# Product Specification — CE Competitive Research

## Input
- A URL (any website — submitted by any user)

---

## Stage 1 — Brand Ingestion
**System analyzes the submitted URL and extracts:**
- Brand name, tagline, positioning statement
- Market category / vertical
- Target audience signals
- Color palette, typography, visual style
- Tone of voice
- Key product/service offerings
- Geographic focus

**Output:** Structured brand profile JSON

---

## Stage 2 — Competitive Discovery (Shallow Report)
**System identifies up to 20 competitive brands:**
- Direct competitors (same category, same audience)
- Adjacent competitors (category overlap)
- Aspiration brands (where the brand wants to be)

**Per competitor (shallow):**
- Name, URL, one-line description
- Market position
- Estimated threat level (low / medium / high)
- Logo + primary color
- Tagline

**User actions on shallow report:**
- Edit any field (name, description, position)
- Remove a brand from the list
- Add a brand manually
- Select 10 brands for deep research

---

## Stage 3 — Deep Research (10 Selected Brands)
**Per selected brand (deep):**
- Visual identity breakdown (color, type, photography style)
- Positioning + messaging analysis
- Website UI/UX patterns
- Content strategy signals
- Social media presence
- Strengths + weaknesses
- Threat score (0–100)

---

## Stage 4 — Scoring + Gallery
**Submitted brand scored against competition:**
- Positioning differentiation score
- Visual distinctiveness score
- Digital presence score
- Overall competitive score (0–100)

**Gallery output:**
- Grid of brand cards (submitted brand + 10 deep-researched)
- Each card: logo, score, 3 key signals
- Click → full brand deep-dive document

---

## Stage 5 — Branding Proposal
**AI-generated brief for the submitted brand:**
- Positioning recommendation
- Differentiation opportunity
- Visual direction (mood, color strategy, type direction)
- Tone of voice direction
- Key messages to own
- What to avoid (what competitors own)

---

## Data Models (to build in data/schemas/)
- `brand-profile.json` — Submitted brand
- `competitor.json` — Each competitor (shallow + deep)
- `competition-map.json` — Full 20-brand map
- `report.json` — The assembled report
- `branding-proposal.json` — The final brief

---

## Tech Stack (Proposed)
- **Research engine:** Gemini (Google Search grounding) + Claude (analysis)
- **Scraping:** Playwright
- **Frontend:** HTML/CSS/JS (CE styleguide) — upgrade to React/Next if needed
- **Storage:** JSON files per report (MVP), DB later
- **Deployment:** CE staging → production