health-score-monitor

健康度评分监控

通过综合评分、趋势跟踪和自动告警，构建系统化的客户健康度监控，主动提升客户成功。

何时使用此技能

• 设计健康度评分框架
• 设置监控仪表板
• 创建告警阈值
• 分析整个客户组合的健康度趋势
• 优化现有健康度模型

方法论基础

基于 Gainsight Health Score Design 和 Totango Customer Success metrics，重点关注：

• 多维度评分
• 领先指标与滞后指标
• 评分标准化
• 趋势分析
• 告警优先级排序

Claude 负责什么，您决定什么

Claude 负责	您决定
设计评分框架	维度权重
计算综合评分	告警阈值
识别趋势模式	干预触发条件
建议监控频率	资源分配
推荐改进措施	业务规则例外情况

此技能的功能

1. 框架设计 - 多因素健康度模型
2. 评分计算 - 加权综合评分
3. 趋势分析 - 方向和速度
4. 告警规则 - 何时通知团队
5. 客户组合视图 - 汇总健康度概览

使用方法

Design a health score monitor for my customer portfolio:

Business Context:
- Product type: [SaaS/Platform/Service]
- Contract model: [Annual/Monthly/Multi-year]
- Key value metric: [What shows customer success?]
- CSM:Account ratio: [1:X]

Available Data Points:
- Product: [List usage metrics available]
- Support: [List support metrics available]
- Financial: [List financial signals]
- Relationship: [List engagement data]

Current Challenges:
- [What's not working with current approach?]

说明

第 1 步：定义健康度维度

标准四支柱模型：

维度	权重	回答的问题
产品	30-40%	他们在使用吗？
支持	15-25%	他们满意吗？
财务	20-25%	他们在付费或增长吗？
关系	20-25%	我们与他们保持联系吗？

根据您的业务调整权重：

• 高接触服务：提高关系维度权重
• 按使用量定价：提高产品维度权重
• 支持密集型业务：提高支持维度权重

第 2 步：为各维度选择指标

产品健康度指标：

指标	类型	评分方式
DAU/MAU	领先	基准值百分比
功能采用率	领先	已使用功能的百分比
产品使用时长	领先	分钟数与平均值对比
关键功能使用情况	领先	是/否或频率
使用趋势	领先	上升/持平/下降

支持健康度指标：

指标	类型	评分方式
CSAT 评分	滞后	1-5 分制
工单量	领先	与基准值对比
升级处理	领先	数量（负向）
回复情绪	领先	正面/中性/负面
解决时长	滞后	与 SLA 对比

财务健康度指标：

指标	类型	评分方式
付款状态	滞后	正常/逾期
扩展	领先	销售管道/讨论中
合同类型	滞后	多年期加分
续约日期	背景信息	剩余天数
ARR 趋势	滞后	增长/持平/下降

关系健康度指标：

指标	类型	评分方式
核心支持者参与度	领先	积极/被动/无
高管赞助人	领先	是/否
NPS/CSAT	滞后	评分
QBR 参与情况	领先	定期/偶尔/无
上次联系	领先	距联系的天数

第 3 步：创建评分规则

评分矩阵示例：

Product Score (0-100):
+30: DAU/MAU > 70%
+20: DAU/MAU 50-70%
+10: DAU/MAU 30-50%
+0:  DAU/MAU < 30%

+25: Feature adoption > 60%
+15: Feature adoption 40-60%
+5:  Feature adoption < 40%

+20: Usage trending up
+10: Usage stable
-10: Usage trending down

+15: Key feature used weekly
+5:  Key feature used monthly
-10: Key feature abandoned

Max: 100 (or normalize)

第 4 步：设置告警阈值

条件	告警级别	操作
评分下降超过 15 分	🔴 严重	CSM 立即审查
评分低于 40	🟠 高	需要干预
评分为 40-60	🟡 中	主动联系客户
距续约不到 60 天且评分低于 60	🔴 严重	制定挽留计划
WoW 使用量下降超过 30%	🟡 中	安排回访电话

第 5 步：定义监控频率

活动	频率	负责人
刷新评分	每日/每周	自动执行
审查告警	每日	CSM
客户组合审查	每周	CS 经理
校准模型	每季度	CS Ops

示例

示例 1：SaaS 健康度评分设计

输入：

Design health score for our B2B SaaS platform:

Business Context:
- Product: Project management software
- Contract: Annual, avg $50K
- Value metric: Projects completed, team adoption
- CSM ratio: 1:40

Available Data:
- Product: Logins, projects created, tasks completed, integrations used
- Support: Zendesk tickets, CSAT, NPS
- Financial: ARR, payment status, expansion pipeline
- Relationship: Meeting attendance, email engagement

Challenges:
- Too many "yellow" accounts, can't prioritize
- Alerts are noisy, CSMs ignoring them

输出：

## Health Score Framework: Project Management SaaS

### Design Principles

Given 1:40 ratio and "alert fatigue" issue:
1. **Prioritize leading indicators** - Catch issues early
2. **Reduce noise** - Tighter alert thresholds
3. **Enable triage** - Clear Red/Yellow/Green actions

---

### Scoring Model

#### Dimension Weights
| Dimension | Weight | Rationale |
|-----------|--------|-----------|
| Product | 40% | Usage = value realization |
| Support | 15% | Lower touch model |
| Financial | 25% | Contract health critical |
| Relationship | 20% | Moderate touch needed |

---

#### Product Score (40 points max)

| Metric | Criteria | Points |
|--------|----------|--------|
| **Weekly Active Users** | | |
| | >70% of seats | 12 |
| | 50-70% | 8 |
| | 30-50% | 4 |
| | <30% | 0 |
| **Projects Created (monthly)** | | |
| | >5 per user | 10 |
| | 2-5 per user | 6 |
| | <2 per user | 2 |
| **Tasks Completed** | | |
| | >80% completion rate | 8 |
| | 50-80% | 5 |
| | <50% | 0 |
| **Integrations** | | |
| | 2+ active | 6 |
| | 1 active | 3 |
| | None | 0 |
| **Usage Trend** | | |
| | Increasing | 4 |
| | Stable | 2 |
| | Declining | -4 |

---

#### Support Score (15 points max)

| Metric | Criteria | Points |
|--------|----------|--------|
| **CSAT** | | |
| | >4.5 | 6 |
| | 4.0-4.5 | 4 |
| | 3.5-4.0 | 2 |
| | <3.5 | 0 |
| **Ticket Volume** | | |
| | Below baseline | 5 |
| | At baseline | 3 |
| | 2x+ baseline | 0 |
| **Escalations** | | |
| | None (90 days) | 4 |
| | Resolved | 2 |
| | Open | -4 |

---

#### Financial Score (25 points max)

| Metric | Criteria | Points |
|--------|----------|--------|
| **Payment Status** | | |
| | Current | 10 |
| | 30 days late | 5 |
| | 60+ days | 0 |
| **Expansion** | | |
| | Closed this year | 8 |
| | In discussion | 5 |
| | No opportunity | 2 |
| **Contract Type** | | |
| | Multi-year | 7 |
| | Annual | 4 |
| | Month-to-month | 0 |

---

#### Relationship Score (20 points max)

| Metric | Criteria | Points |
|--------|----------|--------|
| **Champion Status** | | |
| | Active advocate | 8 |
| | Engaged | 5 |
| | Passive | 2 |
| | Gone/None | 0 |
| **NPS** | | |
| | 9-10 (Promoter) | 6 |
| | 7-8 (Passive) | 4 |
| | 0-6 (Detractor) | 0 |
| **Last Touch** | | |
| | <30 days | 6 |
| | 30-60 days | 4 |
| | 60-90 days | 2 |
| | >90 days | 0 |

---

### Health Bands

| Score | Status | Count Target | CSM Action |
|-------|--------|--------------|------------|
| 80-100 | 🟢 Healthy | 60% | Quarterly touch, expansion |
| 60-79 | 🟡 Monitor | 25% | Monthly touch, watch trends |
| 40-59 | 🟠 At Risk | 12% | Bi-weekly, intervention plan |
| 0-39 | 🔴 Critical | 3% | Weekly, executive escalation |

**Target Distribution** at 1:40 ratio:
- 24 Healthy (quarterly = 8 touches/month)
- 10 Monitor (monthly = 10 touches)
- 5 At Risk (bi-weekly = 10 touches)
- 1 Critical (weekly = 4 touches)
- **Total: 32 touch points/month** (manageable)

---

### Alert Rules (Noise Reduction)

**Only Alert When:**

| Alert | Condition | Frequency |
|-------|-----------|-----------|
| 🔴 Critical New | Score drops below 40 | Immediate |
| 🔴 Renewal Risk | Score <60 AND renewal <60 days | Immediate |
| 🟠 Score Drop | Drops >15 points in 7 days | Daily digest |
| 🟡 Watch | Usage down >30% WoW | Weekly digest |

**Suppress Alerts For:**
- Accounts <30 days old (onboarding)
- Accounts in active implementation
- Known seasonal low periods
- Manual "snooze" by CSM

---

### Dashboard Views

**Portfolio Summary:**

┌─────────────────────────────────────┐ │ 我的客户组合：40 个账户 │ ├─────────────────────────────────────┤ │ 🟢 24 (60%) │ 🟡 10 (25%) │ │ 🟠 5 (12%) │ 🔴 1 (3%) │ ├─────────────────────────────────────┤ │ ⚠️ 本周有 3 个账户评分下降 │ │ 📈 5 个账户有所改善 │ └─────────────────────────────────────┘

**Account Drill-Down:**

┌─────────────────────────────────────┐ │ 账户：TechCorp | 评分：62 🟡 │ ├─────────────────────────────────────┤ │ 产品： 28/40 ████████░░ │ │ 支持： 12/15 █████████░ │ │ 财务： 14/25 ██████░░░░ │ │ 关系： 8/20 ████░░░░░░ │ ├─────────────────────────────────────┤ │ 趋势：↘️ -8 分（30 天） │ │ 续约：90 天 │ │ 行动：安排跟进沟通 │ └─────────────────────────────────────┘

示例 2：健康度趋势分析

输入：

Analyze health trends for my portfolio:

Q4 Scores:
- Oct: 23 Green, 12 Yellow, 4 Orange, 1 Red
- Nov: 20 Green, 14 Yellow, 5 Orange, 1 Red
- Dec: 18 Green, 13 Yellow, 7 Orange, 2 Red

Also:
- 3 accounts churned in Q4
- 2 expansions closed
- NPS survey: 42 average (was 48)

输出：

## Q4 Health Trend Analysis

### Portfolio Drift Summary

| Status | Oct | Nov | Dec | Change |
|--------|-----|-----|-----|--------|
| 🟢 Green | 23 (58%) | 20 (50%) | 18 (45%) | ↘️ -13% |
| 🟡 Yellow | 12 (30%) | 14 (35%) | 13 (33%) | → +3% |
| 🟠 Orange | 4 (10%) | 5 (13%) | 7 (18%) | ↗️ +8% |
| 🔴 Red | 1 (2%) | 1 (2%) | 2 (5%) | ↗️ +3% |

**Trend: ⚠️ Deteriorating**
- 5 accounts moved from Green to lower status
- Orange/Red grew from 12% to 23%
- Portfolio health declining month-over-month

---

### Churn Correlation

| Churned Account | Last Score | Days at Red |
|-----------------|------------|-------------|
| Account A | 28 | 45 days |
| Account B | 35 | 30 days |
| Account C | 41 | 22 days |

**Insight**: All churned accounts were Red/Orange for 20+ days
**Action**: Accounts at Orange >14 days need intervention

---

### Key Drivers of Decline

**Analyzing accounts that dropped:**

| Factor | Accounts Affected | Avg Point Drop |
|--------|-------------------|----------------|
| Usage decline | 8 | -12 pts |
| Champion change | 3 | -18 pts |
| Support issues | 4 | -8 pts |
| Payment delays | 2 | -6 pts |

**Primary Driver**: Usage decline (likely seasonal + holiday)

---

### NPS Correlation

| NPS Segment | Avg Health Score | Q4 Change |
|-------------|------------------|-----------|
| Promoters (9-10) | 78 | -3 |
| Passives (7-8) | 58 | -6 |
| Detractors (0-6) | 38 | -10 |

**Insight**: Detractor scores dropping fastest
**Action**: Prioritize intervention for Detractors

---

### Q1 Recommendations

**Immediate (Week 1):**
1. Save plan for 2 Red accounts
2. Intervention for 7 Orange accounts
3. Outreach to 3 champion-change accounts

**Short-term (Month 1):**
1. Re-engagement campaign for low-usage accounts
2. Proactive support reach-out to ticket-heavy accounts
3. NPS follow-up calls with Detractors

**Strategic (Quarter):**
1. Investigate seasonal patterns (plan for Q4 2026)
2. Champion backup program implementation
3. Revisit Orange threshold (too many?)

---

### Target for Q1

| Status | Dec | Q1 Target | Delta |
|--------|-----|-----------|-------|
| 🟢 Green | 18 | 22 | +4 |
| 🟡 Yellow | 13 | 14 | +1 |
| 🟠 Orange | 7 | 3 | -4 |
| 🔴 Red | 2 | 1 | -1 |

**Success = Move 5 accounts up at least one tier**

技能边界

此技能擅长的事项

• 设计健康度框架
• 计算综合评分
• 识别趋势和模式
• 设置警报阈值

此技能无法完成的事项

• 访问你的实际数据
• 在你的系统中实施
• 了解你的具体业务规则
• 替代数据工程

需要升级至人工处理的情形

• 阈值决策
• 基于流失数据校准权重
• 调整警报规则
• 跨职能协同

迭代指南

后续提示词

• “对于企业客户和 SMB，我应该如何采用不同的维度权重？”
• “对于按使用量计费的定价模式，我应该添加哪些指标？”
• “创建可将噪声减少 50% 的警报规则。”
• “为高接触度服务型业务设计健康度评分。”

参考资料

• Gainsight 健康度评分最佳实践
• Totango 客户健康度方法论
• ChurnZero 评分框架
• 客户成功基准

相关技能

• churn-prediction - 深入分析流失情况
• account-health - RevOps 视角
• expansion-signals - 聚焦增长

技能元数据

• 领域：客户成功
• 复杂度：高级
• 模式：centaur
• 价值实现时间：框架设计需 2-4 小时
• 前置条件：评估数据可用性