跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
数据与分析

analytics

Product analytics implementation — event tracking, funnel analysis, A/B testing, Segment/Amplitude/Mixpanel/PostHog, privacy-compliant instrumentation.

arbazkhan971 · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add arbazkhan971/godmode --skill analytics

技能说明


name: analytics description: Product analytics implementation — event tracking, funnel analysis, A/B testing, Segment/Amplitude/Mixpanel/PostHog, privacy-compliant instrumentation.

Analytics — Analytics Implementation

Activate When

  • User invokes /godmode:analytics
  • User says "add analytics", "track events", "set up tracking"
  • User says "create a funnel", "A/B test", "implement experiments"
  • User says "add Segment", "set up Amplitude", "configure Mixpanel", "add PostHog"
  • User says "privacy-friendly analytics", "GDPR-compliant tracking", "cookieless analytics"
  • When building new features that need usage measurement
  • When /godmode:plan identifies analytics instrumentation tasks

Workflow

Step 1: Analytics Strategy Discovery

Understand what needs to be measured and why:

ANALYTICS DISCOVERY:
Project: <name and purpose>
Goals:
  - <business goal 1 — e.g., increase conversion rate>
  - <business goal 2 — e.g., reduce churn>
  - <business goal 3 — e.g., improve feature adoption>
Key questions to answer:
  - <question 1 — e.g., "Where do users drop off in onboarding?">
  - <question 2 — e.g., "Which features correlate with retention?">
  - <question 3 — e.g., "What is our activation rate?">
Platform: <web | mobile | both | server-side>
Framework: <React | Next.js | Vue | React Native | iOS | Android | backend>
Privacy requirements: <GDPR | CCPA | HIPAA | none | strict — no third-party>
Existing analytics: <none | GA4 | Segment | Amplitude | Mixpanel | custom>
Budget: <free tier | startup plan | enterprise>

If the user hasn't specified, ask: "What do you want to learn about your users? What decisions will this data inform?"

Step 2: Analytics Platform Selection

Choose the right analytics stack:

PLATFORM SELECTION:
| Platform | Best For | Privacy Model |
|--|--|--|
| Segment | Data routing hub, | Third-party, consent needed |
|  | multi-destination CDP | GDPR tools available |
| Amplitude | Product analytics, | Third-party, consent needed |
|  | behavioral cohorts, | SOC 2, GDPR compliant |
|  | journey mapping |  |
| Mixpanel | Event analytics, | Third-party, consent needed |
|  | funnel analysis, | EU data residency available |
|  | retention tracking |  |

Step 3: Event Taxonomy Design

Design a structured, consistent event naming system:

EVENT TAXONOMY:
Naming convention: <Object Action> (e.g., "Button Clicked", "Page Viewed")

Format: <object>_<action> (snake_case) or <Object> <Action> (Title Case)
Selected: <format>

EVENT CATALOG:
| Event Name | Trigger | Properties |
  LIFECYCLE EVENTS
| User Signed Up | Server | method, referral_source, |
|  |  | plan |
| User Logged In | Server | method, mfa_used |

Step 4: Event Property Standards

Define property types and validation:

PROPERTY STANDARDS:

GLOBAL PROPERTIES (sent with every event):
  user_id: string (anonymous ID or authenticated user ID)
  session_id: string (unique per session)
  timestamp: ISO 8601 datetime
  platform: "web" | "ios" | "android"
  app_version: string (semver)
  device_type: "desktop" | "tablet" | "mobile"
  browser: string (web only)
  os: string
  locale: string (BCP 47 — e.g., "en-US")
  experiment_ids: string[] (active A/B test assignments)

USER PROPERTIES (set once, updated on change):

Step 5: Implementation

Implement analytics tracking in the codebase:

Provider Setup

Configure the selected provider SDK (Segment, Amplitude, PostHog, etc.) with environment-keyed initialization. Build a unified abstraction layer (AnalyticsProvider interface with track, identify, page methods) so you swap providers without touching component code.

Step 6: Funnel Analysis Setup

Design and instrument conversion funnels:

FUNNEL DESIGN:
Funnel name: <name — e.g., "Onboarding Funnel">
Goal: <what conversion means — e.g., "User completes onboarding">

FUNNEL STEPS:
| Step | Name | Event | Expected % |
|--|--|--|--|
| 1 | Visit landing page | Page Viewed | 100% |
| 2 | Click sign up | CTA Clicked | 30-40% |
| 3 | Complete registration | User Signed Up | 60-70% |
| 4 | Start onboarding flow | Onboarding Started | 80-90% |
| 5 | Complete onboarding | Onboarding Done | 50-60% |
| 6 | First core action | Feature Used | 40-50% |
| 7 | Activation (aha moment) | Activation Done | 30-40% |

Step 7: A/B Test Instrumentation

Design and instrument experiments:

EXPERIMENT DESIGN:
Name: <experiment name>
Hypothesis: <if we change X, then Y will improve by Z%>
Primary metric: <metric to optimize — e.g., conversion rate>
Secondary metrics: <guardrail metrics to monitor — e.g., session duration, error rate>
Minimum detectable effect: <smallest meaningful change — e.g., 5% relative improvement>
Required sample size: <calculated based on MDE, baseline, significance level>
Duration: <estimated experiment duration>
Significance level: alpha = 0.05
Power: 1 - beta = 0.80

VARIANTS:
| Variant | Description | Traffic % |

Experiment Implementation

// experiments/ab-test.ts
interface Experiment {
  id: string;
  name: string;
  variants: { id: string; weight: number }[];
  isActive: boolean;

Step 8: Analytics Data Modeling

DATA MODEL (3 core tables):
  EVENTS: event_id (PK), event_name (indexed), user_id (indexed), session_id,
          timestamp (partitioned by day), properties (JSONB), context (JSONB)
  USERS:  user_id (PK), traits (JSONB), first_seen, last_seen, event_count
  SESSIONS: session_id (PK), user_id (indexed), started_at, ended_at,
            duration_sec, entry_page, exit_page, device_type, utm_*

COMMON QUERIES:
  - DAU/WAU/MAU: COUNT(DISTINCT user_id) WHERE timestamp >= <period>
  - Retention: cohort analysis grouping by first_seen week
  - Funnel: sequential event matching with time constraints
  - Feature adoption: COUNT(DISTINCT user_id) WHERE event_name = '<feature>'

Step 9: Privacy & Consent

Implement privacy-compliant analytics:

PRIVACY IMPLEMENTATION:
| Requirement | Impleme
阅读完整来源 ↗