Competitor Finder

Identify your competitive set and map their positioning, messaging, pricing, and channel activity. Uses Perplexity for competitive intelligence. Output is the foundation that competitor-analyzer and competitor-visual enrich.

Purpose

Competitor Finder is the starting point of competitive intelligence. It answers:

Who are we actually competing with? (direct AND indirect)
How does each competitor position themselves?
What are their core value propositions and target audiences?
What pricing models and ranges do they use?
Which marketing channels are they active on?
Where are the gaps and opportunities?

The output — research-memory/competitive-intel.md — is the skeleton that two downstream skills build upon:

competitor-analyzer (Firecrawl) adds website messaging detail: headlines, CTAs, pricing pages, social proof
competitor-visual (Playwright) adds design audit: screenshots, color palettes, layout patterns, visual tone

"You can't differentiate if you don't know what you're differentiating FROM." — The Boring Marketer

Modes

Mode	When to Use	Behavior
Full Discovery	No `competitive-intel.md` exists, or it's an empty scaffold	Run all 5 steps from scratch
Refresh	`competitive-intel.md` already has `[competitor-finder]` data	Check `research-log.md` for last scan date → add new competitors or update existing profiles

Auto-Load Protocol

On every invocation, BEFORE any research:

Check research-memory/ directory
If files exist → Read ALL .md files (except README.md)
Priority context from market-landscape.md:
- Market category definition → sets competitor search scope
- Market structure (price tiers, channels) → provides classification frame
- Market size → helps gauge competitive intensity
Check brand-memory/ (read-only) → If exists, use business description, positioning, and target audience as self-reference context
If competitive-intel.md has [competitor-finder] data AND research-log.md shows a previous scan → suggest Refresh mode

Enrichment preservation rule: When loading existing competitive-intel.md, NEVER delete or overwrite sections tagged [competitor-analyzer] or [competitor-visual]. Only touch [competitor-finder] sections.

Input Gathering

Collect from the user conversationally. Do NOT dump a form — ask naturally.

Field	Required	Description
Business / Product description	YES	What does the company/product do?
Known competitors	Optional	Starting points for competitive mapping
Business model	Optional	SaaS, service, e-commerce, education, etc. — frames pricing comparison
Geography / Market scope	Optional	Global or specific region — bounds competitor search
Focus area	Optional	Pricing comparison, messaging differentiation, channel strategy, etc.
Research intensity	Optional	"light" / "standard" / "deep" (default: deep) — 리서치 깊이와 수집량 조절
Language	Optional	결과물 작성 언어 (default: English)

If brand-memory/ exists, pre-fill business description, positioning, and target audience from voice-profile.md and positioning.md — ask user to confirm or correct.

If market-landscape.md exists, show the market category and structure as context: "Based on your market scan, I'll search for competitors in [category]. Does this look right?"

If this is a Refresh, show the current competitive set and ask: "Any new competitors to add? Or specific profiles to update?"

Research Intensity

사용자가 명시적으로 요청하면 리서치 깊이를 조절합니다. 지정하지 않으면 deep (기본값).

Level	search_context_size	수집량	용도
`light`	low	축소 (~50%)	빠른 감 잡기
`standard`	medium	보통 (~75%)	일반 리서치
`deep`	high	전체 (100%)	본격 리서치

"가볍게", "빠르게", "간단히" → light / "보통으로", "적당히" → standard / 별도 지정 없음 → deep

Process

Step 1: Load Context & Set Search Scope

Goal: Establish the competitive search frame using existing research.

Actions:

Load all research-memory/ files
Extract from market-landscape.md (if available):
- Market category → primary search term
- Price tiers → competitor classification frame
- Channels → where to look for competitors
- Adjacent markets → source of indirect competitors
Extract from brand-memory/ (if available):
- Own positioning → identify who occupies similar territory
- Target audience → find who targets the same people
If competitive-intel.md exists with data → switch to Refresh mode

Output: Clear search scope definition (category + constraints).

Step 2: Identify Competitive Set (Direct + Indirect)

Goal: Build a comprehensive list of direct and indirect competitors with URLs.

Tool: perplexity_reason (classification reasoning required)

Query pattern:

Who are the main competitors for a [business description] in the [market category] space?

Identify:
1. Direct competitors (5-8): Same target audience, similar solution, same category
2. Indirect competitors (2-3): Different approach to same problem, OR adjacent category players expanding into this space

For EACH competitor, provide:
- Company name
- Website URL (homepage)
- One-line description of what they do
- Classification: direct or indirect, with reasoning

Parameters:

search_context_size: Research Intensity에 따라 결정 (light→"low" / standard→"medium" / deep→"high")

경쟁사 수집 목표: light=3-5 direct + 1-2 indirect / standard=4-6 direct + 2 indirect / deep=5-8 direct + 2-3 indirect

If user provided known competitors: Include them in the query as starting points and ask Perplexity to validate + expand: "I already know about [names]. Who else competes in this space?"

Output: Competitive set table — names, URLs, descriptions, classifications.

Step 3: Build Competitor Profiles (Positioning, Messaging, Pricing)

Goal: Map each competitor's basic positioning, value proposition, audience, and pricing.

Tool: perplexity_ask (factual data collection)

Query pattern:

For each of these competitors in the [market category]:
[list competitor names from Step 2]

Analyze each one:
1. Positioning: How do they describe themselves? (tagline, hero headline, or elevator pitch)
2. Core value propositions: Top 2-3 benefits they emphasize
3. Target audience: Who are they built for? (job titles, company sizes, demographics)
4. Pricing model: Free / Freemium / Subscription / One-time / Enterprise
5. Price range: Approximate price points or tiers

Cite their website or recent coverage as source for each data point.

Parameters:

search_context_size: Research Intensity에 따라 결정 (light→"low" / standard→"medium" / deep→"high")

light일 경우 상위 경쟁사 위주로 프로필을 작성합니다 (전체 대상이 아닌 핵심 경쟁사 중심).

If competitors are numerous (8+): Split into two queries to avoid shallow coverage.

Output: Positioning & Messaging Matrix for all competitors.

Step 4: Channel Activity + Gap Analysis

Goal: Map competitors' marketing channel presence and identify whitespace opportunities.

Tool: perplexity_ask

Query pattern:

What marketing channels are these [market category] competitors most active on?
[list competitor names]

For each competitor, assess presence and activity level on:
- Content marketing (blog, YouTube, podcast)
- Social media (which platforms, approximate follower counts or engagement level)
- SEO / Paid advertising
- Email / Newsletter
- Community / Events

Rate each: Strong / Moderate / Weak / Absent

Then identify:
1. Underserved channels: Which channels do MOST competitors neglect?
2. Positioning whitespace: What positioning angles are unclaimed?
3. Messaging gaps: What customer needs are competitors NOT addressing in their messaging?

Parameters:

search_recency_filter: "month"
search_context_size: Research Intensity에 따라 결정 (light→"low" / standard→"medium" / deep→"high")

Output: Channel Activity Matrix + Gaps & Opportunities section.

Step 5: Save & Log

Goal: Write all findings to research-memory/competitive-intel.md and log execution.

5a. Write competitive-intel.md

Language rule: 섹션 헤더와 테이블 컬럼명은 영어로 유지합니다. 본문, 셀 값, 설명, 분석 텍스트는 사용자가 지정한 언어로 작성합니다. 언어가 지정되지 않으면 English로 작성합니다.

Use the exact schema in references/competitive-intel-schema.md. Key rules:

Tag every section you write with [competitor-finder]
Leave [competitor-analyzer] and [competitor-visual] sections as empty scaffolds with HTML comments
For Refresh: update only [competitor-finder] sections; NEVER touch other skills' sections

5b. Update research-log.md

Append one row:

| [YYYY-MM-DD] | competitor-finder | Full Discovery / Refresh | [X direct + Y indirect identified, key gaps] | Perplexity |

Perplexity MCP Tool Guide

Tool	When to Use	This Skill
`perplexity_reason`	Classification, reasoning	Step 2: Competitive set identification + direct/indirect classification
`perplexity_ask`	Factual Q&A, current data	Step 3: Competitor profiles, Step 4: Channel activity
`perplexity_search`	Find specific URLs/sources	Only if competitor website URLs need verification

Common parameters:

search_context_size: Research Intensity 레벨에 따라 결정 — 위 Research Intensity 테이블 참조
search_recency_filter: "month" for Step 4 (latest channel activity), default for Steps 2-3

Query best practices:

Include market category in every query (from market-landscape.md or user input)
Always request website URLs — competitor-analyzer and competitor-visual depend on them
Ask for source citations for positioning claims
One topic per query: Don't combine competitor identification + profiling in one call
For 8+ competitors, split profiling into two queries to ensure depth
Language: 사용자가 English 외 언어를 지정한 경우, 모든 query 끝에 "Respond in [language]."를 추가

Quality Checklist

Before saving, verify:

Example (Abbreviated)

Input: "Marketing education business selling AI-powered skill packs to solo marketers and small teams."

Direct Competitors (5):

Marketing Examined (Alex Garcia) — Newsletter-based marketing education

CopyBlogger — Content marketing & copywriting education platform

DigitalMarketer — Comprehensive digital marketing training + certifications

Demand Curve — Growth marketing education (YC-backed)

MarketingProfs — B2B marketing education + events

Indirect Competitors (2):

ChatGPT / Claude — AI directly performing marketing tasks (substitutes education)

Canva — Design tool with built-in marketing education content

Key Gaps:

No competitor offers "AI + marketing" skill packs as a combined product

Newsletter-based education is crowded; actionable skill/template bundles are rare

Most competitors weak on community-driven learning

What This Skill Does NOT Do

Website scraping / messaging detail → Use competitor-analyzer (Firecrawl)
Screenshots / design patterns → Use competitor-visual (Playwright)
Customer profiling → Use audience-profiler
Customer language mining → Use voice-of-customer
Strategic recommendations → Use research-synthesizer (reads this output)

Competitor Finder stays focused on who the competitors are and how they position at a high level — the foundation for deeper competitive analysis.