Competitor Analysis Skill

Purpose

Extracts observable marketing data from competitor websites. Observation only, no interpretation.

Position in Workflow

code

[competitor-analysis] → [keyword-research] → [lead-gen-copy]
         ↓
   Raw observations      →    Patterns    →    Copy decisions

This skill produces RAW DATA. Copy decisions happen in other skills.

Input

yaml

competitor_urls: ["url1", "url2", "url3"]  # min 3, max 5
analysis_focus: full | messaging | trust | pricing | cta
industry: removals | cleaning | trades | legal | healthcare

Output

yaml

competitors:
  - url: "..."
    messaging: { ... }
    trust: { ... }
    pricing: { ... }
    ctas: { ... }
observed_differences: [ ... ]  # What varies between competitors

Hard Rules

Rule	Requirement
Competitor count	Minimum 3, maximum 5
Data source	Visual observation only
No scraping	Manual review, no HTML parsing
No recommendations	Observations only, no "should"
Absence = explicit	Missing element → `not_found: true`

Analysis Categories

1. Messaging

Field	Extract
`headline`	Exact H1 text
`headline_pattern`	benefit / feature / question / statement
`value_prop`	Main promise (1 sentence)
`tone`	formal / professional / casual / friendly
`unique_claims`	List of specific claims

2. Trust Elements

Field	Extract
`review_platform`	Google / Trustpilot / none
`rating`	Number or `not_found`
`review_count`	Number or `not_found`
`credentials`	List: type + name
`guarantees`	List: type + specifics
`team_visible`	true / false

3. Pricing

Field	Extract
`transparency`	hidden / teaser / full
`lowest_shown`	Price or `not_found`
`pricing_model`	fixed / hourly / quote-only
`discounts`	List or `not_found`

4. CTAs

Field	Extract
`primary_text`	Exact button text
`primary_placement`	hero / header / footer
`form_fields_count`	Number
`phone_prominent`	true / false
`mobile_sticky`	true / false

Handling Missing Data

Never leave fields empty. Never guess.

yaml

# If element exists
rating: 4.8

# If element not found
rating: 
  not_found: true
  
# If unclear/ambiguous
pricing_model:
  unclear: true
  observed: "Shows 'from £X' but also 'call for quote'"

Observed Differences Rules

Differences are observations, not opportunities or recommendations.

yaml

# CORRECT - observation only
observed_differences:
  - category: pricing
    observation: "Competitors A, B show no prices. Competitor C shows 'from £299'."
    
# WRONG - contains recommendation
observed_differences:
  - category: pricing
    opportunity: "Show transparent pricing to differentiate"  # ❌ NO

"Opportunity" word is FORBIDDEN in output.

Output Stability

Same input → same output structure. Always.

Scenario	Rule
Field has data	Include with value
Field not found	Include with `not_found: true`
Field unclear	Include with `unclear: true` + `observed`
Never	Omit fields

Forbidden

•❌ HTML scraping or parsing
•❌ Accessing competitor analytics
•❌ Recommendations or "should" statements
•❌ Ranking competitors (better/worse)
•❌ Omitting fields when data missing
•❌ Using word "opportunity"
•❌ Making strategic suggestions

Non-goals

•Does NOT write copy
•Does NOT rank competitors
•Does NOT make recommendations
•Does NOT interpret data
•Does NOT suggest positioning

References

•schema-messaging.md — Full messaging schema
•schema-trust.md — Full trust schema
•schema-pricing-cta.md — Full pricing & CTA schema

Definition of Done

• 3-5 competitors analyzed
• All categories filled (or marked not_found)
• No recommendations in output
• No "opportunity" word used
• Differences are observations only