← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6793.e · 生成模型:GPT-6

满意与留存要按任务解释

理解等级:中级

满意反馈和推荐意愿是不同于使用频率的信号,留存也需按同一批用户与时间窗计算。教育服务可关注持续回来,客服反复返回却可能意味着未解决;选择指标不必五类等权,更不能把NPS当全部幸福程度。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      Happiness metrics are pretty straightforward: they attempt to measure the user’s satisfaction with the chatbot. But just like with human happiness, user happiness is not easily defined and measured. In most cases, to know how a user feels about the bot, you will need to proactively ask them about their experience. One common measure of happiness is the net promoter score (NPS), which is calculated

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第12章《对话引擎与会话设计》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-12.html) - 结论依据:摘录中关于满意与留存要按任务解释的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:Happiness metrics are pretty straightforward: they attempt to measure the user’s satisfaction with the chatbot. But just like with human happiness, user happiness is not easily defined and measured. In most cases, to know how a user feels about the bot, you will need to proactively ask them about their experience. One common measure of happiness is the net promoter score (NPS), which is calculated