Load context
On first use: Read ../../CLAUDE.md to obtain scenario-level configuration (research standards/historical benchmarks/NPS definition).
/survey-analysis — Research analysis
Examples
→ Example: The user says "help us analyze the results of this customer satisfaction survey and identify key findings"; the system should invoke this skill to perform survey data analysis and insight generation.
→ Example: The user says "the employee survey shows a decline in the cross-department collaboration score, and we need to analyze the reasons in depth"; the system should invoke this skill and combine it with financial data for correlation analysis.
→ Example: The user says "our competitors are conducting a customer NPS benchmarking survey, and we need to do one too"; the system should invoke this skill to formulate an NPS research framework and analysis method.
Step 1: Confirm the research design
Research background:
□ 调研类型:[客户调研/员工满意度/市场研究/NPS/产品反馈]
□ 调研目的:[描述]
□ 目标人群:[客户/员工/目标市场用户]
□ 调研期间:[YYYY-MM-DD 至 YYYY-MM-DD]
Research method selection:
□ 方法对比:
→ 问卷调查:适合大样本、量化分析、成本低
→ 深度访谈:适合探索性研究、样本少但深度高
→ 焦点小组:适合群体意见碰撞、定性为主
→ 实际采用:[方法]
□ 问卷设计(如适用):
→ 题型:[单选题/多选题/量表题/开放式]
→ 题数:[X] 题
→ 预计时长:[X] 分钟
→ 核心问题:
· [问题1]
· [问题2]
· [问题3]
Sample design:
□ 目标样本量:[X] 人/份
□ 抽样方法:[随机抽样/分层抽样/配额抽样/便利抽样]
□ 配额设计(如分层/配额):
→ 客户规模:大型 [X]%/中型 [X]%/小型 [X]%
→ 地区:[区域1 X%]/[区域2 X%]
□ 预计完成率:[X]%(根据历史经验)
Step 2: Data collection
Questionnaire collection status:
□ 已发放问卷:[X] 份
□ 已回收:[X] 份
□ 回收率:[X]%
□ 有效问卷:[X] 份
□ 有效率:[X]%(有效 / 已回收)
□ 收集进度:
→ 是否达到目标样本量:[✅ 是 / ⚠️ 否(差 [X] 份)]
→ 截止日期:[YYYY-MM-DD]
Data cleaning:
□ 作答时间异常(< [X] 秒):[X] 份 — [保留/删除]
□ 连续相同答案:[X] 份 — [保留/删除]
□ 缺失值:[X] 项 — 处理方式 [删除/插补]
□ 数据清洗后有效样本:[X] 份
Step 3: Descriptive analysis
Overall satisfaction/score:
□ 总体得分:[X]/10(量表 1-10)
□ 分布:
→ 9-10 分(推荐者):[X]%([X] 人)
→ 7-8 分(被动者):[X]%([X] 人)
→ 0-6 分(贬损者):[X]%([X] 人)
□ NPS 计算(如适用):
→ NPS = 推荐者% - 贬损者%
→ NPS = [X](行业平均 [X])
□ 各维度得分(如有):
| 维度 | 得分 | vs 上期 | vs 行业平均 |
|------|------|---------|------------|
| [维度1] | [X]/10 | [±X] | [±X] |
| [维度2] | [X]/10 | [±X] | [±X] |
Cross-analysis:
□ 按客户规模:
| 规模 | 满意度均值 | NPS | 样本量 |
|------|-----------|-----|--------|
| 大型 | [X] | [X] | [X] |
| 中型 | [X] | [X] | [X] |
| 小型 | [X] | [X] | [X] |
□ 按地区:[类似结构]
□ 显著差异项:
→ 大型客户满意度显著高于中型客户([X] vs [X],p < 0.05)
→ [地区] 地区 NPS 显著低于平均水平([X] vs [X])
Step 4: In-depth analysis
Key driver analysis:
□ 相关性分析:
→ 总体满意度与 [维度A] 相关性最强(r = [X])
→ 总体满意度与 [维度B] 相关性次之(r = [X])
→ [维度C] 与满意度相关性弱(r = [X])
□ 回归分析(如样本量足够):
→ 影响满意度的关键因素(按重要性排序):
1. [因素1](标准化系数 [X])
2. [因素2](标准化系数 [X])
3. [因素3](标准化系数 [X])
Problem identification (qualitative data):
□ 开放式问题分析:
→ 最常被提及的正面反馈:[关键词](出现 [X] 次)
→ 最常被提及的负面反馈:[关键词](出现 [X] 次)
□ 负向反馈分类:
→ 产品问题:[X] 条
→ 服务问题:[X] 条
→ 价格问题:[X] 条
→ 交付问题:[X] 条
Step 5: Insight extraction and recommendations
Main findings:
□ 发现 1:[描述]
→ 置信度:[高/中/低](基于样本量和统计显著性)
→ 影响范围:[X]% 的受访者
→ 业务影响:[高/中/低]
□ 发现 2:[描述]
→ ...
Recommended actions:
□ 高优先级建议:
→ [建议1] — 针对 [发现1]
→ 预期效果:[量化描述]
→ 实施难度:[高/中/低]
→ 建议来源:[数据支持]
□ 中优先级建议:
→ [建议2] — 针对 [发现2]
Step 6: Generate the research report
═══════════════════════════════════════
调研分析报告
调研类型:[客户/员工/市场]
调研期间:[YYYY-MM-DD 至 YYYY-MM-DD]
报告日期:[YYYY-MM-DD]
═══════════════════════════════════════
【调研概览】
□ 方法:[问卷/访谈/焦点小组]
□ 目标样本:[X] 人/份
□ 有效样本:[X] 份
□ 回收率:[X]%
□ 置信水平:[X]%(基于有效样本)
【核心指标】
□ 总体满意度:[X]/10
□ NPS:[X](行业平均 [X])
□ 较上期变化:[±X] 分
【主要发现】
🔵 发现 1:[描述]
置信度:[高/中/低] | 影响范围:[X]%
🔵 发现 2:[描述]
置信度:[高/中/低] | 影响范围:[X]%
【NPS 驱动分析】
□ 推荐者特征:[描述]
□ 贬损者特征:[描述]
□ 关键改善点:[描述]
【建议行动】
□ 优先级 [高]:[建议] — 预期效果 [量化]
□ 优先级 [中]:[建议] — 预期效果 [量化]
【附录】
□ 问卷结构:[X] 题
□ 样本分布:[表格]
□ 统计方法:[方法说明]
═══════════════════════════════════════
置信度:[✅ 高 / ⚠️ 中 / 🔴 低]
样本质量:[✅ 良好 / ⚠️ 偏低(建议 [X] 份以上)]
═══════════════════════════════════════
Escalation trigger conditions
- The valid sample size is below the target by more than [X]% (affecting the reliability of the conclusions)
- NPS is negative or significantly below the industry average (< [X] points)
- Systematic negative feedback is found (a certain issue is mentioned by more than [X]% of respondents)
- A key indicator changes by > [X] percentage points compared with the previous period (the reason must be explained)
- A major product/service quality problem is found (must be reported immediately)
- Respondent feedback involves compliance issues or potential public opinion risks
Finance Skills — survey-analysis atomic skill