lemuriaOS
seoseoanalytics

Core Web Vitals Optimization

Diagnose Core Web Vitals failures using field and lab data, then produce a developer-ready optimization spec with prioritized fixes.

Context

Use this when GSC flags CWV issues or when page speed is impacting organic performance. Core Web Vitals (LCP, CLS, INP) are a confirmed Google ranking signal. This skill produces fix specifications that developers can implement directly, bridging the gap between SEO diagnosis and engineering execution.

Procedure

  1. Pull CWV field data from GSC: identify page groups with Poor or Needs Improvement status.
  2. Run Lighthouse or PageSpeed Insights on representative pages from each failing group.
  3. Diagnose LCP issues: identify the largest contentful paint element, check server response time (TTFB), resource load chain, and render-blocking resources.
  4. Diagnose CLS issues: identify layout shift sources (images without dimensions, dynamic content injection, font loading).
  5. Diagnose INP issues: identify slow event handlers, long tasks, and main thread blocking.
  6. Write fix specifications with code-level guidance for each issue.
  7. Prioritize fixes by: traffic impact x severity x implementation effort.

Output Format

# Core Web Vitals Report: [Domain]

## Summary
| Metric | Status | Threshold | Current (p75) | Pages Affected |
|--------|--------|-----------|---------------|---------------|
| LCP | Poor/NI/Good | 2.5s | | |
| CLS | Poor/NI/Good | 0.1 | | |
| INP | Poor/NI/Good | 200ms | | |

## LCP Issues
| Page Group | LCP Element | Current | Root Cause | Fix |
|-----------|------------|---------|-----------|-----|
| | [element] | [Xs] | [TTFB/resource blocking/image size] | [Specific fix] |

## CLS Issues
| Page Group | Shift Source | CLS Value | Root Cause | Fix |
|-----------|------------|-----------|-----------|-----|
| | [element] | | [Missing dimensions/dynamic inject/font] | |

## INP Issues
| Page Group | Slow Handler | INP Value | Root Cause | Fix |
|-----------|-------------|-----------|-----------|-----|
| | [event] | [Xms] | [Long task/main thread blocking] | |

## Priority Fix List
| # | Fix | Metric | Pages | Expected Improvement | Effort |
|---|-----|--------|-------|---------------------|--------|
| 1 | | | | | L/M/H |

## Implementation Notes
[Code-level guidance for top fixes]

QA Rubric (scored)

  • Diagnostic accuracy (0-5): root causes identified at resource or code level, not just metric-level symptoms.
  • Fix specificity (0-5): developer can implement from the spec without additional research.
  • Prioritization logic (0-5): fixes ordered by traffic impact and severity.
  • Measurement plan (0-5): expected improvements are quantified and verifiable post-fix.

Examples (good/bad)

  • Good: "LCP on /blog/* pages is 4.2s (p75). Root cause: hero image is 2.4MB uncompressed PNG served without srcset. Fix: convert to WebP, add width/height attributes, implement responsive srcset. Expected LCP improvement: 4.2s → ~2.0s."
  • Bad: "Pages are slow. Optimize images and reduce JavaScript." (no specific metric, no root cause, no measurable target)

Variants

  • Quick diagnostic variant: CWV summary + top 3 fixes for the worst-performing metric only.
  • Full optimization variant: all three metrics analyzed with code-level fix specs and implementation timeline.