claude-plugin-wordpress-manager 2.4.0 → 2.6.0

package/skills/wp-content-optimization/references/meta-optimization.md

@@ -0,0 +1,243 @@
+# Meta Description and Title Tag Optimization
+
+## Overview
+
+Title tags and meta descriptions are the first impression users see in search results. Optimizing these elements directly impacts click-through rate (CTR). Claude generates optimized meta elements by combining best practices with GSC CTR data, then produces A/B variants for testing. This process uses existing WP REST Bridge and GSC tools — no additional APIs needed.
+
+## Title Tag Best Practices
+
+### Rules
+
+| Rule | Standard |
+|------|----------|
+| Maximum length | 60 characters (Google truncates at ~60) |
+| Keyword placement | Primary keyword in the first 3-4 words |
+| Uniqueness | Every page must have a unique title |
+| Brand inclusion | Optional, at end with pipe separator |
+| Keyword repetition | Keyword appears once only |
+| Readability | Must make sense as a standalone phrase |
+
+### Title Tag Formula
+```
+[Primary Keyword] — [Benefit or Differentiator] | [Brand]
+```
+
+### Character Count Optimization
+```
+Ideal:   50-60 characters (full display in all search results)
+OK:      40-50 characters (may look short but fully visible)
+Warning: 60-65 characters (partial truncation on some devices)
+Error:   65+ characters (truncated, meaning lost)
+Too Short: <30 characters (underutilizing title space)
+```
+
+### Examples
+```
+Good (55 chars): "WordPress Speed Optimization — 10 Proven Techniques | Brand"
+Good (48 chars): "How to Speed Up WordPress in 2026 | Brand"
+Bad (72 chars):  "The Complete and Ultimate Guide to WordPress Speed Optimization Tips 2026"
+Bad (18 chars):  "Speed Tips"
+```
+
+## Meta Description Guidelines
+
+### Rules
+
+| Rule | Standard |
+|------|----------|
+| Maximum length | 155-160 characters |
+| Minimum length | 120 characters (shorter looks incomplete) |
+| Keyword inclusion | Primary keyword once, naturally |
+| Call to action | Include an action verb (learn, discover, get, try) |
+| Uniqueness | Every page must have a unique description |
+| Emotional trigger | Include benefit or pain point |
+| Specificity | Use numbers, data, or specific outcomes |
+
+### Meta Description Formula
+```
+[Action verb] [what the reader gets/learns]. [Specific detail or data point]. [CTA].
+```
+
+### Examples
+```
+Good (152 chars):
+"Learn 10 proven WordPress speed optimization techniques that cut load time
+by 50%. Step-by-step guide with benchmarks. Start optimizing today."
+
+Bad (89 chars):
+"This article is about WordPress speed optimization. Read more on our website."
+
+Bad (185 chars):
+"In this comprehensive and detailed guide we will cover everything you need to
+know about optimizing your WordPress website for speed including caching and
+image optimization techniques."
+```
+
+## Using GSC CTR Data for Optimization
+
+### Identifying Low-CTR Pages
+
+#### Step 1: Pull Page Performance Data
+```
+Use gsc_page_performance to get CTR data for all pages.
+Sort by impressions (descending) to focus on high-visibility pages first.
+```
+
+#### Step 2: CTR Benchmarks by Position
+
+| Average Position | Expected CTR | Below Average |
+|-----------------|-------------|---------------|
+| 1 | 25-35% | <20% |
+| 2 | 12-18% | <10% |
+| 3 | 8-12% | <6% |
+| 4-5 | 5-8% | <4% |
+| 6-10 | 2-5% | <2% |
+
+#### Step 3: Flag Optimization Candidates
+Pages with high impressions but below-average CTR are prime candidates:
+```
+Example:
+Page: /wordpress-speed-guide
+Position: 3
+Impressions: 5,000/month
+CTR: 4.2% (expected 8-12% for position 3)
+→ META OPTIMIZATION NEEDED — potential to double clicks
+```
+
+#### Step 4: Analyze Current Meta
+```
+Use get_content to fetch the current title tag and meta description.
+Evaluate against the rules above.
+Identify specific issues (too long, missing keyword, no CTA, etc.).
+```
+
+### CTR Impact Estimation
+```
+Current: Position 3, 5000 impressions, 4.2% CTR = 210 clicks/month
+After optimization to 8% CTR: 5000 × 0.08 = 400 clicks/month
+Potential gain: +190 clicks/month (+90% improvement)
+```
+
+## A/B Variant Generation
+
+### Process
+
+#### Step 1: Generate Two Variants
+Claude generates two meta description variants using different approaches:
+
+**Variant A: Benefit-focused**
+```
+"Cut your WordPress load time by 50% with these 10 proven optimization
+techniques. Includes caching, image compression, and CDN setup. Free guide."
+```
+
+**Variant B: Problem-focused**
+```
+"Slow WordPress site losing visitors? Fix it with 10 tested speed
+techniques. Average improvement: 50% faster. Step-by-step instructions."
+```
+
+#### Step 2: Score Both Variants
+
+| Criteria | Variant A | Variant B |
+|----------|-----------|-----------|
+| Keyword presence | Yes (1st sentence) | Yes (1st sentence) |
+| Length | 148 chars ✓ | 142 chars ✓ |
+| CTA | "Free guide" | "Step-by-step" |
+| Emotional trigger | Benefit (cut load time) | Pain (losing visitors) |
+| Specificity | 50%, 10 techniques | 50% faster |
+| Readability | Clear | Clear |
+
+#### Step 3: Apply and Monitor
+```
+Apply the highest-scoring variant via update_content.
+Set a review date 2-4 weeks out.
+Compare CTR data in GSC after the review period.
+```
+
+### Title Tag A/B Variants
+Same process for title tags:
+```
+Original: "WordPress Speed Guide"
+Variant A: "WordPress Speed Optimization — 10 Techniques That Work"
+Variant B: "How to Speed Up WordPress by 50% (2026 Guide)"
+```
+
+## Rich Snippet Optimization
+
+### Content-Type Specific Guidance
+
+| Content Type | Rich Snippet Opportunity | Meta Optimization |
+|-------------|------------------------|-------------------|
+| How-to articles | HowTo schema | Include "how to" in title, steps count in meta |
+| FAQ pages | FAQPage schema | Include question in title, "answers" in meta |
+| Product pages | Product schema (rating, price) | Include price/rating in meta |
+| Review posts | Review schema (star rating) | Include rating in title/meta |
+| List posts | ItemList schema | Include count in title ("10 Best...") |
+
+### Meta for Featured Snippets
+To optimize for position zero:
+- Structure content to answer specific questions
+- Use the question as an H2 heading
+- Provide a concise answer in the first paragraph (40-60 words)
+- Follow with detailed explanation
+- Include a summary table or list
+
+## Step-by-Step Workflow
+
+### Step 1: Identify Optimization Targets
+```
+Use gsc_page_performance to find pages with:
+- High impressions (>500/month)
+- Below-average CTR for their position
+- Sort by potential impact (impressions × CTR gap)
+```
+
+### Step 2: Fetch Current Meta Elements
+```
+Use get_content for each target page.
+Extract: title tag, meta description, H1, URL slug.
+```
+
+### Step 3: Analyze and Score Current Meta
+Claude evaluates each element against rules above:
+- Title: length, keyword position, clarity, brand
+- Description: length, keyword, CTA, emotional trigger, specificity
+
+### Step 4: Generate Optimized Variants
+For each page, Claude produces:
+- 2 title tag variants
+- 2 meta description variants
+- Score comparison with current version
+- Estimated CTR impact
+
+### Step 5: Apply Changes
+```
+Use update_content to apply the winning variants.
+Document changes for tracking:
+- Page URL
+- Previous title/meta
+- New title/meta
+- Date changed
+- Baseline CTR from GSC
+```
+
+### Step 6: Monitor Results
+After 2-4 weeks:
+```
+Use gsc_page_performance to pull updated CTR data.
+Compare against baseline.
+Iterate if CTR did not improve.
+```
+
+## Best Practices
+
+- Prioritize high-impression pages first (biggest impact from small CTR gains)
+- Never duplicate meta descriptions across pages
+- Avoid meta descriptions that do not match page content (causes bounce)
+- Include the year in title tags for time-sensitive content (guides, reviews)
+- Use active voice and power verbs in meta descriptions
+- Test one change at a time to isolate impact
+- Review and refresh meta descriptions quarterly
+- Consider search intent when crafting meta — informational vs transactional queries need different approaches
+- Do not force keywords into meta descriptions unnaturally — readability trumps keyword presence

package/skills/wp-content-optimization/references/readability-analysis.md

@@ -0,0 +1,201 @@
+# Readability Analysis
+
+## Overview
+
+Readability analysis measures how easy content is to read and understand. Claude performs Flesch-Kincaid scoring, sentence length analysis, passive voice detection, paragraph assessment, and jargon identification directly on WordPress content — no external readability tools required. The goal is web-optimized content that matches the target audience's reading level.
+
+## Flesch-Kincaid Scoring
+
+### Formula
+```
+Flesch Reading Ease = 206.835 - (1.015 × ASL) - (84.6 × ASW)
+
+Where:
+  ASL = Average Sentence Length (words per sentence)
+  ASW = Average Syllables per Word
+```
+
+### Score Interpretation
+
+| Score | Grade Level | Audience | Action |
+|-------|-------------|----------|--------|
+| 90-100 | 5th grade | Very easy, children | Too simple for most web content |
+| 80-89 | 6th grade | Easy, conversational | Good for broad consumer content |
+| 70-79 | 7th grade | Fairly easy | Good for general web content |
+| 60-69 | 8th-9th grade | Standard | **Target for most WordPress content** |
+| 50-59 | 10th-12th grade | Fairly difficult | Acceptable for professional audience |
+| 30-49 | College | Difficult | Only for specialized/technical content |
+| 0-29 | Graduate | Very difficult | Rewrite for web consumption |
+
+### Target Scores by Content Type
+
+| Content Type | Target Score | Rationale |
+|--------------|-------------|-----------|
+| Blog posts | 60-70 | General audience, scannable |
+| Product descriptions | 65-75 | Clear, benefit-focused |
+| Technical documentation | 45-60 | Professional audience |
+| Landing pages | 70-80 | Must be instantly clear |
+| Email newsletters | 65-75 | Quick reading format |
+
+## Sentence Length Analysis
+
+### Guidelines
+
+| Metric | Target | Flag Threshold |
+|--------|--------|----------------|
+| Average sentence length | 15-20 words | >20 words average |
+| Maximum sentence length | 35 words | >40 words |
+| Short sentence ratio | 20-30% under 10 words | <10% short sentences |
+| Variety | Mix of short, medium, long | All same length |
+
+### Sentence Length Distribution
+Ideal content alternates between sentence lengths for rhythm:
+```
+Short sentence (5-10 words).     ← Punchy, creates emphasis
+Medium sentence (11-20 words).   ← Core information delivery
+Long sentence (21-30 words).     ← Complex ideas, supporting detail
+Short sentence (5-10 words).     ← Reset reader attention
+```
+
+### Fixing Long Sentences
+Strategies for breaking up long sentences:
+1. **Split at conjunctions** — break at "and", "but", "while", "although"
+2. **Remove relative clauses** — move "which" and "that" clauses to new sentences
+3. **Use bullet lists** — convert compound sentences into scannable lists
+4. **Front-load the point** — put the main idea first, details after
+
+### Example
+```
+Before (38 words):
+"WordPress is a content management system that allows users to create websites
+and blogs with themes and plugins, which can be customized to match any brand
+identity and extended with additional functionality."
+
+After (two sentences, 15 + 16 words):
+"WordPress is a content management system for creating websites and blogs.
+Themes and plugins let you customize the design and extend functionality."
+```
+
+## Passive Voice Detection
+
+### What to Flag
+Passive voice constructions where the subject receives the action:
+```
+Passive: "The plugin was installed by the user."
+Active:  "The user installed the plugin."
+```
+
+### Target
+- Maximum 10% of sentences in passive voice
+- 5% or less is ideal for web content
+
+### Common Passive Patterns
+| Passive Pattern | Active Rewrite |
+|----------------|----------------|
+| "was created by" | "[subject] created" |
+| "is recommended" | "we recommend" |
+| "can be configured" | "you can configure" |
+| "has been updated" | "we updated" / "[subject] updated" |
+| "should be installed" | "install" (imperative) |
+
+### When Passive Is Acceptable
+- Scientific or technical writing where the actor is irrelevant
+- When the object is more important than the subject
+- Policy or legal statements
+
+## Paragraph Length Guidelines
+
+| Metric | Target | Flag |
+|--------|--------|------|
+| Sentences per paragraph | 2-4 | >5 sentences |
+| Words per paragraph | 40-80 | >100 words |
+| One-sentence paragraphs | Occasional for emphasis | >3 consecutive |
+| Wall of text | Never | Any paragraph >150 words |
+
+### Web-Specific Rules
+- Break up paragraphs more aggressively than print
+- Use subheadings every 200-300 words
+- Include visual breaks (images, lists, blockquotes) every 300-400 words
+- One idea per paragraph
+
+## Jargon and Complexity Detection
+
+### Flag These Patterns
+1. **Industry jargon** — terms unknown to general audience (e.g., "canonical URL", "transclusion")
+2. **Acronyms without definition** — first use must define the acronym
+3. **Multi-syllable alternatives** — when simpler words exist
+4. **Nominalization** — turning verbs into nouns ("optimization" instead of "optimize")
+5. **Double negatives** — "not uncommon" instead of "common"
+
+### Simplification Table
+
+| Complex | Simple |
+|---------|--------|
+| utilize | use |
+| implement | add / set up |
+| functionality | feature |
+| methodology | method |
+| in order to | to |
+| at this point in time | now |
+| a large number of | many |
+| in the event that | if |
+| prior to | before |
+| subsequent to | after |
+
+## Step-by-Step Workflow
+
+### Step 1: Fetch Content
+```
+Use get_content with the post ID to retrieve full body content.
+Strip HTML tags for text-only analysis.
+```
+
+### Step 2: Compute Metrics
+Claude analyzes the plain text and computes:
+- Total word count
+- Total sentence count
+- Average sentence length (words/sentences)
+- Estimated Flesch-Kincaid score
+- Passive voice percentage
+- Average paragraph length
+- Jargon terms identified
+
+### Step 3: Generate Report
+```
+Readability Report for: "Post Title"
+═══════════════════════════════════
+Flesch-Kincaid Score: 58 (target: 60-70) ⚠ Below target
+Average Sentence Length: 22 words (target: <20) ⚠ Too long
+Passive Voice: 15% (target: <10%) ⚠ Too high
+Avg Paragraph Length: 95 words (target: <80) ⚠ Long paragraphs
+Jargon Terms Found: 3 (canonical, transclusion, REST endpoint)
+
+Top Issues:
+1. 5 sentences over 35 words — break into shorter sentences
+2. 3 paragraphs over 100 words — split with subheadings
+3. "canonical URL" used without definition — add brief explanation
+```
+
+### Step 4: Suggest Improvements
+Claude generates specific rewrite suggestions:
+- Rewrites for the 3 longest sentences
+- Paragraph split points with suggested subheadings
+- Passive-to-active voice conversions
+- Jargon simplifications or definitions to add
+
+### Step 5: Apply Changes (Optional)
+```
+Use update_content to apply approved changes.
+Preserve all HTML structure, links, and media.
+Only modify text content for readability improvements.
+```
+
+## Best Practices
+
+- Always analyze the full post body, not just excerpts
+- Consider the target audience when evaluating scores (technical docs have lower target)
+- Preserve the author's voice — readability improvements should not flatten style
+- Do not oversimplify to the point of losing nuance or accuracy
+- Run readability analysis after SEO optimization to ensure keyword insertion did not harm readability
+- Check readability on both desktop and mobile (shorter paragraphs matter more on mobile)
+- Use transition words (however, therefore, for example) to maintain flow after sentence splitting

package/skills/wp-content-optimization/references/seo-content-scoring.md

@@ -0,0 +1,245 @@
+# SEO Content Scoring
+
+## Overview
+
+SEO content scoring evaluates how well a piece of WordPress content is optimized for search engines. Claude performs keyword density analysis, heading hierarchy assessment, internal and external linking checks, image alt text coverage, and content structure evaluation. When GSC data is available, Claude cross-references actual search queries to validate keyword targeting.
+
+## Keyword Density Analysis
+
+### Target Ranges
+
+| Keyword Type | Target Density | Minimum | Maximum |
+|-------------|---------------|---------|---------|
+| Primary keyword | 1-2% | 0.5% | 3% |
+| Secondary keywords | 0.5-1% each | 0.3% | 1.5% |
+| LSI/related terms | Natural occurrence | — | — |
+
+### Density Calculation
+```
+Keyword Density = (Number of keyword occurrences / Total word count) × 100
+```
+
+### Placement Rules
+1. **Title (H1)** — primary keyword must appear (preferably front-loaded)
+2. **First paragraph** — primary keyword within first 100 words
+3. **Subheadings** — primary or secondary keyword in at least 1 H2
+4. **Body distribution** — keyword spread throughout, not clustered
+5. **Last paragraph** — primary keyword in conclusion
+6. **URL slug** — primary keyword in the URL
+
+### Keyword Stuffing Detection
+Flag when:
+- Primary keyword density exceeds 3%
+- Same keyword appears in consecutive sentences
+- Keyword is forced into unnatural phrasing
+- Multiple exact-match keyword repetitions in one paragraph
+
+### Example Analysis
+```
+Post: "WordPress Speed Optimization Guide" (1500 words)
+Primary keyword: "wordpress speed optimization"
+
+Occurrences: 18 times
+Density: 18/1500 × 100 = 1.2% ✓ Within target
+
+Distribution:
+- Title: ✓ Present
+- First paragraph: ✓ Present (word 12)
+- H2 headings: ✓ 1 of 4 H2s contains keyword
+- Last paragraph: ✓ Present
+- URL: ✓ /wordpress-speed-optimization-guide
+```
+
+## H2/H3 Hierarchy Coverage
+
+### Structure Requirements
+
+| Element | Requirement | Check |
+|---------|------------|-------|
+| H1 | Exactly 1 per page, contains primary keyword | Required |
+| H2 | 2-6 per post, secondary keywords in 50%+ | Required |
+| H3 | As needed under H2, long-tail keywords | Recommended |
+| H4+ | Rare, deep detail only | Optional |
+| Skip levels | Never (H1 → H3 without H2) | Flag as error |
+
+### Heading Optimization Checklist
+- [ ] One H1 only, matching the post title
+- [ ] H2s outline the main sections (scannable table of contents)
+- [ ] At least one H2 contains the primary keyword
+- [ ] H3s provide detail under their parent H2
+- [ ] No heading level is skipped
+- [ ] Headings are descriptive (not generic like "More Info")
+- [ ] Secondary keywords distributed across H2/H3 headings
+
+### Example Heading Audit
+```
+H1: WordPress Speed Optimization: Complete Guide          ✓ Primary keyword
+  H2: Why WordPress Speed Matters                         ✓ Descriptive
+    H3: Impact on SEO Rankings                            ✓ Secondary keyword
+    H3: Impact on User Experience                         ✓ Related term
+  H2: How to Measure WordPress Speed                      ✓ Secondary keyword
+  H2: 10 WordPress Speed Optimization Techniques          ✓ Primary keyword
+    H3: Enable Caching                                    ✓ Specific
+    H3: Optimize Images                                   ✓ Specific
+    H3: Minimize CSS and JavaScript                       ✓ Specific
+  H2: Conclusion                                          ⚠ Generic, add keyword
+```
+
+## Internal Linking Analysis
+
+### Minimum Standards
+
+| Content Length | Internal Links | Rationale |
+|---------------|---------------|-----------|
+| < 500 words | 1-2 links | Short content, few linking opportunities |
+| 500-1000 words | 2-3 links | Standard blog post |
+| 1000-2000 words | 3-5 links | In-depth article |
+| 2000+ words | 5-8 links | Pillar content, extensive topic |
+
+### What to Check
+1. **Link count** — meets minimum for content length
+2. **Anchor text** — descriptive, keyword-rich (not "click here")
+3. **Link relevance** — linked pages are topically related
+4. **Link distribution** — links spread throughout content, not clustered
+5. **Orphan detection** — identify posts with zero inbound internal links
+6. **Reciprocal links** — pillar pages link to cluster posts and vice versa
+
+### Link Opportunity Detection
+Claude identifies linking opportunities by:
+- Finding mentions of topics that have dedicated pages on the site
+- Detecting keyword phrases that match other post titles
+- Identifying related content that could provide context
+- Suggesting links to pillar pages from cluster content
+
+## External Linking
+
+### Guidelines
+
+| Metric | Target |
+|--------|--------|
+| External links per post | 1-2 minimum |
+| Link targets | Authoritative, relevant sources (.gov, .edu, industry leaders) |
+| Link freshness | Source content should be current (not outdated) |
+| Nofollow | Use for sponsored or untrusted links |
+| Open in new tab | Yes for external links (target="_blank") |
+
+### What to Avoid
+- Linking to direct competitors' commercial pages
+- Linking to low-authority or spammy sites
+- Excessive external links (>5 per 1000 words)
+- Broken external links (check periodically)
+
+## Image Alt Text Coverage
+
+### Requirements
+| Check | Standard |
+|-------|----------|
+| All images have alt text | 100% coverage required |
+| Alt text includes keyword | At least 1 image per post |
+| Alt text is descriptive | Describes the image content |
+| Alt text length | 5-15 words |
+| Decorative images | Empty alt="" (not missing alt) |
+
+### Example
+```
+Good: alt="WordPress performance optimization dashboard showing page load times"
+Bad:  alt="image1" or alt="" (on informational image) or alt missing
+```
+
+## Combining with GSC Data
+
+When Google Search Console is available, enrich the SEO scoring with real search data:
+
+### Step 1: Fetch Search Queries for the Page
+```
+Use gsc_search_analytics with the page URL to get actual queries driving traffic.
+```
+
+### Step 2: Compare Target vs Actual Keywords
+```
+Target keyword: "wordpress speed optimization"
+Actual top queries from GSC:
+  1. "wordpress speed" (pos 8, 500 impressions)
+  2. "wordpress performance" (pos 12, 300 impressions)
+  3. "how to speed up wordpress" (pos 15, 200 impressions)
+
+Analysis: Content ranks for related terms but not the exact target.
+Action: Strengthen primary keyword presence, add "how to speed up wordpress" as H2.
+```
+
+### Step 3: Identify Keyword Gaps
+- Queries with high impressions but low CTR → meta description issue
+- Queries with high position (>10) → content depth issue
+- Queries the page ranks for that are not in the content → add sections
+
+## SEO Content Score Card
+
+### Scoring Template
+```
+SEO Content Score for: "Post Title"
+═════════════════════════════════════
+Keyword Density:      1.2%    ✓ (target: 1-2%)
+Keyword in H1:        Yes     ✓
+Keyword in First 100: Yes     ✓
+Keyword in URL:       Yes     ✓
+
+H2 Count:             4       ✓ (target: 2-6)
+H2 with Keywords:     2/4     ✓ (target: 50%+)
+H3 Count:             6       ✓
+Heading Hierarchy:    Valid   ✓ (no skipped levels)
+
+Internal Links:       2       ⚠ (target: 3-5 for 1200 words)
+External Links:       1       ✓ (target: 1-2)
+Anchor Text Quality:  Good    ✓
+
+Image Alt Coverage:   3/4     ⚠ (1 image missing alt text)
+Keyword in Alt:       1/4     ✓
+
+Overall Score:        7.5/10
+Priority Fixes:
+1. Add 1-2 more internal links to related content
+2. Add alt text to image in section 3
+```
+
+## Step-by-Step Workflow
+
+### Step 1: Fetch Content
+```
+Use get_content to retrieve the full post HTML.
+Parse headings, links, images, and body text.
+```
+
+### Step 2: Identify Target Keyword
+Either provided by user or extracted from:
+- Post title analysis
+- GSC top query for the URL
+- Content theme analysis
+
+### Step 3: Run All Checks
+Claude analyzes in sequence:
+1. Keyword density and placement
+2. Heading hierarchy and keyword coverage
+3. Internal and external link audit
+4. Image alt text coverage
+5. GSC data cross-reference (if available)
+
+### Step 4: Generate Score Card
+Produce the formatted score card with pass/fail for each criterion.
+
+### Step 5: Prioritize Fixes
+Rank issues by impact:
+1. Missing keyword in title/H1 (highest impact)
+2. Keyword density out of range
+3. No internal links
+4. Missing alt text
+5. Heading hierarchy issues
+6. External link additions
+
+## Best Practices
+
+- Run SEO scoring before publishing (pre-publish checklist)
+- Re-score after content updates to ensure changes did not break optimization
+- Use GSC data to validate keyword targeting against real search behavior
+- Do not sacrifice readability for SEO score — readability comes first
+- Update scoring analysis quarterly as search patterns evolve
+- Track score changes over time to measure optimization impact