亚马逊选品运营智能分析师
核心理念
从"问答式" → "顾问式"
从"被动响应" → "主动发现"
从"固定报告" → "按需交付"
用户交互流程（极简）
Step 1: 用户输入（一句话）
用户: "我想做沙滩裤"
用户: "分析women active shorts市场"
用户: "帮我找选品机会"
Step 2: 系统快速扫描（30秒）
系统自动：
- 调用
get_complete_market_intelligence()获取市场数据 - 计算机会评分
输出示例：
[Quick Scan Result]
[HIGH] Score: 72/100 - Worth entering!
[Market Size] 25/30
Monthly: 52,847 units
Level: large
[Competition] 20/30
CR3: 56.14%
Level: medium
[Trend] 15/20
Direction: up
Growth: +8.5%/month
[Price Opportunity] 12/20
Gap: Yes
Step 3: 智能判断与交互
场景A: 高机会（评分≥60）
[HIGH OPPORTUNITY]
Worth entering! This market shows strong potential.
To provide deep analysis, I need:
1. Your target price? (market avg $28-35)
2. Your estimated cost?
[Start Deep Analysis] [View Quick Data]
场景B: 中机会（评分40-59）
[MEDIUM OPPORTUNITY]
Opportunity exists but with risks:
1. High competition, need strong differentiation
2. Market trend not strong
Key Questions:
1. Do you have unique design or supply chain advantage?
2. Can you control cost under $10?
3. Do you have enough budget for promotion?
[Proceed with Caution] [View Alternatives]
场景C: 低机会（评分<40）
[LOW OPPORTUNITY]
Not recommended:
1. Small market or high competition
2. Hard for new brands
3. Limited profit
[Suggest Alternatives]
Step 4: 深度分析（2分钟，可选）
用户提供价格/成本后，启动完整7步分析：
[Deep Analysis] Starting...
[Step 1] Market Overview... OK
[Step 2] Brand Structure... OK
[Step 3] Price Band Analysis... OK
[Step 4] Profit Calculation... OK
[Step 5] Product Attributes... OK
[Step 6] Market Trend... OK
[Step 7] Seller Distribution... OK
[Analysis Complete]
[Decision] GO
[Score] 65/100
Key Findings:
- Market Size: 52,847 units/month, $1.2M revenue
- Competition: CR3=56%, medium concentration
- Your Position: $32 price in optimal $28-35 band
- Profit: $4.20/unit, 13.1% margin
- Break-even: 180 units/month
Action Items:
1. Contact suppliers for cost confirmation
2. Finalize differentiation design
3. Sample production and testing
4. First batch inventory
输出分层
新手模式（Beginner）
[Conclusion]
This market is WORTH ENTERING (Score: 72/100)
[Key Numbers]
- Monthly Sales: 52,847 units
- Your Profit: $4.20/unit
- Break-even: 180 units/month
[Action Checklist]
☐ Contact 3 suppliers for quotes
☐ Confirm cost under $12
☐ Design differentiation features
☐ Order 300 units first batch
[Risk Alert]
⚠️ Competition is medium (CR3=56%)
⚠️ Need strong differentiation
[Next Step]
[Start Supplier Research] [View Competitors] [Generate Listing]
标准模式（Standard）
完整7步分析 + 数据表格 + 可视化图表
专家模式（Expert）
原始数据下载 + 自定义分析参数
技术实现
核心类
class AmazonSellerV2:
def analyze(self, user_input: str,
price: Optional[float] = None,
cost: Optional[float] = None) -> Dict:
"""
Main entry point
Flow:
1. Extract keyword from user_input
2. Quick scan (30s) → opportunity score
3. Branch based on score:
- ≥60: Ask for price/cost → deep analysis
- 40-59: Cautious analysis
- <40: Reject with alternatives
"""
pass
机会评分算法
def calculate_opportunity_score(market_data: Dict) -> int:
"""
Calculate opportunity score (0-100)
Components:
- Market Size: 30 points
- Competition: 30 points
- Trend: 20 points
- Price Opportunity: 20 points
"""
score = 0
# Market Size (30pts)
units = market_data.get("monthly_units", 0)
if units > 50000: score += 30
elif units > 20000: score += 25
elif units > 10000: score += 20
elif units > 5000: score += 15
else: score += 10
# Competition (30pts)
cr3 = market_data.get("cr3", 100)
if cr3 < 40: score += 30
elif cr3 < 50: score += 25
elif cr3 < 60: score += 20
elif cr3 < 70: score += 15
else: score += 10
# Trend (20pts)
trend = market_data.get("trend_direction", "unknown")
if trend == "up": score += 20
elif trend == "stable": score += 15
else: score += 10
# Price Opportunity (20pts)
bands = market_data.get("price_bands", [])
if len(bands) >= 3: score += 20
elif len(bands) >= 2: score += 15
else: score += 10
return score
用户体验优化
1. 渐进式信息收集
不是：一次性询问所有参数 而是：先快速判断，有机会再问关键信息
2. 实时反馈
Analyzing...
[====> ] 25%
[=========> ] 50%
[===============> ] 75%
[==================] 100% ✓
3. 交互式调整
User: "What if cost drops to $10?"
System: [Recalculate instantly]
Profit: +40% → $5.89/unit
Margin: 18.4%
依赖
unified_data_layer_v2.py- 统一数据层（26 APIs）sellersprite_mcp.py- MCP客户端- SellerSprite API access
版本
V2 - 2026-04-12
- Complete refactor from Q&A to advisor mode
- Eliminated Excel upload requirement
- Added intelligent opportunity scoring
- Implemented layered output