Product Selection

Product Selection

Comprehensive product selection workflow with evidence-first methodology. Guides you through industry analysis, consumer research, product discovery, and supplier matching based on user's starting point.

When to Use

• User wants product selection, trending products, category analysis, or follow-sell opportunities
• Questions like: "what to sell", "which products are trending", "how to choose a category", "find hot products for [market]"
• User asks about selection methodology or data sources
• Coordinates with product-supplier-sourcing skill for concrete product/supplier retrieval

Execution Flow (start point depends on context)

Key Rules:

• Step2 (consumer research) is critical for category selection within an industry — skip only if user specifies both industry AND specific category
• Step4 (supplier) runs ONLY when user explicitly mentions "supplier", "factory", "manufacturer", or "sourcing". Terms like "find products" or "hot products" do NOT trigger Step4.

Core Principles

1. User intent overrides skill: Follow user's specific requirements (platform, market, timeframe, format) even if they differ from this guide
2. Evidence first: Every conclusion needs data support; cite sources and methods
3. Multi-source validation: Use ≥2 data source types for trend/industry conclusions (e.g., platform data + third-party tools + social/reviews)
4. Explicit data gaps: State what data is missing and suggest next collection steps instead of guessing

Step1: Industry Research

Trigger: User has NOT specified industry/category.

Goal: Identify promising industries/categories for product selection.

Selection Criteria (priority order)

Data Sources

Output Format:

Industry: [Category Name]
Conclusion: [Promising/Not Recommended]
Evidence:
  - Traffic: [Trend data + source]
  - Sales: [Trend data + source]
  - Concentration: [Top 10 share + source]
  - New listings: [Rate + source]
Data Gaps: [List any missing data points]

Step2: Consumer Research

Trigger:

• Required when industry is not pre-specified (pairs with Step1)
• Also required when industry is specified but specific product category is not — consumer insights guide category selection within the industry

Goal: Identify consumer pain points, unmet needs, and product opportunities to guide category/product selection.

Research Method

1. Extract complaints/needs from:

   • E-commerce reviews (Amazon, eBay, Walmart, etc.)
   • Social discussions (Reddit, Facebook groups, TikTok comments)
   • Q&A platforms (Quora, specialized forums)
   • App reviews (for digital/smart products)

2. Use specialized skills for review analysis:

   • review-summarizer skill: For scraping and summarizing reviews from Amazon, Google, Yelp, TripAdvisor
     • Generates sentiment analysis, pros/cons, complaint frequency
     • Supports multi-platform comparison
   • review-analyst-agent skill: For deep analysis and prioritization
     • Identifies top complaints with frequency/severity
     • Extracts feature requests
     • Provides actionable improvement recommendations
     • Use when you need structured analysis with priority matrix

3. Categorize pain points:

   • Quality issues (durability, materials, craftsmanship)
   • Design/functionality gaps (missing features, poor UX)
   • Price sensitivity (overpriced, poor value perception)
   • Delivery/service problems (shipping, returns, support)
   • Safety/compliance concerns

4. Map to product opportunities:

   • Improvement opportunities: Better quality/features than existing products
   • Gap-filling opportunities: Unmet needs in the market
   • Differentiation opportunities: Unique value proposition
   • Follow-sell opportunities: Products with complaints that can be addressed

Integration with Review Analysis Skills

When to use each skill:

Example workflow:

1. User specifies industry: "Home & Kitchen"
2. Use review-analyst-agent to analyze top products in that industry
3. Extract top complaints (e.g., "poor battery life", "leaks easily")
4. Map complaints to product categories (e.g., "rechargeable kitchen scales", "leak-proof containers")
5. Proceed to Step3 with category focus

Output Format:

### Consumer Research Summary

**Sources Analyzed**: 
- [Platform1]: [Number] reviews
- [Platform2]: [Number] posts/comments

**Top Pain Points** (by frequency):

1. [Pain Point1] — Frequency: [X]% of complaints
   - Sources: [Platform + sample reviews/posts]
   - Severity: [High/Medium/Low]
   - Product Category Opportunity: [Specific category to target]
   - Product Direction: [How to address this]
  
2. [Pain Point2] — Frequency: [X]% of complaints
   - Sources: [Platform + sample reviews/posts]
   - Severity: [High/Medium/Low]
   - Product Category Opportunity: [Specific category to target]
   - Product Direction: [How to address this]

**Feature Requests**: 
- [Request1] — Frequency: [X] mentions
- [Request2] — Frequency: [X] mentions

**Recommended Categories for Step3**:
1. [Category1] — Addresses pain points: [#1, #3]
2. [Category2] — Addresses pain points: [#2, #4]

Step3: Product Selection

Trigger: Industry specified OR after completing Steps 1-2.

Goal: Identify specific products to follow-sell, develop, or list.

Approach1: E-commerce Platform Analysis

Target: Find products with proven demand but manageable competition.

Selection Logic:

• Sweet spot: BSR rank 100-5,000 (varies by category)
• Avoid: Top 20 (too competitive), BSR > 10,000 (uncertain demand)
• Check: Review count (300-1,000 = validated), rating (4.0-4.5 = room for improvement)

Tools & Platforms:

Concrete Steps:

1. Filter category by Step1/2 criteria (industry + consumer pain points)
2. Export products in target BSR range (use product_supplier_search if available, or manual export)
3. Competitive analysis (use competitive-landscape skill for deeper analysis):
   • Identify top 5-10 competitors in the BSR range
   • Analyze differentiation opportunities (product features, pricing, positioning)
   • Use review-analyst-agent to find their weaknesses (top complaints)
   • Map improvement opportunities (e.g., "Competitor has battery complaint → source better battery")
4. Calculate: Estimated sales × margin - fees = profit
5. Rank by: (Profit potential × Demand validation × Differentiation feasibility) / Competition level

Integration with Competitive Analysis:

• Use competitive-landscape skill when you need to:
  • Understand competitive intensity (Porter's Five Forces)
  • Identify positioning gaps (positioning map)
  • Develop differentiation strategy
• Use review-analyst-agent to find competitor weaknesses:
  • Extract top complaints from competitor products
  • Identify improvement opportunities
  • Prioritize which issues to solve in your product selection

Approach2: Advertising & Traffic Analysis

Target: Products with strong recent marketing momentum.

Selection Logic:

• Ad frequency: Seen in ads ≥10 times in past 30 days (use Pipiads/Minea)
• Engagement: High CTR/comments/shares relative to ad spend
• Freshness: Launched within past 6 months (less saturated)

Tools:

Concrete Steps:

1. Search category keywords in ad intelligence tools
2. Filter: 10+ appearances in past 30 days, engagement rate > 5%
3. Cross-check with product_supplier_search (intent_type=product) to find supply options
4. Validate on destination marketplace (check if already saturated)

Approach3: Crowdfunding Signal Mining

Target: Innovative products for early follow-sell or white-label opportunities.

Selection Logic:

• Funding level: 200-500% of goal (validated interest, not yet mass market)
• Backer count: 500-5,000 (niche demand, scalable)
• Timeline: Projects ending in 1-3 months (time to source & launch)

Platforms:

• Kickstarter / Indiegogo (innovation signal)

⚠️ Caution:

• High risk: Crowdfunding success ≠ marketplace success
• IP concerns: Check for patents before copying
• Long development: Expect 6-12 month sourcing/development cycle

When to use: Only for users with product development capability and higher risk tolerance.

Step4: Supplier Matching

⚠️ CRITICAL: Execute ONLY when user explicitly requests suppliers/factories/manufacturers/sourcing.

Goal: Match selected products to qualified suppliers.

Matching Process

1. Prepare product specs:

   • Product name/category
   • Key features (from Step3 selection)
   • Target price range
   • MOQ constraints

2. Search suppliers:

   • Primary: Call product_supplier_search (intent_type=supplier, query in English)
     • This searches alibaba.com by default (per product-supplier-sourcing skill)
     • Output must include:
       • Clickable supplier links: [Supplier Name](https://alibaba.com/company/...)
       • Product thumbnails if available: <img src="..." width="80">
       • Direct store links for each supplier
   • Secondary: If alibaba.com unsuitable, try:
     • 1688.com (domestic China supply, lower MOQ)
     • Made-in-China, Global Sources (alternative B2B platforms)
     • Direct factory outreach (trade shows, referrals)

3. Qualification criteria:

   • Trade assurance / verified status
   • MOQ matches your budget
   • Production capacity (order size flexibility)
   • Sample policy (can order samples before bulk)
   • Communication responsiveness

Output Format:

Product: [Name from Step3]
Target Specs: [Key features/requirements]

Supplier1: **[Company name](https://alibaba.com/company/...)** 
  Platform: alibaba.com
  MOQ: [Quantity + price]
  Lead time: [Days]
  Qualification: [Verified/Trade Assurance/Years]
  Store Link: [Visit Store](https://alibaba.com/company/...)

Supplier2: ...

Next Steps:
  - Request samples from top 2-3 suppliers
  - Compare quality/pricing
  - Negotiate terms (MOQ, payment, customization)

Visual Presentation (if products shown):

• Include product thumbnails: <img src="..." width="80">
• Verify image URLs are valid (no broken links)
• Make all product/supplier names clickable
• See product-supplier-sourcing skill for detailed formatting requirements

Selecting the Right Approach

Default: If user doesn't specify, recommend Approach1 (lowest risk, most accessible data).

Tool Integration

Using product_supplier_search (from product-supplier-sourcing skill)

When in Step3 (Product Selection):

intent_type: "product"
query: "[product category/name] [key attributes]" (in English)

Example: query: "wireless earbuds, waterproof, noise cancelling"

When in Step4 (Supplier Matching):

intent_type: "supplier"
query: "[product category] [optional: region/industry]" (in English)

Example: query: "consumer electronics, Shenzhen"

Important: This tool searches alibaba.com only. If user requests other platforms (1688, Amazon, etc.), see product-supplier-sourcing skill for fallback strategies.

Output Requirements (see product-supplier-sourcing skill for full details)

When presenting product or supplier search results:

1. Include product thumbnails:

   • Display images inline with controlled size: <img src="..." width="80">
   • Check image URL validity before displaying
   • If image unavailable, use placeholder: 📦 or [No Image]

2. Include clickable links:

   • Product name: [Product Name](https://alibaba.com/product/...)
   • Supplier: [Supplier Name](https://alibaba.com/company/...)
   • Detail link: [View Details](https://alibaba.com/product/...)

3. Example product table format:

   | Thumbnail | Product | Price | MOQ | Supplier | Link |
   |-----------|---------|-------|-----|----------|------|
   | <img src="..." width="80"> | **[Product A](link)** | $5-$8 | 100 pcs | [Supplier X](link) | [Details](link) |
   

4. Image validation checklist:

   • ✅ Verify image URL starts with https:// or http://
   • ✅ Use width="80" for consistent sizing
   • ✅ If image URL invalid, omit image but keep product link
   • ❌ Never display broken image links
   • ❌ Never exceed width="150" in tables

Output Structure

Comprehensive Selection Report

## 1. Selection Scope
- Market: [Target marketplace/region]
- Category: [If specified, or "TBD after industry research"]
- Goal: [Follow-sell / Brand building / Innovation]
- Supplier required: [Yes/No]

## 2. Industry Analysis (if Step1 executed)
### Selected Industry: [Name]
- Traffic trend: [Data + source]
- Sales trend: [Data + source]
- Competition: [Concentration data]
- New listings: [Rate + source]
- **Conclusion**: [Why this industry is promising/risky]
- **Data gaps**: [What's missing]

## 3. Consumer Insights (if Step2 executed)
### Analysis Sources
- [Platform1]: [Number] reviews/posts analyzed
- [Platform2]: [Number] reviews/posts analyzed
- **Skills used**: [review-summarizer / review-analyst-agent]

### Top Pain Points (by frequency)
1. [Pain point] — [X]% of complaints
   - **Severity**: [High/Medium/Low]
   - **Product category opportunity**: [Specific category]
   - **Product direction**: [How to address]
   - **Sample evidence**: "[Quote from review]"

2. [Pain point] — [X]% of complaints
   - [Same structure]

### Feature Requests
- [Request1] — [X] mentions
- [Request2] — [X] mentions

### Recommended Categories for Selection
1. [Category1] — Addresses pain points: [#1, #3]
2. [Category2] — Addresses pain points: [#2, #4]

## 4. Product Candidates (Step3 - REQUIRED)
### Approach Used: [E-commerce / Advertising / Crowdfunding]

### Competitive Landscape
**Skills used**: [competitive-landscape / review-analyst-agent]

- **Porter's Five Forces summary** (if deep analysis):
  - New entrants: [Low/Medium/High]
  - Competition intensity: [Low/Medium/High]
  - Key finding: [Insight]

- **Competitor weakness analysis** (from reviews):
  - Competitor A: Top complaint = "[Issue]" ([X]% of reviews)
  - Competitor B: Top complaint = "[Issue]" ([X]% of reviews)
  - **Opportunity**: Address these issues in product selection

### Product Selection Results

| Thumbnail | Product | BSR/Rank | Est. Sales | Price | Competition | Improvement Opportunity | Link |
|-----------|---------|----------|------------|-------|-------------|------------------------|------|
| <img src="..." width="80"> | **[Product Name](link)** | [Data] | [Data] | [$] | [Low/Med/High] | [What to improve vs. competitors] | [View Details](link) |

**Note**: 
- Include product thumbnails with `width="80"` for visual scanning
- Verify image URLs are valid before displaying
- All product names must be clickable links to source pages
- Add "View Details" link for quick access

**Recommendation**: [Top 3 products with reasoning]
- Product1: [Why] — Addresses pain point: [#1], competitor weakness: [Issue]
- Product2: [Why] — Addresses pain point: [#2]
- Product3: [Why] — Differentiation opportunity: [Gap in positioning]

**Risk notes**: [Competition, seasonality, compliance issues]

## 5. Supplier Options (only if Step4 executed)
[See Step4 output format above]

## 6. Next Steps
- [ ] [Specific action1]
- [ ] [Specific action2]
- [ ] Deep dive on competitor reviews (use `review-analyst-agent`)
- [ ] Validate positioning gap (use `competitive-landscape`)

Citation Requirements

Every conclusion must include:

• Data source: Tool/platform name
• Timeframe: When data was collected
• Method: How the conclusion was derived

Example: ❌ "This product is trending" → ✅ "This product has 150% sales growth in past 30 days (Source: Jungle Scout, accessed 2024-03-07)"

Validation Checklist

Before finalizing output:

• Starting point determined correctly:
  • No industry specified → Step1 + 2
  • Industry specified, no category → Step2 + 3
  • Industry + category specified → Step3
• Step2 (consumer research) executed when needed (unless user provides both industry AND category)
• Step4 executed ONLY if user explicitly requested suppliers
• Review analysis skills used in Step2 (review-summarizer or review-analyst-agent)
• Competitive analysis considered in Step3 (competitor weaknesses, positioning gaps)
• All conclusions have data sources cited
• Multi-source validation used for trend claims (≥2 sources)
• Data gaps explicitly stated (not hidden)
• Quantitative criteria used (not vague terms like "popular")
• Risk factors mentioned (competition, seasonality, compliance)
• Actionable next steps provided
• Product thumbnails included in output tables (with width="80")
• All image URLs validated (no broken links)
• All product/supplier names are clickable links to source pages
• Image size controlled (not oversized in tables)

Common Pitfalls to Avoid

❌ Don't:

• Skip Step2 (consumer research) when user only specifies industry — category selection needs consumer insights
• Run supplier search (Step4) when user only asked for "products"
• State "this is a hot product" without sales/trend data
• Recommend products without analyzing competitor weaknesses
• Ignore data gaps and sound overconfident
• Mix up product-supplier-sourcing (alibaba.com) with 1688/other platforms
• Select products without checking review pain points
• Present products without images or links — always include thumbnails and clickable links
• Use broken/invalid image URLs — check validity before displaying
• Display oversized images — use width="80" in tables

✅ Do:

• Start from the right step based on user context (see Execution Flow)
• Use review-analyst-agent to understand consumer pain points in Step2
• Use competitive-landscape to identify positioning gaps in Step3
• Analyze competitor reviews to find improvement opportunities
• Cite specific data sources and timeframes
• Quantify criteria (BSR ranges, sales thresholds, complaint frequencies)
• State what you can't verify and suggest how to get that data
• Align tool capabilities with user expectations
• Include product thumbnails (<img src="..." width="80">) for visual scanning
• Make all product/supplier names clickable links to source pages
• Verify image URLs before displaying (avoid broken links)
• Control image size in tables (width="80" recommended)

Dependencies

Core Skills

• product-supplier-sourcing skill: Provides product_supplier_search tool (alibaba.com) for product/supplier retrieval

Consumer Research Skills (Step2)

• review-summarizer skill: Multi-platform review scraping and summarization (Amazon, Google, Yelp, TripAdvisor)
• review-analyst-agent skill: Deep review analysis with priority matrix and actionable recommendations

Competitive Analysis Skills (Step3)

• competitive-landscape skill: Porter's Five Forces, positioning maps, differentiation strategies

Optional Tools

• chart-design-guide.md: Reference for visualizing trends (if generating charts from collected data)

Advanced: Data Visualization

If you collect sufficient time-series or comparison data and need to visualize trends:

See chart-design-guide.md in this skill directory for:

• Chart type selection (trend lines, bar charts, radar charts)
• Seaborn/Matplotlib implementation
• When to use tables vs. charts

When to visualize:

• Comparing 5+ categories (use bar chart)
• Showing 3-month+ trends (use line chart)
• Multi-factor product comparison (use radar chart)

When NOT to visualize:

• Only 2-3 data points (use table)
• Data from screenshots/third-party tools (just cite the source)