageObject","url":"https://camilo-cf.github.io/images/Foto_Camilo_perfil.jpg"},"name":"Camilo Andrés Cáceres Flórez"},"url":"https://camilo-cf.github.io/docs/root-cause-report/"}</script>
<!-- End Jekyll SEO tag -->
` in head.html | ✓ Working | - |
| JSON-LD schema | `_includes/seo_schema.html` | ✓ Correct | - |
| Hreflang tags | `_includes/hreflang_links.html` | ✓ Correct | - |
| Canonical tags | `jekyll-seo-tag` (auto) | ✓ Working | - |
| Sitemap | `jekyll-sitemap` plugin | ✓ Generates /sitemap.xml | - |
| Robots.txt | `/robots.txt` | ✓ Exists | - |
| Redirect stubs | `jekyll-redirect-from` | ⚠ Unpolished | P2 |
| Legacy seo.html | `_includes/seo.html` | ⚠ Unused? | P3 (verify/delete) |

### What's Being Indexed (Hypothesis)

**Should Be Indexed:**
- ✓ `/en/`, `/es-419/`, `/pt-br/` (locale homes)
- ✓ `/en/cv/`, `/en/about/`, `/en/pillars/`, `/en/case-studies/`
- ✓ Blog posts (`/en/blog/YYYY/MM/DD/title/`)
- ✓ Root `/` (x-default with hreflang)

**Should NOT Be Indexed (Fix Required):**
- ✗ `/cv/`, `/contact/`, `/resume/` (redirect stubs) → Add noindex
- ✗ `/about_me/` (old content) → Redirect to `/en/about/`

**Needs Update (Content Refresh):**
- ⚠ `/talks/` (2012-2017 archive) → Redirect to `/en/speaking/` OR add "Archive" notice + link to speaking page
- ⚠ Individual publication pages → Polish template, contextual CTA

---

## Implementation Roadmap

### PHASE 0: Root Cause Complete ✓
- This document

### PHASE 1: Critical SEO Fixes (Week 1)
**PR #1: SEO Snippet Mismatch + Meta Descriptions**
1. Add `description` frontmatter to:
   - `_pages/cv.md`
   - `_pages/about.md`
   - `/en/index.md` (already has excerpt, verify)
   - `/es-419/index.md`, `/pt-br/index.md`
2. Rewrite `_pages/about_me.md` first paragraph OR redirect to `/en/about/`
3. Add OG default image: `/images/og-default.jpg`
4. Update `_config.yml` line 80: `og_image: "/images/og-default.jpg"`
5. Submit sitemap to Google Search Console

**PR #2: Redirect Polish + Noindex**
1. Add noindex to redirect stubs (custom redirect layout OR manual)
2. Fix `/about_me/` redirect
3. Test all legacy URLs redirect correctly

### PHASE 2: Authority Features (Week 2)
**PR #3: Speaking / Media Kit Pages**
1. Create `/en/speaking/index.md` with:
   - 4-6 talk proposals
   - Bio (short/medium/long)
   - Headshot link
   - "Book me" CTA
2. Create `/es-419/charlas/index.md` (localized)
3. Create `/pt-br/palestras/index.md` (localized)
4. Update navigation.yml: Replace "Talks" with "Speaking"
5. Update `/talks/` to redirect OR show "Archive (2012-2017) | See current speaking topics"

**PR #4: Impact Page**
1. Create `/en/impact/index.md` with:
   - 5 quantified outcomes
   - Links to case studies
   - CTA: speaking + hiring + contact
2. Create `/es-419/impacto/index.md` (localized)
3. Create `/pt-br/impacto/index.md` (localized)

### PHASE 3: Conversion Optimization (Week 3)
**PR #5: Persona Routing + Newsletter**
1. Add persona routing to `/en/index.md`, `/es-419/index.md`, `/pt-br/index.md`:
   - Hiring → CV + case studies
   - Engineering → Pillars
   - Speaking → Speaking page
2. Create `/en/newsletter/index.md` (localized)
3. Create `/es-419/newsletter/index.md` (localized)
4. Create `/pt-br/newsletter/index.md` (localized)

**PR #6: Lead Magnet (Optional)**
1. Create `/en/resources/ads-ml-checklist/index.md`
2. Add to navigation or pillar CTAs

### PHASE 4: Template Polish (Week 4)
**PR #7: Contextual CTAs**
1. Create `_includes/cta_publication.html`
2. Create `_includes/cta_speaking.html`
3. Create `_includes/cta_case_study.html`
4. Update `_layouts/single.html` to use contextual CTAs

**PR #8: Cookie Banner Polish**
1. Remove debug text
2. Add footer privacy settings link

### PHASE 5: CI & Testing (Week 5)
**PR #9: Playwright Tests**
1. Setup Playwright
2. Add quality gate tests:
   - Welcome page has persona routing ✗ (after PR #5)
   - `/en/speaking/` exists with ≥3 talk proposals ✗ (after PR #3)
   - `/en/contact/` email not in HTML ✓ (already implemented, verify)
   - `/robots.txt` exists ✓
   - `/sitemap.xml` exists ✓
3. Add to `.github/workflows/ci.yml`

**PR #10: README + Docs**
1. Update README with:
   - Local dev instructions
   - How to add new talk/post in all locales
   - Cookie consent gating for embeds
2. Document persona routing customization

---

## Appendix: File Locations Reference

### Critical Files for SEO

_config.yml # Site description (line 15), og_image (line 80) _includes/head.html # SEO stack (seo tag, hreflang, schema) _includes/seo_schema.html # JSON-LD Person schema _includes/hreflang_links.html # Language variant tags _includes/seo.html # Legacy (verify if used)

_pages/cv.md # Add description frontmatter _pages/about.md # Add description frontmatter _pages/about_me.md # Redirect to /en/about/ OR rewrite

/en/index.md # Verify excerpt/description /es-419/index.md # Add description /pt-br/index.md # Add description

/robots.txt # ✓ Exists /sitemap.xml # ✓ Auto-generated by jekyll-sitemap


### Files to Create

/en/speaking/index.md # Media kit /es-419/charlas/index.md # Localized /pt-br/palestras/index.md # Localized

/en/impact/index.md # Proof of outcomes /es-419/impacto/index.md # Localized /pt-br/impacto/index.md # Localized

/en/newsletter/index.md # Conversion /es-419/newsletter/index.md # Localized /pt-br/newsletter/index.md # Localized

/images/og-default.jpg # 1200x630px OpenGraph image

_layouts/redirect.html # Custom redirect template (optional) _includes/cta_publication.html # Contextual CTA _includes/cta_speaking.html # Contextual CTA _includes/cta_case_study.html # Contextual CTA

.github/workflows/ci.yml # CI with Playwright tests/smoke.spec.js # Playwright tests playwright.config.js # Playwright config ```

END OF ROOT CAUSE REPORT

Next Steps:

Review this report with stakeholder (Camilo)
Prioritize PRs based on business goals (conference speaking = prioritize PR #3)
Begin Phase 1 implementation
Measure: Google Search Console snippet updates, social share previews, speaking inquiries

Approval Required Before Proceeding to Implementation.

Continue the conversation

Need a sounding board for ML, GenAI, or measurement decisions? Reach out or follow along with new playbooks.

Contact Subscribe via RSS or email See a case study