competitor-research · diff
v2.7 to v1.0
73 added, 761 removed. Audit A to A.
---
name: competitor-research
- version: "2.7"
+ version: '1.0'
+ last_updated: 2026-05-19
author: genesys-growth
- last_updated: 2026-03-31
- description: |
- Use when client needs deep analysis of a specific competitor.
- Use when positioning, sales-enablement, or product-messaging need competitive context.
- Use when user mentions competitor name or asks for "competitive analysis".
- Use when preparing battlecard content or competitive positioning.
- Use with --refresh flag to update an existing competitor dossier (living document mode).
-
- # Agentic Layer Fields
- dependencies:
+ description: Run deep research on a single competitor. Produces a dossier with executive summary, position + messaging, status-quo alternatives, key differentiators, pricing position, where-they-win / where-they-lose, threat assessment, and what to watch. Writes to marketing/competitors/MMYY-{competitor-slug}.md. Triggers - "competitor research", "competitive analysis", "research competitor X", "battlecard prep", "analyze [competitor name]"
+ goal: Produce a deep dossier on a single competitor: positioning, messaging, status-quo, differentiators, pricing, where-they-win/lose, threat assessment.
+ outcome: marketing/competitors/MMYY-{competitor-slug}.md with executive summary + 11 dimensions of analysis.
+ primitive: research
+ ontology_type: competitor-intel
+ review_gate: 2
+ inputs:
required: []
recommended:
- - company-context
-
+ []
outputs:
- - type: competitor-profile
- feeds_into:
- - positioning
- - sales-enablement
- - product-messaging
- - type: comparison-matrix
- feeds_into:
- - positioning
- - sales-enablement
- - type: aggregate-insights
+ - type: competitor-intel
feeds_into:
- positioning
- product-messaging
+ - competitor-aggregate
- sales-enablement
-
+ owned_by_agent: researcher
+ mcps_used:
+ - exa
triggers:
- auto_suggest_when:
- - "user mentions competitor name or competitive analysis"
- - "positioning skill needs competitive context"
- - "sales-enablement needs battlecard content"
- - "user has completed 2+ competitor deep dives for same client"
- - "user asks for 'aggregate analysis' or 'cross-competitor insights'"
- - "user asks to 'refresh' or 'update' competitor research"
- auto_run_when: null # Always require confirmation
-
- run_modes:
- - quick-scan # Weekly: news, G2 check
- - deep-refresh # Monthly: full 13-dimension research + aggregate
-
- review_gate: 1 # Quick review (research output)
-
- output_path: "context/"
+ slash_commands:
+ - /competitor-research
+ status: draft
---
- # Competitor research
-
- Run deep competitor analysis across 13 research dimensions for B2B SaaS companies. Produces structured insights with explicit confidence levels, source citations, and actionable data gaps.
-
- **Knowledge type:** `competitor-intel`
- **Maturity on first run:** emergent → validated after client review
-
- ## Attribution standard
-
- **Inline attribution (simplified):**
-
- Use concise inline source references for key data points. Do NOT use the verbose `[VERIFIED: url, date]` block syntax inline — it clutters the output.
-
- - `(Source: Crunchbase)` — direct named source
- - `(Source: Vendr)` — Vendr contract data (pricing fallback)
- - `(Source: competitor website)` — direct competitor page
- - `(Source: Reddit, r/[subreddit])` — community signal, treat as Low confidence unless corroborated
- - `(Source: G2)` — review platform data
- - Omit attribution only for facts that are self-evidently sourced (e.g., a feature listed on the competitor's features page)
-
- **End-of-document audit trail:**
-
- Every output ends with a **Sources & data quality** section (see output template). This table consolidates:
- - All URLs with access dates
- - Confidence level per dimension
- - Data gaps with unverified claims and follow-up actions
+ # competitor-research — Stage 1 research skill
- **Quality threshold:** minimum 50% verified claims, maximum 20% estimated.
+ The first skill in the article's Example 1 chain. Reads a competitor URL + name, produces a structured dossier, writes the file to `marketing/competitors/MMYY-{slug}.md`. Run once per competitor (3-5 competitors is typical for the first cycle). After all per-competitor files exist, run `/competitor-aggregate` to synthesize the canonical threat matrix.
---
- ## Run Modes
-
- This skill supports two run modes for different cadences:
-
- ### Quick scan (weekly)
-
- **Duration:** ~30 min | **Trigger:** `/competitor-research --quick [competitor]`
-
- Refreshes fast-moving data only. Does NOT re-research all 13 dimensions.
-
- **Steps:**
- 1. Read existing competitor output file
- 2. Search the web for recent news — `"[competitor]" announcement OR launch OR funding` (last 7 days)
- 3. Check for new blog posts since last refresh
- 4. G2 rating check — has rating or review count changed?
- 5. Compare findings against existing output
- 6. Update "Recent Changes" section at top with date-stamped entries
- 7. Refresh only the sections where data changed (Company, Launches, Reviews)
- 8. Update "Last refreshed" date
-
- **Output:** Updated competitor file with new "Recent Changes" entries. Sections without changes remain untouched.
-
- ### Deep refresh (monthly)
-
- **Duration:** ~90 min | **Trigger:** `/competitor-research --refresh [competitor]`
-
- Full 13-dimension research cycle. Rewrites entire output with latest data.
-
- **Steps:**
- 1. Read existing competitor output file
- 2. Run full Phase 1-3 process (all 13 dimensions)
- 3. TrustPilot customer monitoring (optional, for B2B SaaS with customer-facing products)
- 4. Compare ALL findings against existing output
- 5. Produce "Recent Changes" section highlighting what's different from last refresh
- 6. Rewrite all sections with latest data
- 7. Archive change entries older than 4 weeks to bottom
- 8. Run Phase 4 aggregate analysis if 2+ competitors refreshed this cycle
- 9. Update "Last refreshed" date
-
- **Output:** Fully refreshed competitor dossier with change log.
-
- ### First run (new competitor)
-
- **Duration:** ~90 min | **Trigger:** `/competitor-research [competitor]` (no --refresh flag)
-
- Standard full process — Phases 1-3 (or 1-4 for aggregate). No existing file to compare against.
-
- ---
+ ## When to use
- ## Living Dossier Refresh Protocol
+ - Day 1 of the article's Example 1 ("foundational workflow"): you have a list of 3-5 competitors and need structured intel on each
+ - Before refreshing `/positioning` — positioning must read from current competitor intel
+ - Before writing battlecards or comparison pages
+ - Quarterly refresh of any specific competitor that's made a major move (funding, launch, executive change)
- When running in `--quick` or `--refresh` mode:
+ ## When NOT to use
- 1. **Read first.** Before any research, read the existing competitor output file. This is your baseline.
- 2. **Compare, don't append.** New findings replace old data in the relevant section. Don't just add — update.
- 3. **"Recent Changes" at the top.** Every refresh adds date-stamped entries to the top of the file:
- ```
- ## Recent changes
+ - Aggregating across multiple competitors (use `/competitor-aggregate` after this skill produces per-competitor files)
+ - Building battlecards (use `/sales-enablement` — different shape, sales-team-facing)
+ - ICP research (use `/icp-research`)
+ - General company background unrelated to competition (use `/company-context`)
- **2026-03-31:** Headcount grew from 150 → 180 (Source: LinkedIn). New blog post on AI guardrails.
- **2026-03-24:** Series C announced — $45M led by Sequoia (Source: TechCrunch).
- **2026-03-17:** G2 rating dropped from 4.6 to 4.4 — 3 new negative reviews about onboarding.
- ```
- 4. **Archive after 4 weeks.** Move entries older than 4 weeks to "Archived change logs" section at the bottom.
- 5. **"Last refreshed" date** at the top of every output.
- 6. **No-change = no entry.** If a dimension hasn't changed, don't add a change log entry for it.
+ ## How it works
- ---
+ 1. Inputs: competitor name + competitor URL (required). Optional: industry context, key differentiator hypothesis you want to test.
+ 2. Reads (optional Exa MCP for fresh web data): competitor homepage, pricing page, about page, recent blog posts, recent LinkedIn activity.
+ 3. Produces a 12-section dossier (matching the PulseAnalytics example at `marketing/competitors/0526-competitor-a.md`):
+ - Executive summary (3-4 sentences)
+ - Anchor + category
+ - Position + messaging
+ - Status-quo alternatives they position against
+ - Key differentiators (claimed)
+ - Pricing position
+ - Where they win
+ - Where they lose (your wedge)
+ - Threat assessment (PRIMARY / DIRECT / STEALTH / LOW / DEFUNCT)
+ - Win-loss patterns (from your deal history if available)
+ - What to watch
+ - Refresh cadence + owner
+ 4. Every claim tagged with confidence per `.claude/rules/ontology.md`: `[VERIFIED: source, url, date]`, `[INFERRED: from X + Y]`, `[ESTIMATED: reasoning]`, `[UNAVAILABLE]`.
+ 5. Writes the file to `marketing/competitors/MMYY-{competitor-slug}.md`. MMYY is current month+year (e.g., `0526` for May 2026).
- ## Process Flowchart
+ ## Invoke
- ```text
- ┌──────────────────────────────────────────────────────────────┐
- │ COMPETITOR RESEARCH PROCESS │
- └──────────────────────────────────────────────────────────────┘
- │
- ▼
- ┌──────────────────────────────────────────────────────────────┐
- │ PHASE 1: INPUT VALIDATION │
- │ □ Required: Competitor name, Website URL │
- │ □ Optional: Market category, client context, specific Qs │
- │ → If ambiguous name: Confirm competitor identity with user │
- └──────────────────────────────────────────────────────────────┘
- │
- ▼
- ┌──────────────────────────────────────────────────────────────┐
- │ PHASE 2: DIMENSION RESEARCH (13 dimensions) │
- │ Step 2.1: Company → funding, team, revenue │
- │ Step 2.2: Product → features, differentiators │
- │ Step 2.3: ICP → segments, personas, customers │
- │ Step 2.4: Pricing → plans, model, value metric │
- │ Step 2.5: Reviews → G2, Capterra, sentiment │
- │ Step 2.6: Content → blog, formats, lead magnets, events │
- │ Step 2.7: Launches → recent announcements │
- │ Step 2.8: SEO/AEO → SERP positions │
- │ Step 2.9: Technographics → integrations (limited) │
- │ Step 2.10: Openings → hiring signals │
- │ Step 2.11: GTM → sales motion, outbound signals, messaging │
- │ Step 2.12: LinkedIn/Social → organic strategy, founder │
- │ Step 2.13: Paid advertising → LinkedIn/Meta/Google ads │
- │ ✓ Checkpoint: All dimensions researched, sources documented │
- └──────────────────────────────────────────────────────────────┘
- │
- ▼
- ┌──────────────────────────────────────────────────────────────┐
- │ PHASE 3: SYNTHESIS & GAPS │
- │ Step 3.1: Assign confidence levels (High/Med/Low) │
- │ Step 3.2: Write executive summary │
- │ Step 3.3: Document data gaps with follow-up actions │
- │ Step 3.4: Run quality verification │
- │ ✓ Checkpoint: No unsourced claims, gaps documented │
- └──────────────────────────────────────────────────────────────┘
- │
- ▼
- ┌──────────────────────────────────────────────────────────────┐
- │ PHASE 4: AGGREGATE ANALYSIS (Multi-Competitor) │
- │ Trigger: 2+ competitors researched for same client │
- │ Step 4.1: Identify cross-competitor patterns │
- │ Step 4.2: Build threat matrix │
- │ Step 4.3: Analyze market positioning dynamics │
- │ Step 4.4: Feature parity analysis │
- │ Step 4.5: Credibility signal audit │
- │ Step 4.6: Extract strategic recommendations │
- │ ✓ Checkpoint: Aggregate insights documented │
- └──────────────────────────────────────────────────────────────┘
- │
- ▼
- ┌──────────────────────────────────────────────────────────────┐
- │ SELF-EVALUATION │
- │ □ Completeness: All 13 dimensions addressed? │
- │ □ Evidence: Every data point has source or "Not available"? │
- │ □ Guardrails: No invented data? No unsourced claims? │
- │ → If issues found: Flag low-confidence areas │
- └──────────────────────────────────────────────────────────────┘
- │
- ▼
- ┌──────────────────────────────────────────────────────────────┐
- │ REVIEW GATE: Level 1 (Quick Review) │
- │ Present: Executive summary, confidence breakdown │
- │ Actions: [Approve] [Dig deeper on X] [Research more] │
- └──────────────────────────────────────────────────────────────┘
- │
- ▼
- ┌──────────────────────────────────────────────────────────────┐
- │ CHAIN SUGGESTIONS │
- │ → "Want me to research additional competitors?" │
- │ → "Ready to run positioning with this competitive context?" │
- │ → "Should I generate battlecard content from this?" │
- │ → "Ready for aggregate analysis across all competitors?" │
- │ → Save as reference example if positive feedback │
- └──────────────────────────────────────────────────────────────┘
```
-
- ---
-
- ## The Iron Law
-
- **NO DATA POINT WITHOUT SOURCE.**
-
- Every claim has a URL + access date. Every metric cites its origin. "Not available" is always better than invented data.
-
- **No exceptions:**
-
- - "This is widely known" → Still needs a source. Cite it.
- - "I read this somewhere" → Find the URL or mark "Not available".
- - "I can estimate this" → Estimates need explicit confidence levels and reasoning.
- - "The user needs this fast" → Fast + wrong = useless. Take the time.
-
- ---
-
- ## Red Flags (Stop and Verify)
-
- About to write revenue/funding without source → STOP. Cite Crunchbase, PitchBook, or mark "[Data needed]".
-
- About to state competitor feature without verifying → STOP. Check their website/docs first.
-
- About to invent customer logos → STOP. Only include logos you found on their site.
-
- About to guess pricing → STOP. Research their pricing page or mark "Pricing not public".
-
- About to claim market position without evidence → STOP. Cite G2, analyst report, or mark confidence level.
-
- ---
-
- ## Claude Code Triggers
-
- **Invoke this skill when user says:**
- - "Research this competitor: [name/URL]"
- - "Competitor analysis for [company]"
- - "Competitive intelligence on [company]"
- - "Analyze [company] as a competitor"
- - "Battlecard research for [competitor]"
- - "Competitive landscape for [market]"
- - "Compare [competitor 1] vs [competitor 2]"
- - "What's [competitor] doing?"
-
- **Do NOT invoke when:**
- - User wants company qualification/traction → Use `company-context` skill
- - User wants product messaging only → Use `product-messaging` skill
- - User wants to research own company → Use `company-context` or `product-messaging`
- - Quick question about one feature → Answer directly without full 13-dimension framework
-
- ---
-
- ## Input Requirements
-
- ### Required Inputs
-
- | Input | Description | Source |
- |-------|-------------|--------|
- | **Competitor name** | Exact company/product name | User provides |
- | **Website URL** | Competitor's website | User provides or confirm |
-
- ### Optional Inputs (improve quality)
-
- | Input | How It Helps |
- |-------|--------------|
- | Market category | Helps with disambiguation and keyword selection |
- | Client context | Current positioning helps sharpen analysis |
- | Specific questions | Focus research on areas of interest |
- | Research depth | Deep (default) or comparison matrix |
-
- ### Input Validation Checklist
-
- Before proceeding, verify:
- - [ ] Competitor name is unambiguous (if not, confirm with user)
- - [ ] Website URL is correct and accessible
- - [ ] Research mode confirmed (single deep dive or comparison matrix)
-
- **If inputs are missing:** Ask for website URL. If name is ambiguous (e.g., "Bolt" could be multiple companies), confirm which competitor.
-
- ---
-
- ## Process (Step-by-Step)
-
- ### Phase 1: Confirmation & Setup
-
- **Purpose:** Verify competitor identity and establish research scope.
-
- **Steps:**
-
- 1. **Step 1.1: Confirm competitor identity**
- - Verify exact company/product name
- - Confirm website URL is correct
- - Check for name disambiguation issues
- - **Output:** Confirmed competitor details
-
- 2. **Step 1.2: Determine research mode**
- - Default: Single competitor deep dive (13 dimensions)
- - Alternative: Comparison matrix (3-6 competitors, core dimensions)
- - **Output:** Research mode confirmed
-
- **Phase 1 Checkpoint:**
- - [ ] Competitor name unambiguous
- - [ ] Website URL verified
- - [ ] Research mode selected
-
- ### Phase 2: Dimension Research
-
- **Purpose:** Systematically research all 13 dimensions.
-
- **Steps:**
-
- 1. **Step 2.1: Company research (Dimension 1)**
- - Search the web for company data: `"[competitor]" crunchbase funding`, `"[competitor]" series raised valuation`
- - Fetch about page from competitor website
- - Extract: Founded, HQ, team size, funding, revenue, headcount growth trend, recent news
- - **Output:** Company profile with sources
-
- 2. **Step 2.2: Product research (Dimension 2)**
- - Fetch homepage and features page
- - Fetch documentation if available
- - Extract: Core description, key features, differentiators, platform
- - **Output:** Product profile with sources
-
- 3. **Step 2.3: ICP research (Dimension 3)**
- - Fetch customers/case studies page
- - Fetch solutions/industries page
- - Search: `site:reddit.com "[competitor]" use case who uses` — validate segments and uncover niche use cases not on their website
- - Extract: Company size, industries, personas, geos, named customers
- - **Output:** ICP profile with sources
-
- 4. **Step 2.4: Pricing research (Dimension 4)**
- - **Step 1:** Fetch `/pricing` page directly
- - **Step 2 (if hidden/gated):** Search `site:g2.com "[competitor]" pricing` — extract plan names, price ranges from review snippets
- - **Step 3 (if still missing):** Search `site:vendr.com "[competitor]"` — Vendr publishes negotiated SaaS contract data; mark findings as `(Source: Vendr)`
- - **Step 4 (if still missing):** Search `site:reddit.com "[competitor]" pricing cost per seat` — extract user-reported pricing, negotiation outcomes, plan details; mark as `(Source: Reddit)`, Low confidence
- - **Step 5 (if all fail):** Mark "Pricing not public — pricing page, G2, Vendr, and Reddit checked"
- - Extract: Model, plans, free tier, enterprise availability, value metric
- - **Output:** Pricing profile with sources
-
- 5. **Step 2.5: Reviews research (Dimension 5)**
- - Search: `site:g2.com "[competitor]" reviews`
- - Search: `site:capterra.com "[competitor]" reviews`
- - Search: `site:reddit.com "[competitor]" review` — unfiltered sentiment, bugs, feature complaints
- - Search: `site:reddit.com "[competitor]" problems issues` — surfaces pain points not visible on review platforms
- - Extract: Ratings, review count, sentiment themes, pros/cons, Reddit community signals
- - Note: Reddit findings = Low confidence unless corroborated by G2/Capterra patterns
- - **TrustPilot customer monitoring (optional, deep refresh only):**
- - Build list of competitor's known customers from case studies page and G2 reviewers
- - For top 10-15 high-profile customers, check TrustPilot: `"[customer name]" site:trustpilot.com`
- - Look for reviews mentioning "customer support", "chat", "bot", "AI", "automated"
- - Score each customer: Green Healthy / Yellow Warning / Red Problem
- - Flag Red customers prominently — these are direct proof points for sales battlecards
- - Skip this step if competitor doesn't deploy customer-facing products
- - **Output:** Reviews profile with sources + TrustPilot monitor table (if applicable)
-
- 6. **Step 2.6: Content research (Dimension 6)**
- - Fetch blog and resources page
- - Extract: Topics, formats, frequency, lead magnets
- - **Output:** Content profile with sources
-
- 7. **Step 2.7: Launches research (Dimension 7)**
- - Search: `"[competitor]" launch announcement [current year]`
- - Search: `site:producthunt.com "[competitor]"`
- - Extract: Product launches, feature announcements (last 3 months)
- - **Output:** Launches profile with sources
-
- 8. **Step 2.8: SEO/AEO research (Dimension 8)**
- - Run 5-10 SERP checks for key category keywords
- - Query patterns: `"[category keyword]"`, `"[competitor name] vs [category]"`, `"best [category] tool"`
- - Extract: competitor ranking position, featured snippet ownership, title/meta copy
- - Use `site:[competitor.com]` search to estimate indexed page count
- - If you have access to SEO tools (Ahrefs, Semrush), pull domain rating, organic traffic estimate, top keywords, and referring domains for higher confidence data
- - **Output:** SEO profile with confidence level based on data source
-
- 9. **Step 2.9: Technographics research (Dimension 9)**
- - Fetch integrations page
- - Check job postings for tech stack hints
- - Note: Full data requires BuiltWith access
- - Extract: Integrations, visible tech signals
- - **Output:** Technographics profile (limited without premium tools)
-
- 10. **Step 2.10: Openings research (Dimension 10)**
- - Fetch careers page
- - Search: `site:linkedin.com/jobs "[competitor]"`
- - Extract: Open roles, hiring departments, seniority, locations
- - **Output:** Openings profile with sources
-
- 11. **Step 2.11: GTM research (Dimension 11)**
- - Analyze website messaging and CTAs
- - Check founder LinkedIn activity
- - Search job postings for SDR/BDR/outbound roles — mention of sequencing tools (Outreach, Salesloft, Apollo) signals active outbound
- - Search: `site:reddit.com "[competitor]" sales demo cold email` — surfaces buyer-reported sales experience
- - Extract: Sales motion (PLG/sales-led/hybrid), primary CTA, channel mix, outbound signals, messaging themes
- - **Output:** GTM profile with sources
-
- 12. **Step 2.12: LinkedIn / Social research (Dimension 12)**
- - Check linkedin.com/company/[competitor] — capture followers, about, recent posts
- - Search: `"[competitor]" site:linkedin.com/posts` for recent post titles and topics
- - Check founder/CMO LinkedIn profiles for post cadence and themes
- - Extract: Follower count + YoY growth, post frequency, content types, founder activity
- - Note: Impression and reach data require LinkedIn Analytics access — mark [UNAVAILABLE]
- - **Output:** LinkedIn/Social profile with sources
-
- 13. **Step 2.13: Paid advertising research (Dimension 13)**
-
- **IMPORTANT:** All three ad libraries are JS-rendered — raw HTML fetches will fail or return empty. Use browser-based scraping tools if available.
-
- **Google Ads Transparency Center (highest priority for B2B SaaS):**
- - Navigate to `https://adstransparency.google.com/?region=anywhere`
- - Search for competitor domain
- - Extract: active campaign count, ad copy themes, CTA patterns, geographic targeting
- - **If scraping fails:** Mark "Google Ads data unavailable — JS wall. Manual check at adstransparency.google.com required"
-
- **LinkedIn Ads Library:**
- - Check: `https://www.linkedin.com/ad-library/search?companyIds=[handle]`
- - Extract: ad types, creative themes, CTA patterns, running dates
- - **If no ads found:** Mark as verified absence — "No active LinkedIn Ads found (verified YYYY-MM-DD)"
-
- **Meta Ads Library:**
- - Fetch: `https://www.facebook.com/ads/library/?q=[competitor name]&search_type=keyword_unordered&active_status=active`
- - Extract: active ad count, creative formats, copy themes, CTA patterns
-
- - "No active ads found" is a high-confidence data point — treat as verified GTM signal, not a data gap
- - **Output:** Paid advertising profile per platform with confidence levels
-
- **Phase 2 Checkpoint:**
- - [ ] All 13 dimensions researched
- - [ ] Sources documented for each finding
- - [ ] "Not available" noted for missing data
-
- ### Phase 3: Synthesis & Gaps
-
- **Purpose:** Compile findings and document limitations.
-
- **Steps:**
-
- 1. **Step 3.1: Assign confidence levels**
- - High: Primary source, recent, verifiable
- - Medium: Secondary source, analyst estimate
- - Low: Inferred, single unverified source
- - **Output:** Confidence levels assigned
-
- 2. **Step 3.2: Compile executive summary**
- - 2-3 paragraphs summarizing key findings
- - Highlight competitive strengths and weaknesses
- - **Output:** Executive summary
-
- 3. **Step 3.3: Document data gaps**
- - What couldn't be found
- - What requires premium tools
- - Suggested follow-up actions
- - **Output:** Data gaps section
-
- 4. **Step 3.4: Quality verification**
- - Run quality checklist
- - Verify no unsourced claims
- - Confirm surprising claims corroborated
- - **Output:** Verified output
-
- **Phase 3 Checkpoint:**
- - [ ] Confidence levels assigned
- - [ ] Executive summary written
- - [ ] Data gaps documented
- - [ ] No unsourced claims
-
- ### Phase 4: Aggregate Analysis (Multi-Competitor)
-
- **Purpose:** Synthesize cross-competitor patterns into strategic intelligence after completing 2+ individual deep dives for the same client.
-
- **Trigger conditions:**
- - 2+ competitors researched for the same client project
- - User explicitly requests aggregate analysis or cross-competitor insights
-
- **Steps:**
-
- 1. **Step 4.1: Pattern identification**
- - Read all competitor deep dives in the client's competitors folder
- - Extract recurring themes across dimensions
- - Identify market dynamics
- - **Output:** Cross-competitor patterns summary
-
- 2. **Step 4.2: Threat matrix construction**
- - Rank competitors by threat level: PRIMARY / ENTERPRISE TIER / DIRECT ICP / STEALTH WATCH / LOW / DEFUNCT
- - Justify each ranking with evidence
- - Identify exploitable vulnerabilities per competitor
- - **Output:** Threat matrix table
-
- 3. **Step 4.3: Market positioning dynamics**
- - Map competitors on relevant spectrums (SMB↔Enterprise, Organic↔VC/PR, Feature depth, etc.)
- - Identify white space and crowded zones
- - **Output:** Positioning maps with narrative analysis
-
- 4. **Step 4.4: Feature parity analysis**
- - Build capability comparison table (all competitors × key features)
- - Classify each feature by market maturity: Commoditized / Differentiator / Emerging
- - **Output:** Feature parity matrix with maturity classification
-
- 5. **Step 4.5: Credibility signal audit**
- - Compare funding, awards, press, enterprise logos, founder visibility, content strategy
- - Identify where the client is ahead and behind on credibility signals
- - **Output:** Credibility comparison table with gap analysis
-
- 6. **Step 4.6: Strategic recommendations**
- - Extract 3-5 actionable recommendations from pattern analysis
- - Each recommendation includes: rationale, evidence, suggested actions, effort/impact scoring
- - **Output:** Strategic recommendations with evidence
-
- **Phase 4 Checkpoint:**
- - [ ] All individual deep dives referenced and cross-checked
- - [ ] Threat rankings justified with evidence (not opinion)
- - [ ] Feature parity table uses "Unknown" where client capabilities are unverified
- - [ ] Strategic recommendations are actionable (not generic "be better")
-
- ---
-
- ## Core Frameworks
-
- ### 13 Research Dimensions
-
- | # | Dimension | What to Find | Primary Sources |
- |---|-----------|--------------|-----------------|
- | 1 | **Company** | Revenue, funding, team size, HQ | Crunchbase, Tracxn, LinkedIn |
- | 2 | **Product** | Features, capabilities, differentiators | Website, docs, G2 |
- | 3 | **ICP** | Segments, geos, company sizes | Case studies, solutions pages |
- | 4 | **Pricing** | Plans, amounts, value metrics | Pricing page, G2 |
- | 5 | **Reviews** | Ratings, sentiment, themes | G2, Capterra, TrustRadius |
- | 6 | **Content** | Topics, formats, cadence, owned events | Blog, resources, LinkedIn, event pages |
- | 7 | **Launches** | Recent product/feature announcements | LinkedIn, Product Hunt, press releases |
- | 8 | **SEO/AEO** | Rankings, traffic, gaps | SERP analysis (limited) |
- | 9 | **Technographics** | Tech stack, integrations | Integrations page (limited) |
- | 10 | **Openings** | Job postings, hiring signals | Careers, LinkedIn Jobs |
- | 11 | **GTM** | Sales motion, inbound/outbound mix, outbound signals, messaging | Website, job descriptions, LinkedIn |
- | 12 | **LinkedIn / Social** | Organic posting strategy, content types, frequency, founder activity, follower count | LinkedIn company page, founder profiles |
- | 13 | **Paid advertising** | Active ad campaigns, creative themes, ICP signals in copy, CTAs, spend signals | LinkedIn Ads Library, Meta Ads Library, Google Ads Transparency Center |
-
- ### Confidence Scoring
-
- | Level | Definition | Example |
- |-------|------------|---------|
- | **High** | Primary source, recent, verifiable | Company blog, press release, official pricing page |
- | **Medium** | Secondary source, analyst estimate | Sacra estimates, TechCrunch reporting, Tracxn data |
- | **Low** | Inferred, single unverified source | Forum posts, old articles, indirect mentions |
-
- ### Competitor Disambiguation
-
- Many company names are ambiguous. Common issues:
-
- | Ambiguous Name | Possible Meanings |
- |----------------|-------------------|
- | Bolt | Ride-sharing, Fintech, Bolt.new (AI coding) |
- | Cursor | AI code editor, Design tool |
- | Base | Database, Base44 (AI app builder) |
- | Linear | Issue tracking, Other |
-
- **Resolution approach:**
- 1. Use domain-qualified searches: "bolt.new" not "bolt"
- 2. Include category context: "Cursor AI coding" not "Cursor"
- 3. Check for parent companies: Bolt.new is part of StackBlitz
- 4. When in doubt, confirm with user before starting
-
- ---
-
- ## Output Format
-
- ### Standard Output Structure (Single Competitor)
-
- **IMPORTANT**: All competitor files must follow this exact structure. The title is always `# Competitor research: [Name]`.
-
- ```markdown
- # Competitor research: [Competitor Name]
-
- **Website**: [URL]
- **Category**: [Market category]
- **Research date**: [YYYY-MM-DD]
- **Last refreshed**: [YYYY-MM-DD]
-
- ---
-
- ## Recent changes
-
- **[YYYY-MM-DD]:** [Change description] (Source: [source])
-
- ---
-
- ## Executive summary
-
- [2-3 paragraphs summarizing key findings, competitive strengths, weaknesses, and strategic implications]
-
- ---
-
- ## Confidence summary
-
- | Confidence level | Data points |
- |------------------|-------------|
- | Verified | [N] |
- | Inferred | [N] |
- | Estimated | [N] |
- | Unavailable | [N] |
-
- ---
-
- ## 1. Company
-
- | Metric | Finding | Source |
- |--------|---------|--------|
- | Founded | [Year] | (Source: Crunchbase) |
- | HQ | [Location] | (Source: LinkedIn) |
- | Team size | [Count] | (Source: LinkedIn) |
- | Total funding | [Amount] | (Source: Crunchbase) |
- | Latest round | [Stage, amount, date] | (Source: TechCrunch) |
- | Revenue estimate | [Amount/range] | (Source: Sacra) |
-
- ## 2. Product
- ## 3. ICP
- ## 4. Pricing
- ## 5. Reviews
- ## 6. Content
- ## 7. Launches
- ## 8. SEO/AEO
- ## 9. Technographics
- ## 10. Openings
- ## 11. GTM
- ## 12. LinkedIn / Social
- ## 13. Paid advertising
-
- ## Sources & data quality
-
- | Dimension | Source | URL | Access date | Confidence |
- |-----------|--------|-----|-------------|------------|
-
- ### Data gaps & unverified claims
-
- | Dimension | Claim or gap | Why unverified | Follow-up action |
- |-----------|--------------|----------------|------------------|
-
- ---
-
- ## Related context
-
- **Built from:**
- - `context/MMYY-company-context.md` (comparison baseline)
- - Competitor website, public sources
-
- **Feeds into:**
- - `marketing/product-marketing/MMYY-positioning.md` (differentiation)
- - `marketing/product-marketing/MMYY-product-messaging.md` (competitive alternatives)
+ /competitor-research Bizible — https://bizible.com
```
- [See full output template in Process section above for detailed structure per dimension]
-
- ---
-
- ## Anti-Hallucination Guardrails
-
- 1. **Never invent data.** Mark "Not available" or "Unknown" if not found
- 2. **Always cite sources inline.** Use `(Source: name)` for key data points
- 3. **Consolidate sources at the end.** Full URLs, access dates, and confidence levels go in the "Sources & data quality" table
- 4. **Acknowledge premium tool limits.** SEO and technographics need paid tools — state which tool was used
- 5. **Verify surprising claims.** If data seems unusual, search for corroboration
- 6. **Explicit gaps over false data.** "Not available" is always better than guessing
- 7. **Reddit is Low confidence by default.** Corroborate with G2/Capterra before upgrading to Medium
- 8. **Vendr pricing = Medium confidence.** User-negotiated data may differ from list price — always note this
- 9. **Ads "no data found" is not a data gap.** If scraping returns no ads, mark as "no active ads verified" — it's a high-confidence GTM signal
-
- ### Missing Data Labels
-
- | Label | When to Use |
- |-------|-------------|
- | "Not available" | Data doesn't exist publicly |
- | "Unknown" | Data may exist but not found |
- | "Requires premium tool" | Data exists but needs paid access (Ahrefs, BuiltWith) |
- | "Conflicting sources" | Multiple sources disagree |
-
- ---
-
- ## Quality Checklist (Pre-Delivery)
-
- ### Data Quality
- - [ ] Every data point has source URL or explicit "Not available"
- - [ ] Confidence levels assigned to key metrics
- - [ ] No invented or estimated numbers without clear labeling
- - [ ] Disambiguation confirmed for ambiguous names
- - [ ] Surprising claims corroborated with second source
-
- ### Coverage Quality
- - [ ] All 13 dimensions addressed (single competitor mode)
- - [ ] Core 5 dimensions addressed (comparison matrix mode)
- - [ ] Executive summary provides strategic insight
- - [ ] Data gaps section completed with follow-up actions
-
- ### Format Quality
- - [ ] Output header with date
- - [ ] Tables properly formatted
- - [ ] Confidence summary included
- - [ ] Freshness noted for time-sensitive data
-
- ---
-
- ## Self-Evaluation Protocol
-
- After generating competitor research output, automatically run these checks:
-
- ### Completeness Check
- - [ ] All 13 dimensions addressed?
- - [ ] Executive summary provides strategic insight?
- - [ ] Data gaps section complete with follow-up actions?
-
- ### Evidence Quality Check
- - [ ] Every data point has source URL or explicit "Not available"?
- - [ ] Confidence levels assigned to all key metrics?
- - [ ] No invented or estimated numbers without clear labeling?
-
- ### Guardrail Check
- - [ ] No invented data? Every claim sourced?
- - [ ] Disambiguation confirmed for ambiguous names?
- - [ ] Surprising claims corroborated with second source?
-
- ### Self-Roast Questions
+ Or open-ended:
+ ```
+ /competitor-research analyze our top competitor — Bizible — and write the dossier
+ ```
- 1. "If the user cross-checked this research, would any claims fail verification?"
- 2. "Am I being honest about data gaps, or hiding them in vague language?"
- 3. "Does the executive summary provide real strategic insight or just summarize facts?"
- 4. "Would a competitive intelligence analyst consider this thorough?"
- 5. "Are my confidence levels accurate, or am I being overconfident?"
+ ## Example output (against the PulseAnalytics example seed)
- ---
+ See [`marketing/competitors/0526-competitor-a.md`](../../../marketing/competitors/0526-competitor-a.md) for a fully-populated example of what this skill produces. Notice:
- ## Integration with Other Skills
+ - Anchor + threat level surfaced in the executive summary
+ - Claims tagged with confidence levels
+ - `[UNAVAILABLE]` notation where real data isn't accessible to the skill (e.g., your private win-loss data)
+ - Refresh cadence + owner at bottom
- ### Competitor-research feeds INTO these skills
+ ## Dependencies
- | Skill | What It Receives | How It Uses It |
- |-------|------------------|----------------|
- | **positioning** | Competitive alternatives, market context | Alternative mapping, differentiation angles |
- | **product-messaging** | Competitor weaknesses, positioning gaps | Sharpens client messaging, contrast points |
- | **sales-enablement** | Full competitor profiles, weaknesses | Battlecard content, objection handling |
+ - **Reads from:** `marketing/positioning/positioning.md` (optional — gives the skill context on what differentiation matters); `marketing/icp/ICP.md` (optional — informs "where they win vs. lose" framing).
+ - **Reads via Exa MCP (optional):** competitor's website, recent news, public reviews on G2/Reddit/Trustpilot.
+ - **Writes to:** `marketing/competitors/MMYY-{slug}.md`
- ### Competitor-research receives FROM these skills
+ ## Customization
- | Skill | What It Provides | How Competitor-research Uses It |
- |-------|------------------|--------------------------------|
- | **company-context** | Client context, market position | Provides comparison baseline |
+ Change the dimension count or section structure by editing this SKILL.md body. The 12-section default tracks the Genesys ontology (`.claude/rules/ontology.md` § competitor-intel). Add a custom dimension (e.g., "Open-source posture" if you're in a category where that matters) by appending a new section after section 8.
- ### Recommended workflow sequences
+ ## Where this fits in the Example 1 chain
- **Sequence 1: Comprehensive competitive analysis**
- ```text
- company-context → competitor-research → positioning
```
-
- **Sequence 2: Full PMM stack**
- ```text
- icp-research + competitor-research → positioning → product-messaging
+ Day 1-3: /competitor-research × N (this skill) → per-competitor files in marketing/competitors/
+ Day 4: /competitor-aggregate → synthesizes per-competitor files into the canonical threat matrix
+ Day 5: /icp-research → reads competitor outputs to inform ICP
```
- ---
+ After all 3 competitor files exist, run `/competitor-aggregate` to lock the canonical version.
- ## Changelog
+ ## Refresh cadence
- | Version | Date | Changes |
- |---------|------|---------|
- | 2.7 | 2026-03-31 | Living dossier: added run modes, refresh protocol, TrustPilot monitoring |
- | 2.4 | 2026-03-17 | Expanded from 11 → 13 dimensions |
- | 2.3 | 2026-02-06 | Phase 4: Aggregate Analysis |
- | 2.0 | 2026-01-16 | Refactored to v2.0: structured phases, confidence scoring |
- | 1.0 | Original | Initial skill creation with 11 dimensions |
+ Per competitor: quarterly minimum. Trigger sooner on funding, launch, or repositioning. Use the `--refresh` flag (when you implement it) to update an existing dossier rather than create a new file.