Competitor Signals

Phase 1: Collect Context

Step 1: Gather Competitor Information

Ask the user:

"To find leads from competitor activity, I need:

1. Who are your competitors? (product names and company names)
2. Do you know their Product Hunt slugs? (the URL path on producthunt.com/posts/SLUG)
3. Any specific competitor launches or announcements you've seen recently?
4. Are there competitors or signals you specifically want to track? (e.g., a competitor just raised funding, launched a new feature, or got press coverage)"

Step 2: Discover Competitors (if user needs help)

If the user doesn't have a complete competitor list, help them discover competitors:

2a. Product Hunt search:

• Search producthunt.com for the user's product category
• Note: PH doesn't have a great search API — use web search: "site:producthunt.com [product category]"

2b. G2/Capterra category pages:

• Search: "[product category] G2" or "[product category] Capterra"
• These pages list all competitors in a category with rankings

2c. "Alternatives to" sites:

• Search: "[known competitor] alternatives"
• Sites like alternativeto.net, slant.co, stackshare.io list competitors

2d. Ask the user:

"Based on my research, here are competitors I've found in your space: [list]. Are there any I'm missing? Any you'd like to exclude (e.g., not really competitors, too different in market segment)?"

Step 3: Find Product Hunt Slugs

For each competitor, find their PH launches:

• Search: "site:producthunt.com [competitor name]"
• Or browse: producthunt.com/products/[competitor-name]
• Note the slug from the URL: producthunt.com/posts/SLUG
• A competitor may have multiple launches (initial launch + feature launches)

Step 4: Identify Competitor Web Pages to Scrape

For each competitor, identify pages the agent should scrape:

Case studies page: [competitor].com/customers or [competitor].com/case-studies

• Extract: company names, logos, quotes, person names, titles
• These are PROVEN BUYERS in the category

Testimonials page: Often on the homepage or a dedicated page

• Extract: person name, title, company, quote
• These are current users who publicly endorsed the competitor

Blog: [competitor].com/blog

• Guest posts by customers are case studies in disguise
• "How [Company X] uses [Competitor]" = case study

Present all discovered pages to the user for review.

Phase 2: Agent-Driven Scraping

Step 5: Scrape Competitor Websites

Before running the tool, the agent should manually scrape competitor case studies and testimonials. This is agent-driven because every competitor website has a different format.

For each competitor's case study page:

1. Navigate to the page using web fetch or Chrome DevTools
2. Extract all customer company names and any associated person names/quotes
3. Note the case study URL for reference

For each competitor's testimonials page:

1. Extract: person name, title, company, quote text
2. These are high-value signals — these people actively chose to endorse the competitor

Save all scraped data to ${CLAUDE_SKILL_DIR}/../.tmp/competitor_manual_signals.json:

[
    {
        "person_name": "Sarah Chen",
        "company": "TechCorp",
        "signal_type": "case_study_company",
        "signal_label": "Competitor Case Study",
        "competitor": "Twilio",
        "context": "How TechCorp scaled video calls to 100K users with Twilio",
        "url": "https://twilio.com/case-studies/techcorp",
        "profile_url": "",
        "date": "",
        "source": "Manual",
        "engagement": 0
    }
]

Step 6: Check Tech Press

Search for recent articles about competitors:

• "[competitor] TechCrunch"
• "[competitor] The New Stack"
• "[competitor] InfoQ"
• "[competitor] DevOps.com"
• "[competitor] launch announcement"
• "[competitor] raises funding"

For articles found:

• Note the article URL and key companies/people mentioned
• If the article has comments, check for people expressing opinions
• Add notable findings to the manual signals JSON

Phase 3: Execute Tool

Step 7: Save Config

cat > ${CLAUDE_SKILL_DIR}/../.tmp/competitor_signals_config.json << 'CONFIGEOF'
{
    "competitors": ["Twilio", "Agora", "Vonage", "Daily.co"],
    "product_hunt_slugs": ["twilio-video", "agora-2", "daily-co"],
    "days": 90,
    "manual_signals_file": "${CLAUDE_SKILL_DIR}/../.tmp/competitor_manual_signals.json",
    "skip": []
}
CONFIGEOF

Step 8: Run the Tool

python3 ${CLAUDE_SKILL_DIR}/scripts/competitor_signals.py \
    --config ${CLAUDE_SKILL_DIR}/../.tmp/competitor_signals_config.json \
    --output ${CLAUDE_SKILL_DIR}/../.tmp/competitor_signals.csv

The tool will:

1. Try Product Hunt API first (if PRODUCTHUNT_TOKEN is set)
2. Fall back to Apify PH scraper if API names are redacted
3. Search HN for all competitor names (stories + comments, last 90 days)
4. Load manual signals (case studies, testimonials, press)
5. Detect "switching signals" (highest priority — people saying they're moving to/from a competitor)
6. Deduplicate and score
7. Export CSV with switching signals highlighted

Phase 4: Analyze & Recommend

Step 10: Analyze Results

10a. Switching Signals (HIGHEST PRIORITY)

• These are people who publicly said they're switching from or evaluating alternatives to a competitor
• List every switching signal with full context
• These leads should be contacted IMMEDIATELY — they're in active evaluation
• Outreach angle: "I noticed you mentioned looking for alternatives to [competitor] — here's how we compare"

10b. Case Study Companies

• These are PROVEN BUYERS in the category
• They've already committed budget to the problem space
• The decision-maker already said yes once — they'll consider alternatives if you offer something better
• Recommend enriching these companies via SixtyFour to find the current decision-maker

10c. Testimonial Authors

• Current users of the competitor who are vocal about it
• They may be satisfied (hard sell) OR they may have moved on since the testimonial
• Good for understanding what the competitor does well (competitive intel)
• If the testimonial mentions specific pain points or limitations, that's an opening

10d. Product Hunt Activity

• Commenters asking questions = evaluating the category
• Commenters with negative feedback = potentially dissatisfied
• Upvoters = interested in the space (weaker signal, higher volume)

10e. HN Discussion

• Commenters engaging with competitor stories = following the space
• People sharing experiences (positive or negative) = active users or evaluators

10f. Competitor-Level Analysis

• Which competitor generates the most signals? (largest audience = most opportunity)
• Which competitor has the most negative signals? (weakest competitor = easiest to displace)
• Are there any surprises? (unknown competitor getting a lot of attention?)

Step 11: Recommend Next Steps

1. Switching signals (immediate outreach):

   • Enrich these people via SixtyFour NOW
   • They're in active evaluation — speed matters
   • Personalize based on what they said ("You mentioned [specific pain]...")

2. Case study companies (account-based approach):

   • These companies have budget for this category
   • Use SixtyFour /enrich-company to understand them
   • Find the decision-maker (not the person in the case study, who may have left)
   • Outreach angle: "Companies like yours in [industry] are switching to us because..."

3. PH commenters asking questions:

   • They're early in evaluation
   • Can reply directly on Product Hunt (public, non-intrusive)
   • Or enrich and reach out privately

4. Cross-reference with other signals:

   • If a company appears in competitor case studies AND in job signals (hiring for the role) -> they're invested but possibly scaling beyond the competitor
   • If a person appears in competitor PH comments AND in community signals -> they're deeply researching the space

Step 12: Ask for Go-Ahead

"Would you like me to:

1. Enrich the switching signal leads immediately (highest priority)
2. Enrich the case study companies and find decision-makers
3. Cross-reference with data from other signal skills
4. Scrape additional competitor pages for more signals
5. Export for manual review first"

Signal Scoring

Output Schema (Single Sheet)

Cost Estimates

Lookback Period

Default: 90 days. Competitor launches and case studies have a longer shelf life than Reddit posts. Someone who commented on a competitor's PH launch 60 days ago is still a viable lead.