Series ID: EDKS-BASIC-ZH-0119 · Lesson No.119 · Expert Maintenance: Governance and Accountability
Decision Logs, Governance and Accountability | 决策记录、治理与问责
Good governance makes decisions inspectable: what was decided, by whom, using which evidence, under which authority, and with what follow-up.
Decision logs preserve reasoning so future teams can distinguish a bad process from bad luck, revisit assumptions when conditions change, and understand why a seemingly obvious alternative was not selected.
高质量治理写作不是增加官僚流程，而是让决策可追踪：谁有权限？当时知道什么？考虑过哪些方案？为什么选择这一条？谁负责执行？什么时候必须重新评估？
Chinese Edition Hub · 中文版入口 · ← Lesson No.118 · Writing Change Plans, Rollouts and Rollbacks
The governance map | 治理地图
- decision
- authority
- evidence
- alternatives
- rationale
- challenge
- owner
- accountability
- review trigger
- audit trail
1. Decision | 决策
State the decision in one sentence.
2. Decision required | 需要决定什么
Before the meeting, define the exact choice that needs authority.
3. Authority | 权限
Who is authorised to make the decision?
4. Recommendation vs decision | 建议 vs 决定
An expert may recommend; a manager, committee or owner may decide.
5. Decision owner | 决策责任人
The person or body accountable for the choice.
6. Implementation owner | 执行责任人
May be different from the decision owner.
7. Evidence available at the time | 当时可获得的证据
Record what was known when the decision was made.
8. Hindsight bias | 后见之明偏误
Do not judge the original decision using evidence that appeared later without separating the two.
9. Assumptions | 假设
Record assumptions that materially support the decision.
10. Confidence | 置信度
State where evidence was strong, mixed or incomplete.
11. Alternatives | 备选方案
Record credible options, including status quo where relevant.
12. Rejected alternative | 未选方案
Document why it was rejected if future teams may revisit it.
13. Criteria | 判据
What dimensions were used to compare options?
- risk
- cost
- speed
- reliability
- fairness
14. Weighting | 权重
If one criterion mattered more, state that explicitly.
15. Rationale | 理由
Connect evidence + criteria + decision.
16. Weak rationale | 弱理由
We chose B because it seemed best.
17. Stronger rationale | 更强理由
We chose B because reliability was the mandatory criterion, and B was the only option meeting the agreed threshold during the pilot.
18. Challenge | 质疑机制
Good governance allows material challenge before final commitment.
19. Dissent | 异议
Record significant dissent where it affects future review.
20. Dissent ≠ disloyalty | 异议不等于不合作
A documented challenge can improve decision quality.
21. Conflict of interest | 利益冲突
Where applicable, disclose interests that could affect judgement.
22. Recusal | 回避
In some governance settings, a conflicted person may need to abstain from a decision.
23. Governance framework | 治理框架
Use the organisation’s actual framework; do not invent formal-sounding roles.
24. RACI-style roles | 责任角色
Some teams distinguish:
- Responsible
- Accountable
- Consulted
- Informed
Use only if that model helps and roles are defined.
25. Responsible vs accountable | 执行 vs 问责
Responsible does the work; accountable owns the outcome/decision in many frameworks.
26. Avoid role ambiguity | 避免角色模糊
If two people believe the other owns the decision, governance has failed.
27. Decision log | 决策日志
| Field | Example |
|---|---|
| Decision | Proceed with staged rollout |
| Owner | Programme lead |
| Evidence | Pilot + risk review |
| Review trigger | Error rate > 3% |
28. Date/time | 日期时间
Record when the decision became effective.
29. Version | 版本
If the decision supersedes an older one, link the history.
30. Decision status | 决策状态
- proposed
- approved
- implemented
- superseded
- retired
31. Review date | 复审日期
Some decisions should expire or be revisited.
32. Review trigger | 复审触发器
Conditions can be more useful than arbitrary dates.
33. Sunset clause | 到期机制
Temporary rules can end automatically unless renewed.
34. Exception approval | 例外批准
Document who can approve exceptions and under what conditions.
35. Exception log | 例外记录
Repeated exceptions may indicate the rule itself needs revision.
36. Audit trail | 审计轨迹
Preserve enough record to reconstruct the decision process.
37. Audit trail ≠ transcript | 审计轨迹不等于逐字记录
Capture material evidence, decisions, owners and rationale.
38. Accountability | 问责
Accountability means ownership of actions and decisions—not automatic blame for every bad outcome.
39. Outcome vs decision quality | 结果 vs 决策质量
A sound decision under uncertainty can still produce an adverse outcome.
40. Process accountability | 过程问责
Was the agreed process followed?
41. Evidence accountability | 证据问责
Were material facts ignored, distorted or unverified?
42. Action accountability | 行动问责
Were agreed actions completed?
43. Learning accountability | 学习问责
Did the system update after new evidence or failure?
44. Governance without paralysis | 治理不能让系统瘫痪
Controls should be proportionate to decision stakes.
45. Low-risk reversible decision | 低风险可逆决策
May need lighter approval.
46. High-risk irreversible decision | 高风险不可逆决策
May justify stronger evidence, challenge and sign-off.
47. Escalation | 升级
Define when local authority is insufficient.
48. Delegation | 授权
Authority can be delegated, but boundaries should be explicit.
49. Delegation limit | 授权边界
State what the delegate may decide without further approval.
50. Governance communication | 治理沟通
People need to know:
- who decides
- who acts
- who reviews
- who must be informed
51. Decision announcement | 决策公告
State decision, rationale, effective date, owner and next step.
52. Do not relitigate after decision without trigger | 没有新触发条件不要无限重审
Stable governance needs closure.
53. But do not freeze bad decisions | 也不要把决定冻结
Review when evidence or conditions materially change.
54. Governance drift | 治理漂移
Informal habits can override formal roles over time.
55. Periodic governance review | 定期治理复盘
Check whether authority and process still match reality.
56. Mandarin transfer: “负责”可对应 responsible/accountable/owner | 要分清角色
57. Mandarin transfer: “批准”可对应 approve/authorise/sign off | 依权限选择
58. Practice A | 练习 A
Create a decision log for one real choice.
59. Practice B | 练习 B
Separate recommendation, decision and implementation ownership.
60. Practice C | 练习 C
Write a short dissent note that challenges evidence without attacking people.
61. Error clinic | 常见问题
| Problem | Repair |
|---|---|
| No clear decision owner. | Name authority. |
| Rationale vague. | Link evidence/criteria. |
| Rejected option forgotten. | Record alternatives. |
| Accountability becomes blame. | Separate process/action/outcome. |
| Old decision persists after context changed. | Add review trigger. |
62. First weak link diagnosis | 第一个卡点诊断
- people unsure who decides → authority.
- future team repeats debate → decision history.
- poor outcome creates blame → decision-quality distinction.
- actions not completed → ownership.
- bad assumptions persist → review trigger.
63. Seven-day training cycle | 七天训练
| Day 1 | decision statement | 决策陈述 |
| Day 2 | authority/roles | 权限角色 |
| Day 3 | rationale | 理由 |
| Day 4 | alternatives/dissent | 备选异议 |
| Day 5 | accountability | 问责 |
| Day 6 | review triggers | 复审 |
| Day 7 | full decision log | 综合 |
64. Self-test | 自测
Document a decision so a future reader can reconstruct authority, evidence, alternatives, rationale, dissent, ownership and review conditions.
65. For parents and teachers | 给家长和老师
Students can practise governance through group projects: clarify who decides, who does the work, and how disagreements are recorded.
66. Final real-world challenge | 最终真实任务
- Choose one real decision.
- State decision required.
- Name authority.
- List evidence.
- List alternatives.
- State criteria.
- Write rationale.
- Record dissent.
- Assign actions.
- Set review trigger.
Next: Lesson No.120 | 下一课
The next lesson integrates Lessons 101–119 into one operating capstone: research, risk, change, governance, cross-cultural coordination, documentation, recovery and long-term maintenance in one complex English workflow.
Lesson No.120 · Expert Maintenance Capstone: Operating in English Across Systems · 专家维护顶点：用英语跨系统运作
Reference floor: expert-maintenance governance. Decision records preserve authority, evidence, rationale and accountability without confusing a bad outcome with an automatically bad decision.
