Auto-sync: 2026-04-21 08:02
This commit is contained in:
30
wiki/concepts/Analytics-Feedback-Loop.md
Normal file
30
wiki/concepts/Analytics-Feedback-Loop.md
Normal file
@@ -0,0 +1,30 @@
|
||||
---
|
||||
title: "Analytics Feedback Loop"
|
||||
type: concept
|
||||
tags: [marketing, data-analysis, optimization, ai-agent]
|
||||
sources: [marketing-carousel-growth-engine]
|
||||
last_updated: 2026-04-21
|
||||
---
|
||||
|
||||
## Definition
|
||||
数据驱动的持续优化循环,通过收集、分析和应用性能数据不断提升内容效果。
|
||||
|
||||
## Cycle Process
|
||||
1. **Fetch**:获取分析数据(Upload-Post API)
|
||||
2. **Extract**:提取洞察(表现最佳的 hook、发布时间、视觉风格)
|
||||
3. **Update**:更新 learnings.json 知识库
|
||||
4. **Plan**:规划下一个轮播内容
|
||||
|
||||
## Tracked Metrics
|
||||
- Hook 性能:问题型 vs 声明型 vs 痛点型
|
||||
- 发布时机:最佳日期/小时
|
||||
- 视觉风格:相关 slide prompts 与 engagement 关联
|
||||
- 参与率:likes + comments + shares / views
|
||||
|
||||
## Storage
|
||||
- `/tmp/carousel/learnings.json`
|
||||
- 滚动 100 期历史记录
|
||||
|
||||
## Aliases
|
||||
- 分析反馈循环
|
||||
- 数据驱动优化
|
||||
Reference in New Issue
Block a user