anysite-competitor-analyzer

name

anysite-competitor-analyzer

description

Deep competitive intelligence combining web scraping, LinkedIn data, social media monitoring, leadership analysis, GitHub activity, Glassdoor sentiment, and community insights. Analyzes founders/C-level profiles, tracks real-time signals vs quarterly reports, and creates comprehensive competitor profiles. Use when asked to analyze competitors, research leadership teams, investigate market positioning, compare products/pricing, assess strategic threats, or gather intelligence on founders and key executives.

Competitor Analyzer

Systematic framework for gathering and analyzing competitive intelligence using Anysite MCP v2 tools.

Tool Interface (v2)

All data fetching uses the unified v2 meta-tools:

• execute(source, category, endpoint, params) - Fetch data. Returns first 10 items + cache_key. If next_offset is present, use get_page() to load more.
• get_page(cache_key, offset, limit) - Paginate through cached results from a previous execute(). Data cached 7 days.
• query_cache(cache_key, conditions, sort_by, aggregate, group_by) - Filter, sort, count, or aggregate already-fetched data without new API calls.
• export_data(cache_key, format) - Export full dataset as CSV, JSON, or JSONL. Returns a download URL.
• discover(source, category) - Inspect available endpoints and params before calling execute().

Error handling: If execute() returns an error with llm_hint, follow the hint to fix the call. Common issues: wrong URN format, alias not found (search first), fsd_company vs company: prefix mismatch.

When to Use This Skill

Trigger this skill when users ask to:

• "Analyze [competitor name]"
• "Research our competitors"
• "Create a competitive analysis of [company]"
• "How does [competitor] position themselves?"
• "What are [competitor]'s strengths and weaknesses?"
• "Compare our product with [competitor]"
• "Who are our main competitors?"
• "Build a battle card for [competitor]"

Quick Start

For single competitor analysis:

# 1. Generate analysis template
python scripts/analyze_competitor.py "Competitor Name" "https://competitor.com"

# 2. Use Anysite v2 tools to gather data (see workflow below)

# 3. Fill in the JSON template with findings

# 4. Generate final report
python scripts/analyze_competitor.py "Competitor Name" "https://competitor.com" | \
    python -c "import sys,json; exec('from scripts.analyze_competitor import format_markdown_report; print(format_markdown_report(json.load(sys.stdin)))')" \
    > /mnt/user-data/outputs/competitor_report.md

Analysis Workflow

Phase 1: Foundation (15-20 min)

Step 1: Initialize Analysis Structure

Run the analysis script to create structured template:

python scripts/analyze_competitor.py "Competitor Name" "https://competitor.com" > /tmp/analysis.json

Step 2: Web Presence Reconnaissance

Scrape key pages to understand positioning:

# Homepage - core messaging
execute("webparser", "parse", "parse", {
    "url": "https://competitor.com",
    "only_main_content": true,
    "strip_all_tags": true
})

# Pricing - cost structure
execute("webparser", "parse", "parse", {
    "url": "https://competitor.com/pricing",
    "only_main_content": true
})

# About - company background
execute("webparser", "parse", "parse", {
    "url": "https://competitor.com/about",
    "only_main_content": true,
    "extract_contacts": true
})

Extract from homepage:

• H1/H2 headlines → positioning_statement
• Feature bullets → core_features
• Customer logos → customer_logos
• Value prop → value_proposition

Extract from pricing:

• Tier names and prices → pricing.tiers
• Cost per unit → pricing.unit_economics
• Free tier details → pricing.free_tier_limits
• Entry price → pricing.entry_price

Extract from about:

• Company description → company_overview.description
• Location → company_overview.headquarters
• Team size hints → company_overview.employee_count

Phase 2: LinkedIn Intelligence (10-15 min)

Step 3: Find Company Profile

# Search for company
execute("linkedin", "search", "search_companies", {
    "keywords": "competitor name",
    "count": 5
})

# Get detailed profile using slug from search results
execute("linkedin", "company", "company", {
    "company": "company-slug-from-search"
})

Extract:

• follower_count → Online presence indicator
• employee_count → Company size
• description → Self-positioning
• headquarters → Location
• specialties → Keywords they emphasize

Step 4: Analyze Team & Growth

# Check employee growth signals (use search_users with current_company filter)
execute("linkedin", "search", "search_users", {
    "current_company": "company-slug",
    "keywords": "engineer developer",
    "count": 50
})

# Find leadership
execute("linkedin", "search", "search_users", {
    "current_company": "company-slug",
    "title": "CEO founder",
    "count": 10
})

# Get employee stats breakdown (functions, seniority, growth trends)
# First get company URN from company profile, convert fsd_company to company: prefix
execute("linkedin", "company", "company_employee_stats", {
    "urn": {"type": "company", "value": "COMPANY_ID_FROM_URN"}
})

Use findings to assess:

• Team size → growth_indicators.employee_growth
• Eng:sales ratio → GTM strategy signal
• Recent hires → growth phase indicator

Tip: Use query_cache() on the employee search results to filter by title or sort by relevance without re-fetching:

query_cache(cache_key, conditions=[{"field": "headline", "operator": "contains", "value": "engineer"}])

Step 5: Content Strategy

# Analyze posting activity
# Use company: prefix URN from the company profile (convert fsd_company:{id} to company:{id})
execute("linkedin", "company", "company_posts", {
    "urn": {"type": "company", "value": "COMPANY_ID_FROM_URN"},
    "count": 20
})

Analyze posts for:

• Frequency → content_strategy.blog_frequency
• Themes → content_strategy.key_topics
• Engagement → online_presence.linkedin.engagement_quality
• Tone → content_strategy.tone_of_voice

Tip: Use query_cache() to aggregate engagement metrics across fetched posts:

query_cache(cache_key, aggregate=[{"field": "comment_count", "function": "avg"}])

Phase 3: Deep Social & Community Research (20-30 min)

Step 6: Twitter Deep Dive

A. Company Account Analysis

# Get profile stats
execute("twitter", "user", "get", {
    "username": "competitor_handle"
})

# Recent activity (analyze more posts)
execute("twitter", "user_tweets", "get", {
    "username": "competitor_handle",
    "count": 100
})
# If next_offset returned, use get_page() to load more:
# get_page(cache_key, offset=next_offset, limit=50)

Extract from company account:

• Followers → reach indicator
• Tweet frequency → activity level
• Content mix (product updates, thought leadership, customer engagement)
• Response time to mentions
• Tone of voice
• Most engaging tweets (viral content patterns)

Tip: Use query_cache() to find top-performing tweets:

query_cache(cache_key, sort_by=[{"field": "favorite_count", "order": "desc"}])

B. Founder/Executive Twitter Presence

# Find and analyze founder accounts
execute("twitter", "user", "get", {
    "username": "founder_handle"
})

execute("twitter", "user_tweets", "get", {
    "username": "founder_handle",
    "count": 100
})

Leadership Twitter signals:

• Personal brand strength
• Technical credibility (what they share)
• Customer engagement quality
• Industry thought leadership
• Follower quality (who follows them)
• Retweet patterns (what they amplify)

C. Brand Mentions & Sentiment

# Comprehensive mention search
execute("twitter", "search", "search_posts", {
    "query": "competitor_name OR @handle OR #competitor_hashtag",
    "count": 200
})

# Problem/complaint mentions
execute("twitter", "search", "search_posts", {
    "query": "competitor_name (problem OR issue OR bug OR slow OR expensive)",
    "count": 100
})

# Positive sentiment
execute("twitter", "search", "search_posts", {
    "query": "competitor_name (love OR great OR amazing OR best OR solved)",
    "count": 100
})

# Competitive mentions
execute("twitter", "search", "search_posts", {
    "query": "competitor_name vs OR competitor_name alternative OR switching from competitor_name",
    "count": 100
})

Sentiment scoring:

For each mention batch, calculate:
- Positive mentions: praise, recommendations, success stories
- Negative mentions: complaints, frustrations, churn signals
- Neutral mentions: questions, feature discussions
- Competitive mentions: comparisons with alternatives

Sentiment Score = (Positive - Negative) / Total
Range: -1.0 (very negative) to +1.0 (very positive)

Tip: Use query_cache() to filter cached mentions by sentiment keywords without re-fetching:

query_cache(cache_key, conditions=[{"field": "text", "operator": "contains", "value": "love"}])

D. Customer Voice Analysis

# Find actual users
execute("twitter", "search", "search_posts", {
    "query": "using competitor_name OR tried competitor_name",
    "count": 100
})

# Power users
execute("twitter", "search", "search_posts", {
    "query": "@handle thanks OR @handle helped OR @handle support",
    "count": 50
})

Extract:

• Real use cases (what customers actually do)
• Pain points (what they struggle with)
• Success stories (what works well)
• Feature requests (what they want)
• Support quality (how fast company responds)

Step 7: Reddit Deep Community Intelligence

A. Brand Presence Mapping

# General mentions across Reddit
execute("reddit", "search", "search_posts", {
    "query": "competitor_name",
    "count": 100
})

# Industry-specific searches (combine with subreddit keywords)
relevant_topics = [
    "SaaS", "startups", "Entrepreneur",  # Business
    "webdev", "programming", "devops",    # Tech
    "nocode", "automation",               # No-code
    "datascience", "analytics"            # Data
]

for topic in relevant_topics:
    execute("reddit", "search", "search_posts", {
        "query": f"competitor_name {topic}",
        "count": 50
    })

B. Competitive Discussions

# Direct comparisons
execute("reddit", "search", "search_posts", {
    "query": "competitor_name vs",
    "count": 100
})

# Alternative searches
execute("reddit", "search", "search_posts", {
    "query": "alternative to competitor_name",
    "count": 100
})

execute("reddit", "search", "search_posts", {
    "query": "better than competitor_name",
    "count": 50
})

# Problem space
execute("reddit", "search", "search_posts", {
    "query": "[problem they solve] tools OR solutions",
    "count": 100
})

C. Deep Thread Analysis

For high-engagement threads, get comments:

# Get specific post details
execute("reddit", "posts", "posts", {
    "post_url": "reddit.com/r/subreddit/comments/..."
})

# Get all comments
execute("reddit", "posts", "posts_comments", {
    "post_url": "reddit.com/r/subreddit/comments/..."
})

Analyze thread comments for:

• Detailed user experiences
• Technical discussions
• Feature comparisons
• Pricing discussions
• Customer support experiences
• Decision factors (why they chose/didn't choose)

Tip: Use query_cache() on fetched comments to sort by score and find the most upvoted opinions:

query_cache(cache_key, sort_by=[{"field": "score", "order": "desc"}])

D. Sentiment & Voice Analysis

Positive signals:

• "I love [competitor]"
• "Works perfectly for..."
• "Best tool for..."
• "Highly recommend"
• "Switched to [competitor] and..."

Negative signals:

• "Disappointed with..."
• "Overpriced"
• "Customer support is..."
• "Buggy/unreliable"
• "Looking for alternative"
• "Switched away from..."

Neutral/informational:

• "How does [competitor] work?"
• "Anyone tried [competitor]?"
• "Pricing question"
• Feature clarifications

E. Community Size & Engagement

Calculate metrics:

Brand Awareness Score:
- Total unique mentions (last 30 days)
- Number of different subreddits mentioned in
- Average upvotes per mention
- Comment volume per mention

Community Health:
- Positive/Negative mention ratio
- Response rate to questions
- Problem resolution in comments
- Community helping each other

Step 7.5: Cross-Platform Insight Synthesis

Compare Twitter vs Reddit:

Twitter typically shows:

• Official company narrative
• Marketing messaging
• Quick customer service interactions
• Surface-level sentiment
• Broader reach

Reddit typically reveals:

• Unfiltered user opinions
• Detailed technical discussions
• Pricing sensitivity
• Competitive comparisons
• Real problems and workarounds

Look for disconnects:

• Company claims strong product (Twitter) but users complain (Reddit)
• High Twitter engagement but low Reddit mentions → Marketing-driven, not organic
• Reddit loves it but low Twitter presence → Word-of-mouth, under-marketed
• Consistent messaging → Authentic product-market fit

Phase 4: Leadership & Founders Intelligence (15-20 min)

Step 8: Identify Key Leaders

# Find founders and C-level
execute("linkedin", "search", "search_users", {
    "company_keywords": "competitor-name",
    "title": "founder OR CEO OR CTO OR CPO",
    "count": 10
})

# Get detailed profiles
execute("linkedin", "user", "user", {
    "user": "founder-linkedin-username",
    "with_experience": true,
    "with_education": true,
    "with_skills": true
})

Extract for each leader:

• Full career history → their experience and expertise
• Previous companies → track record
• Education background → academic credentials
• Skills → technical depth
• Languages → market reach
• Recommendations → credibility signals

Step 9: Analyze Leadership Activity

# Get personal posts (use fsd_profile URN from user profile)
execute("linkedin", "user", "user_posts", {
    "urn": {"type": "fsd_profile", "value": "USER_URN_VALUE"},
    "count": 50
})

# Check comments on others' posts
execute("linkedin", "user", "user_comments", {
    "urn": {"type": "fsd_profile", "value": "USER_URN_VALUE"},
    "count": 30
})

# See what they're engaging with
execute("linkedin", "user", "user_reactions", {
    "urn": {"type": "fsd_profile", "value": "USER_URN_VALUE"},
    "count": 50
})

Analyze for:

• Posting frequency and themes
• Technical depth in posts
• Market perspective
• Customer engagement
• Thought leadership quality
• Network quality (who engages with them)

Tip: Use query_cache() on leadership posts to aggregate engagement:

query_cache(cache_key, aggregate=[
    {"field": "comment_count", "function": "avg"},
    {"field": "comment_count", "function": "sum"}
])

Step 10: Twitter Leadership Presence

# Founder Twitter activity
execute("twitter", "user", "get", {
    "username": "founder_handle"
})

execute("twitter", "user_tweets", "get", {
    "username": "founder_handle",
    "count": 100
})

Leadership indicators:

• Personal brand strength
• Technical credibility
• Customer relationships
• Industry influence
• Communication style
• Transparency level

Phase 5: Technical & Data Discovery (10-15 min)

Step 11: Documentation Quality

# Scrape docs homepage
execute("webparser", "parse", "parse", {
    "url": "https://competitor.com/docs",
    "only_main_content": true
})

# Check API reference
execute("webparser", "parse", "parse", {
    "url": "https://competitor.com/api",
    "only_main_content": true
})

Assess:

• Documentation completeness
• Code examples presence
• Interactive explorer
• SDK availability → Feed into technical_capabilities

Step 12: GitHub Presence (if applicable)

# Parse GitHub profile page
execute("webparser", "parse", "parse", {
    "url": "https://github.com/competitor-org",
    "only_main_content": true
})

# Check main repository
execute("webparser", "parse", "parse", {
    "url": "https://github.com/competitor-org/main-repo",
    "only_main_content": true
})

Extract:

• Star count (developer interest)
• Fork count (actual usage)
• Commit frequency (development velocity)
• Contributors count (community size)
• Issue response time (support quality)
• Open source components (ecosystem play)

Step 13: Alternative Data Sources

# Glassdoor reviews (if company page exists)
execute("webparser", "parse", "parse", {
    "url": "https://www.glassdoor.com/Reviews/competitor-name",
    "only_main_content": true
})

What to extract:

• Overall rating (employee satisfaction)
• CEO approval rating (leadership quality)
• Salary ranges (compensation level)
• Interview difficulty (hiring standards)
• Work-life balance (culture signal)
• Recent reviews (current state)

Step 14: Integration Ecosystem

# Get sitemap to find all pages
execute("webparser", "sitemap", "sitemap", {
    "url": "https://competitor.com",
    "count": 50
})

# Parse integrations page
execute("webparser", "parse", "parse", {
    "url": "https://competitor.com/integrations",
    "only_main_content": true
})

Extract:

• Integration partners → technical_capabilities.integrations
• Platform focus (Zapier, enterprise tools, etc.)
• API-first vs GUI-first

Phase 6: Synthesis (15-20 min)

Step 15: Competitive Analysis

Compare findings against your own product:

Strengths (what they do well):

• Identify 3-5 clear advantages they have
• Based on features, pricing, market position, or execution

Weaknesses (where they struggle):

• Identify 3-5 clear gaps or problems
• Missing features, high prices, poor UX, etc.

Opportunities (what you can exploit):

• Their weaknesses that you can capitalize on
• Underserved segments they're missing
• Messaging/positioning gaps

Threats (what you need to watch):

• Their strengths that could hurt you
• Recent funding or growth
• Feature development velocity

Step 16: Strategic Insights

Synthesize everything into:

Key Takeaways (3-5 bullets):

• Most important findings
• Clear, actionable insights

Competitive Threats (2-3 bullets):

• What they could do to hurt your position
• Their strategic advantages

Opportunities to Exploit (3-5 bullets):

• How to position against them
• Their vulnerabilities to target
• Market gaps they're missing

Watch Areas:

• Things to monitor quarterly
• Signals of strategic shifts

Step 17: Generate Final Report

Update the JSON template with all findings, then generate markdown:

import json
from scripts.analyze_competitor import save_analysis

# Load populated template
with open('/tmp/analysis.json', 'r') as f:
    data = json.load(f)

# Generate reports
json_path, md_path = save_analysis(data)
print(f"Reports saved:\n  JSON: {json_path}\n  Markdown: {md_path}")

Move final files to outputs:

cp /tmp/analysis.json /mnt/user-data/outputs/
cp /tmp/analysis.md /mnt/user-data/outputs/

Tip: Use export_data() to export any collected dataset for the final report:

export_data(cache_key, "csv")   # Returns download URL for spreadsheet import
export_data(cache_key, "json")  # Returns download URL for programmatic use

Advanced Techniques

Multi-Competitor Analysis

For analyzing 3-5 competitors simultaneously:

1. Run analysis workflow for each competitor
2. Create comparison matrix in spreadsheet format
3. Focus on key differentiators:
   • Pricing comparison table
   • Feature matrix (rows=features, cols=competitors)
   • Market position map (price vs capabilities)
   • Social presence comparison

Tip: Export each competitor's data and use query_cache() to compare metrics:

# After fetching data for each competitor, aggregate and compare
query_cache(cache_key, aggregate=[{"field": "follower_count", "function": "sum"}])
export_data(cache_key, "csv")

Ongoing Monitoring

For quarterly updates (not full re-analysis):

Quick check (30 min):

# 1. Re-scrape pricing
execute("webparser", "parse", "parse", {"url": "https://competitor.com/pricing"})

# 2. Check recent posts (use company: prefix URN)
execute("linkedin", "company", "company_posts", {
    "urn": {"type": "company", "value": "COMPANY_ID"},
    "count": 10
})

# 3. Employee growth
execute("linkedin", "search", "search_users", {
    "current_company": "company-slug",
    "count": 20
})

# 4. Recent mentions
execute("twitter", "search", "search_posts", {
    "query": "competitor",
    "count": 50
})

Update only changed sections in JSON template.

Battle Card Creation

For sales team quick reference:

Focus on:

1. Quick facts (1-2 sentences)
2. Head-to-head feature comparison (table format)
3. Pricing comparison (clear numbers)
4. 3 reasons we win
5. 3 reasons we might lose
6. Talk tracks ("When they say X, we say Y")

Keep to 1-2 pages maximum.

Reference Files

When you need detailed guidance:

• Data collection methodology: See data_collection.md

  • Use when unsure which Anysite tools to use
  • Use when planning data gathering strategy
  • Contains detailed tool parameters and extraction techniques

• Analysis frameworks: See analysis_frameworks.md

  • Use when analyzing specific company types (SaaS vs Enterprise vs Consumer)
  • Use when creating battle cards or competitive matrices
  • Contains templates for different output formats

Common Patterns

Pattern 1: Rapid Assessment (30-45 min)

For quick competitive scan:

1. Homepage + pricing scrape
2. LinkedIn company profile
3. Recent social posts (20 total)
4. Fill core sections only (skip deep dives)
5. Generate brief summary (1 page)

Pattern 2: Deep Intelligence (2-3 hours)

For comprehensive analysis:

1. Full web presence (7-10 pages)
2. Complete LinkedIn intelligence
3. Social media deep dive (50-100 posts/mentions)
4. Community sentiment analysis
5. Technical documentation review
6. Full JSON template populated
7. Detailed markdown report

Pattern 3: Pricing Focus

For pricing-specific analysis:

1. Scrape all pricing pages
2. Calculate unit economics
3. Map tier structures
4. Compare to market
5. Identify pricing strategy
6. Generate pricing comparison table

Pattern 4: Leadership Focus

For founder/team intelligence:

1. Identify all founders and C-level
2. Deep dive into founder LinkedIn profiles
3. Analyze personal posting activity (50+ posts)
4. Track Twitter presence and influence
5. Map previous company experience
6. Assess thought leadership quality
7. Evaluate public credibility

Use when:

• Considering partnerships
• Evaluating acquisition targets
• Assessing strategic threats
• Understanding company DNA

Tips for Effective Analysis

Be Systematic:

• Follow the phase order
• Don't skip LinkedIn intelligence (best growth signals)
• Always check pricing (most volatile data)

Think Strategically:

• Not just "what" they do, but "why"
• Look for patterns in their behavior
• Consider their constraints (funding, team size)

Verify Claims:

• Marketing copy ≠ reality
• Cross-reference multiple sources
• Note confidence levels (verified vs estimated)

Focus on Actionable Insights:

• Don't just describe, analyze implications
• What should YOUR company do based on findings?
• What threats need immediate response?

Document Data Freshness:

• Always note analysis date
• Mark which data is recent vs stale
• Plan update frequency based on importance

Use v2 Efficiency Features:

• Use get_page() instead of re-executing with higher counts
• Use query_cache() to filter/sort/aggregate without new API calls
• Use export_data() to generate shareable CSV/JSON files
• Check llm_hint in error responses for fix guidance

Output Quality Standards

Good competitive analysis includes:

• Clear positioning statement
• Quantified metrics (prices, follower counts, team size)
• Specific examples (actual quotes, feature lists)
• Strategic implications explained
• Data sources noted
• Confidence levels indicated

Avoid:

• Vague assessments ("they seem good at X")
• Unsupported claims ("probably losing money")
• Missing pricing details
• Outdated data without date stamps
• Pure feature lists without analysis

Troubleshooting

"Can't find LinkedIn company":

• Try variations of company name
• Search for CEO name, find company from profile
• Check if they use different legal name

"Pricing page missing/unclear":

• Check /plans, /buy, /subscribe URLs
• Look for pricing calculator
• Note "Contact Sales" as signal (enterprise focus)

"No social media presence":

• Still document the absence (itself a signal)
• Check founder personal accounts
• Look for employee posting activity

"Too much data, overwhelmed":

• Start with Phase 1 & 2 only (foundation + LinkedIn)
• Generate partial report
• Add Phase 3 & 4 if needed for depth

"execute() returned error with llm_hint":

• Read the hint carefully and adjust params
• Common: wrong URN format (use company: prefix, not fsd_company)
• Common: alias not found (use search endpoint first)

"Need more than 10 results from execute()":

• Check if response includes next_offset
• Use get_page(cache_key, offset=next_offset, limit=50) to load more
• Do NOT re-execute with a higher count — use pagination