feature-prioritization
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChinesePrioritize with math, not opinions. RICE scoring forces explicit tradeoffs. The enabler/blocker lens from Linear ensures you're not just building fun things while adoption barriers remain.
用数据而非主观意见来确定优先级。RICE评分模型要求明确做出取舍。Linear的「推动者/阻碍者」视角则能确保你不会在仍存在用户采用障碍的情况下,只去开发那些有趣的功能。
RICE Scoring
RICE评分模型
Score every candidate feature on four dimensions:
- Reach: How many users/accounts will this affect in a set time period? Use real numbers from analytics, not gut feel. "500 users/quarter" not "a lot."
- Impact: How much will this move the target metric per user? Score 0.25 (minimal), 0.5 (low), 1 (medium), 2 (high), 3 (massive). Be honest — most features are a 1.
- Confidence: How sure are you about Reach and Impact? 100% = hard data. 80% = strong evidence. 50% = gut feel. NEVER score 100% without quantitative data.
- Effort: Person-weeks of work. Include design, engineering, QA, and any cross-team coordination. Round up.
RICE = (Reach x Impact x Confidence) / Effort
Example: SSO — Reach: 500 users/qtr, Impact: 2 (high — unlocks enterprise deals), Confidence: 80%, Effort: 4 person-weeks. RICE = (500 x 2 x 0.8) / 4 = 200. Tag: Blocker.
Rank by score. The math won't be perfect, but it forces you to justify each dimension.
为每个候选功能从四个维度进行评分:
- 覆盖范围(Reach): 在特定时间段内,该功能会影响多少用户/账户?请使用分析数据中的真实数字,而非直觉判断。例如写「每季度500名用户」,而非「很多用户」。
- 影响程度(Impact): 该功能每位用户能让目标指标提升多少?评分标准:0.25(极小)、0.5(低)、1(中等)、2(高)、3(极大)。请如实评分——大多数功能的影响程度为1。
- 置信度(Confidence): 你对覆盖范围和影响程度的判断有多确定?100% = 有确凿数据;80% = 有充分证据;50% = 凭直觉。没有量化数据的情况下,绝对不要给100%的评分。
- 投入成本(Effort): 以人周为单位的工作量。包括设计、开发、QA以及跨团队协调的时间。请向上取整。
RICE得分 =(覆盖范围 × 影响程度 × 置信度)÷ 投入成本
示例:SSO功能 —— 覆盖范围:每季度500名用户,影响程度:2(高——能促成企业客户交易),置信度:80%,投入成本:4人周。RICE得分 =(500 × 2 × 0.8)÷ 4 = 200。标签:阻碍者。
按得分进行排名。这个计算结果并非绝对完美,但它能迫使你为每个维度的评分提供依据。
Enablers vs Blockers (Linear)
Linear的「推动者vs阻碍者」视角
After RICE scoring, classify each feature:
- Blocker: Removes a barrier to adoption or retention. Users are churning, stuck, or can't even start because this is missing. Examples: missing SSO for enterprise deals, broken mobile experience, no data export.
- Enabler: Delights existing users or deepens engagement. Users are already successful but this makes them more so. Examples: keyboard shortcuts, advanced filters, integrations.
Rule: Prioritize blockers over enablers when growing. Removing friction > adding delight when you're trying to grow. Flip this when retention is strong but engagement is flat.
完成RICE评分后,对每个功能进行分类:
- 阻碍者(Blocker): 消除用户采用或留存的障碍。用户会因为缺少该功能而流失、停滞不前甚至无法开始使用产品。例如:企业客户所需的SSO功能缺失、移动端体验糟糕、无数据导出功能。
- 推动者(Enabler): 为现有用户带来愉悦体验或加深用户参与度。用户已经能成功使用产品,但该功能能让他们获得更好的体验。例如:键盘快捷键、高级筛选功能、第三方集成。
规则:在增长阶段,优先处理阻碍者而非推动者。 当你试图实现增长时,消除摩擦比增添愉悦更重要。如果用户留存表现良好但参与度平平,则可以反过来优先处理推动者。
Running a Prioritization Session
优先级排序会议流程
- List all candidates with a one-sentence description
- Score each on R, I, C, E independently — don't anchor on each other
- Calculate RICE scores and rank
- Tag each as Blocker or Enabler
- Check: are any Blockers ranked below Enablers? Justify or re-rank.
- Top 3-5 items = your next cycle
- 列出所有候选功能,每个功能配一句描述
- 独立为每个功能的R(覆盖范围)、I(影响程度)、C(置信度)、E(投入成本)评分——不要互相影响
- 计算RICE得分并进行排名
- 将每个功能标记为「推动者」或「阻碍者」
- 检查:是否有阻碍者的排名低于推动者?如有,请说明理由或重新排名
- 排名前3-5的项即为你下一周期的开发内容
Guidelines
注意事项
- ALWAYS show the math. A prioritization without visible scores is just an opinion list.
- NEVER let Confidence = 100% without hard data (analytics, A/B test results, or customer commit).
- ALWAYS separate enablers from blockers before finalizing the rank.
- NEVER prioritize more than one cycle at a time. The world changes. Re-score next cycle.
- CRITICAL: If two items are within 20% RICE score of each other, they're effectively tied. Use the blocker/enabler lens to break ties, or pick the one with higher Confidence.
Built on Intercom's RICE framework and the Linear Method. Skills from productskills.
- 务必展示计算过程。没有明确评分的优先级列表本质上只是主观意见清单。
- 没有确凿数据(分析数据、A/B测试结果或客户承诺)的情况下,绝对不要将置信度设为100%。
- 在最终确定排名前,务必将推动者和阻碍者分开归类。
- 绝对不要一次确定超过一个周期的优先级。市场环境会变化,下一周期请重新评分。
- 关键提示:如果两个功能的RICE得分差距在20%以内,则视为并列。此时可通过「推动者/阻碍者」视角来打破平局,或选择置信度更高的功能。
基于Intercom的RICE框架和Linear方法构建。技能源自productskills。