analytics-metrics-kpi

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Analytics & Metrics Skill

Analytics & Metrics 技能

Become data-driven. Define meaningful metrics, build dashboards, run experiments, and make decisions based on data, not intuition.
成为数据驱动型从业者。定义有意义的指标、搭建仪表盘、开展实验,基于数据而非直觉做出决策。

Metrics Framework (Acquisition → Revenue)

指标框架(获客 → 营收)

North Star Metric

北极星指标

Definition: One metric that best captures the value your product delivers.
Characteristics:
  • Directly tied to business success
  • Driven by product improvements
  • Leading indicator of revenue
  • Understandable to whole company
Examples:
  • Slack: Daily Active Users (DAU)
  • Airbnb: Booked Nights
  • YouTube: Watch Time
  • Uber: Rides Completed
  • Stripe: Payment Volume Processed
定义: 最能体现产品所传递价值的单一指标。
特征:
  • 与业务成功直接挂钩
  • 受产品优化驱动
  • 是营收的领先指标
  • 全公司都能理解
示例:
  • Slack:日活跃用户(DAU)
  • Airbnb:预订晚数
  • YouTube:观看时长
  • Uber:完成订单量
  • Stripe:处理支付交易量

Funnel Metrics (Acquisition)

漏斗指标(获客)

Total Visitors: 100,000/month
↓ 20% conversion
Free Signups: 20,000
↓ 10% free-to-paid
Paid Customers: 2,000

CAC: $50 (marketing + sales spend / customers acquired)
LTCAC: $100 (all customer acquisition costs)
Metrics to Track:
  • Traffic - Total visitors to website/app
  • Signup Rate - % who sign up (target: 10-15%)
  • Free-to-Paid Conversion - % free users who pay (target: 2-5%)
  • CAC - Cost per acquired customer
  • CAC Payback - Months to recover CAC from revenue (target: < 12 months)
Total Visitors: 100,000/month
↓ 20% conversion
Free Signups: 20,000
↓ 10% free-to-paid
Paid Customers: 2,000

CAC: $50 (marketing + sales spend / customers acquired)
LTCAC: $100 (all customer acquisition costs)
需追踪的指标:
  • 流量 - 网站/应用的总访客数
  • 注册转化率 - 完成注册的用户占比(目标:10-15%)
  • 免费转付费转化率 - 付费的免费用户占比(目标:2-5%)
  • CAC - 单个获客成本
  • CAC回收期 - 从营收中收回CAC所需的月数(目标:<12个月)

Activation Metrics

激活指标

Goal: New users become active users
Free Signups: 2,000
↓ 30% onboard successfully
Activated: 600
↓ 60% remain active Day 7
Day 7 Active: 360
Metrics to Track:
  • Onboarding Completion Rate - % who complete setup (target: 50-80%)
  • Time to First Value - Hours to first successful use
  • Feature Adoption - % who try key features
  • Day 1/7/30 Retention - % active those days (target: 40/25/15)
目标: 让新用户成为活跃用户
Free Signups: 2,000
↓ 30% onboard successfully
Activated: 600
↓ 60% remain active Day 7
Day 7 Active: 360
需追踪的指标:
  • 新手引导完成率 - 完成设置的用户占比(目标:50-80%)
  • 首次价值实现时间 - 首次成功使用产品所需的小时数
  • 功能采用率 - 尝试核心功能的用户占比
  • 次日/7日/30日留存率 - 对应日期的活跃用户占比(目标:40/25/15)

Engagement Metrics

参与度指标

Goal: Users regularly use product
Daily/Monthly Metrics:
  • DAU/MAU - Daily/Monthly Active Users
  • DAU/MAU Ratio - Stickiness (target: 20-30%)
  • Feature Usage - % using key features
  • Session Length - Minutes per session
  • Session Frequency - Times per week
Cohort Analysis Example:
Jan Cohort (1,000 signups):
- Day 1: 600 active (60%)
- Day 7: 360 active (36%)
- Day 30: 180 active (18%)
- Month 3: 90 active (9%)

Feb Cohort (1,500 signups):
- Day 1: 1050 active (70%) ← Improving!
- Day 7: 630 active (42%)
- Day 30: 300 active (20%)
目标: 让用户定期使用产品
日/月指标:
  • DAU/MAU - 日/月活跃用户数
  • DAU/MAU比值 - 用户粘性(目标:20-30%)
  • 功能使用率 - 使用核心功能的用户占比
  • 会话时长 - 每次会话的分钟数
  • 会话频率 - 每周使用次数
同期群分析示例:
Jan Cohort (1,000 signups):
- Day 1: 600 active (60%)
- Day 7: 360 active (36%)
- Day 30: 180 active (18%)
- Month 3: 90 active (9%)

Feb Cohort (1,500 signups):
- Day 1: 1050 active (70%) ← 正在提升!
- Day 7: 630 active (42%)
- Day 30: 300 active (20%)

Retention Metrics

留存指标

Goal: Users stay and continue paying
Month 1: 1,000 customers
Month 2: 900 active (90% retained)
Month 3: 810 active (90% of month 2)
Month 12: 314 active (31% annual retention)
Churn Rate: % lost each period
  • Monthly churn: (Customers Lost / Month Start) × 100
  • Annual churn: 1 - (Ending / Starting)
  • Target for SaaS: < 5% monthly churn
NPS (Net Promoter Score)
  • Question: "How likely to recommend (0-10)?"
  • Score = % Promoters (9-10) - % Detractors (0-6)
  • Range: -100 to +100
  • Target: 50+ (world-class)
目标: 留住用户并持续付费
Month 1: 1,000 customers
Month 2: 900 active (90% retained)
Month 3: 810 active (90% of month 2)
Month 12: 314 active (31% annual retention)
流失率: 每个周期内流失的用户占比
  • 月度流失率:(流失用户数 / 月初用户数)×100
  • 年度流失率:1 -(期末用户数 / 期初用户数)
  • SaaS产品目标:月度流失率 <5%
NPS(净推荐值)
  • 问题:“你向他人推荐本产品的可能性有多大(0-10分)?”
  • 得分 = 推荐者占比(9-10分)- 贬损者占比(0-6分)
  • 范围:-100 至 +100
  • 目标:50+(世界级水平)

Revenue Metrics

营收指标

Monthly Recurring Revenue (MRR)
MRR = (Total paid customers) × (average subscription price)
Growth MRR = New MRR + Expansion MRR - Churn MRR
Annual Run Rate (ARR)
ARR = MRR × 12
Average Revenue Per User (ARPU)
ARPU = MRR / Total Users
Customer Lifetime Value (LTV)
LTV = (ARPU × Gross Margin %) / Monthly Churn %

Example:
ARPU: $100
Gross Margin: 80%
Monthly Churn: 5%
LTV = ($100 × 80%) / 5% = $1,600

If CAC = $400: LTV/CAC = 4x ✓ (target: 3x+)
月度经常性收入(MRR)
MRR = (付费用户总数) × (平均订阅价格)
Growth MRR = New MRR + Expansion MRR - Churn MRR
年度经常性收入(ARR)
ARR = MRR × 12
每用户平均收入(ARPU)
ARPU = MRR / 总用户数
客户生命周期价值(LTV)
LTV = (ARPU × 毛利率 %) / 月度流失率 %

示例:
ARPU: $100
Gross Margin: 80%
Monthly Churn: 5%
LTV = ($100 × 80%) / 5% = $1,600

如果CAC = $400: LTV/CAC = 4倍 ✓(目标:3倍+)

Dashboard Architecture

仪表盘架构

Executive Dashboard (C-Level)

高管仪表盘(C级)

Weekly Updates:
  • MRR / ARR (vs target, vs month ago)
  • New customers (weekly, monthly)
  • Churn rate (%)
  • NPS score
  • Engagement (DAU, MAU)
  • Key initiatives status
Frequency: Weekly
每周更新:
  • MRR / ARR(与目标对比、与上月对比)
  • 新增客户数(周度、月度)
  • 流失率(%)
  • NPS得分
  • 用户参与度(DAU、MAU)
  • 核心项目进度
更新频率: 每周

Product Dashboard (Product Team)

产品仪表盘(产品团队)

Daily/Weekly:
  • Funnel metrics (signup → paid)
  • Feature adoption
  • Engagement metrics
  • User feedback score
  • A/B test results
  • Support ticket volume
Frequency: Daily updates
每日/每周更新:
  • 漏斗指标(注册 → 付费)
  • 功能采用率
  • 用户参与度指标
  • 用户反馈得分
  • A/B测试结果
  • 支持工单数量
更新频率: 每日更新

Financial Dashboard (Finance/Operations)

财务仪表盘(财务/运营团队)

Monthly:
  • MRR / ARR
  • Customer acquisition cost
  • Customer lifetime value
  • Gross margin
  • CAC payback period
  • Revenue by segment
  • Churn by cohort
Frequency: Monthly
每月更新:
  • MRR / ARR
  • 客户获取成本
  • 客户生命周期价值
  • 毛利率
  • CAC回收期
  • 细分群体营收
  • 同期群流失情况
更新频率: 每月

Health Dashboard (Operations)

健康度仪表盘(运营团队)

Realtime:
  • System uptime (%)
  • Error rate (%)
  • Response time (p95)
  • Database performance
  • Support ticket response time
  • Support backlog
Frequency: Realtime/hourly
实时更新:
  • 系统可用性(%)
  • 错误率(%)
  • 响应时间(p95)
  • 数据库性能
  • 支持工单响应时间
  • 支持工单积压量
更新频率: 实时/每小时

A/B Testing (Experimentation)

A/B测试(实验)

Test Planning

测试规划

Hypothesis: "If we change X, then Y will improve, because Z"
Example: "If we move signup button above the fold, then conversion will improve 15%, because users won't scroll."
假设: “如果我们修改X,那么Y将得到提升,原因是Z”
示例: “如果我们将注册按钮移至首屏上方,转化率将提升15%,因为用户无需滚动查找。”

Test Structure

测试结构

Experiment Design:
  • Control: Keep current version
  • Treatment: New version
  • Sample size: Enough users to be statistical
  • Duration: 2-4 weeks minimum
  • Metric: Clear success metric
实验设计:
  • 对照组: 保留当前版本
  • 实验组: 新版本
  • 样本量: 足够大以具备统计意义
  • 时长: 至少2-4周
  • 指标: 明确的成功指标

Statistical Significance

统计显著性

Confidence Level: 95% (industry standard)
  • Means 5% chance of false positive
  • Need enough samples (typically 1000-10K per variant)
  • Use calculator for exact sample size
P-Value: Probability result is random chance
  • P < 0.05: Statistically significant
  • P > 0.05: Not significant, inconclusive
置信水平: 95%(行业标准)
  • 意味着有5%的概率出现假阳性结果
  • 需要足够的样本量(通常每个变体1000-10000用户)
  • 使用计算器确定精确样本量
P值: 结果由随机因素导致的概率
  • P < 0.05:具备统计显著性
  • P > 0.05:无显著性,结果不确定

Example A/B Test

A/B测试示例

Hypothesis: Moving signup button above fold increases conversion 15%
Setup:
  • Control: Current design
  • Treatment: Button moved above fold
  • Success metric: Conversion rate (signup / visit)
  • Sample size: 10,000 users per variant
  • Duration: 2 weeks
  • Confidence: 95%
Results:
  • Control: 2.0% conversion (200 signups from 10K visitors)
  • Treatment: 2.8% conversion (280 signups from 10K visitors)
  • Improvement: 40% increase (0.8% / 2% = 40%)
  • P-value: 0.02 (statistically significant!)
  • Decision: SHIP IT - Roll out to 100%
假设: 将注册按钮移至首屏上方可提升15%的转化率
设置:
  • 对照组:当前设计
  • 实验组:按钮移至首屏上方
  • 成功指标:转化率(注册数 / 访客数)
  • 样本量:每个变体10000用户
  • 时长:2周
  • 置信度:95%
结果:
  • 对照组:2.0%转化率(10000访客中200人注册)
  • 实验组:2.8%转化率(10000访客中280人注册)
  • 提升幅度:40%(0.8% / 2% = 40%)
  • P值:0.02(具备统计显著性!)
  • 决策:全量发布 - 推送给所有用户

Test Ideas by Priority

按优先级排序的测试方向

High Priority (Start Here):
  • Signup flow optimization (biggest funnel)
  • Onboarding experience
  • Pricing page clarity
  • Feature discoverability
Medium Priority:
  • UI copy optimization
  • CTA button colors
  • Email subject lines
  • Notification triggers
Low Priority:
  • Micro-copy tweaks
  • Animation effects
  • Color scheme changes
高优先级(从这里开始):
  • 注册流程优化(最大的漏斗环节)
  • 新手引导体验
  • 定价页面清晰度
  • 功能可发现性
中优先级:
  • UI文案优化
  • CTA按钮颜色
  • 邮件主题
  • 通知触发条件
低优先级:
  • 微文案调整
  • 动画效果
  • 配色方案变更

Metric Pitfalls to Avoid

需避免的指标陷阱

Vanity Metrics

虚荣指标

❌ "We have 1M page views!" ✓ "We have 50K daily active users, growing 10% monthly"
❌ “我们有100万页面浏览量!” ✓ “我们有5万日活跃用户,月增长10%”

Actionable vs Non-Actionable

可行动 vs 不可行动

❌ "User satisfaction increased" (what changed?) ✓ "Onboarding completion rate 65% → 78% (↑20%)" (clear action)
❌ “用户满意度提升了”(什么因素导致的?) ✓ “新手引导完成率从65%提升至78%(↑20%)”(明确可行动)

Correlation vs Causation

相关性 vs 因果性

❌ "Ice cream sales correlate with drownings" ✓ Understand actual causation, not just correlation
❌ “冰淇淋销量与溺水事件相关” ✓ 理解真实的因果关系,而非仅看相关性

Look-Alike Metrics

相似指标误区

❌ Track MRR but not Customer LTV (can grow MRR by spending more on acquisition) ✓ Track both acquisition efficiency AND retention
❌ 仅追踪MRR而不追踪客户LTV(可以通过增加获客投入来提升MRR) ✓ 同时追踪获客效率与留存情况

Metrics Review Cadence

指标回顾节奏

Daily:
  • System uptime
  • Error rates
  • Support response time
Weekly:
  • Funnel metrics
  • Feature adoption
  • Key engagement metrics
  • Test results
Monthly:
  • Revenue metrics
  • Cohort analysis
  • Churn breakdown
  • LTV/CAC trends
Quarterly:
  • Strategic metric review
  • Long-term trend analysis
  • Metric changes needed
每日:
  • 系统可用性
  • 错误率
  • 支持响应时间
每周:
  • 漏斗指标
  • 功能采用率
  • 核心参与度指标
  • 测试结果
每月:
  • 营收指标
  • 同期群分析
  • 流失情况细分
  • LTV/CAC趋势
每季度:
  • 战略指标回顾
  • 长期趋势分析
  • 必要的指标调整

Troubleshooting

故障排查

Yaygın Hatalar & Çözümler

常见错误与解决方案

HataOlası SebepÇözüm
Vanity metrics focusWrong KPI selectionNorth Star alignment
Inconclusive A/B testLow sample sizeExtend duration
Data inconsistencyMultiple sourcesSingle source of truth
Dashboard unusedToo complexSimplify to 5-7 KPIs
错误可能原因解决方案
关注虚荣指标KPI选择错误对齐北极星指标
A/B测试结果不确定样本量不足延长测试时长
数据不一致多数据源建立单一数据源
仪表盘无人使用过于复杂简化为5-7个核心KPI

Debug Checklist

调试检查清单

[ ] North Star metric defined mi?
[ ] Metrics business goals'a aligned mi?
[ ] Data collection accurate mi?
[ ] Dashboard refreshed mi?
[ ] A/B test sample sufficient mi?
[ ] Statistical significance achieved mi?
[ ] 是否已定义北极星指标?
[ ] 指标是否与业务目标对齐?
[ ] 数据收集是否准确?
[ ] 仪表盘是否已刷新?
[ ] A/B测试样本量是否充足?
[ ] 是否达到统计显著性?

Recovery Procedures

恢复流程

  1. Data Quality Issues → Flag affected metrics, exclude
  2. Inconclusive A/B → Extend test duration
  3. Misleading Metrics → Add context/segmentation

Master data-driven decision making and grow faster!
  1. 数据质量问题 → 标记受影响的指标,排除其影响
  2. A/B测试结果不确定 → 延长测试时长
  3. 指标存在误导性 → 添加上下文/细分维度

掌握数据驱动决策,实现更快增长!