Consistency Plan Template

Use this template for one action experiment with an end date. It supplements the Rhythm Ledger: the ledger tracks weekly capacity and obstacles, while this experiment card specifies one change, its acceptance criteria, and conditions for stopping. If you already maintain a ledger, reference the relevant records instead of copying them. Return the conclusion to the ledger when the experiment ends.

This is a practical tool, not a psychological assessment or a guarantee that a habit will form within a fixed number of days. Fill in what you actually know and leave unknowns for verification. There is no need to add tasks merely to complete the form.

Read Discipline and Consistency in Practice to choose a problem to investigate. Keep using the arrangements in Daily Systems and Rhythm and Compounding instead of maintaining another schedule. Keep state and work locations in your Learning State, and put weekly conclusions in the Weekly Review. Store personal records in your own notes and use evidence locations that do not disclose sensitive information.

Copyable Template

# Consistency Plan — [One Goal]

Updated: YYYY-MM-DD
Cycle: YYYY-MM-DD to YYYY-MM-DD
Review date or condition:

## Goal and Evidence
- Why this is worth doing now:
- The one outcome for this cycle:
- Current baseline and sample location:
- Acceptance criteria that can be observed or checked:
- Explicitly outside this cycle's scope:

## Realistic Budget
- Existing work, caring, and other responsibilities:
- Available windows and total time for this cycle:
- Time for preparation, execution, and feedback:
- Buffer window, if any:
- Time or resources I will not reallocate:
- Conditions requiring agreement or still uncertain:

## Starting Point
- After [an identifiable event]:
- At [an available place], open [resource/file]:
- First do [one specific action]:
- Main obstacle and the one change for this cycle:
- Contacts or alerts that must remain available:
- Next step to save before finishing:

## Three Versions
| Version | When to use it | Action and stopping condition | What to save |
| --- | --- | --- | --- |
| Minimum | | | |
| Standard | | | |
| Extended, optional | | | |

- Practice or validation the minimum version cannot replace:
- Maximum time per session:
- Conditions for reducing the workload:
- Conditions for pausing:

## Restart After an Interruption
- State file and most recent work locations:
- How to reassess ability at the first restart:
- First bounded task after the break:
- Next reassessment date or condition:
- How to adjust scope or negotiate if a deadline is affected:

## Cycle Log
| Date | Planned action | Actual output and location | Obstacle/cost | Next step |
| --- | --- | --- | --- | --- |

## Cycle Review
- What changed from the baseline, with evidence:
- What I still cannot independently explain or complete:
- Whether the arrangement exceeded the budget or displaced responsibilities:
- Conditions for continuing:
- Conditions for changing the method:
- Conditions for pausing or stopping:
- Decision for this cycle and supporting reasons:
- The one change for the next cycle:

Worked Example: Improving an English Project Introduction in Four Weeks

This example is fictional. Its schedule and workload only demonstrate how to fill in the template. The learner needs to explain an AI note-organizing tool in English during a team discussion. The plan takes work and family responsibilities into account.

# Consistency Plan — English Project Introduction

Updated: 2026-09-19
Cycle: 2026-09-21 to 2026-10-18
Review date or condition: Every Sunday, or earlier if work arrangements change.

## Goal and Evidence
- Why this is worth doing now: Explain the project's purpose and limitations at the next team discussion.
- The one outcome for this cycle: Independently give a 90-second introduction and answer two follow-up questions.
- Current baseline and sample location: The first recording lasts about 50 seconds; examples/baseline.m4a.
- Acceptance criteria: A listener can restate the purpose, intended user, and one limitation. Answer two follow-ups in my own words and explicitly acknowledge unknowns.
- Explicitly outside this cycle's scope: Accent imitation, a full product pitch, and new feature development.

## Realistic Budget
- Existing work, caring, and other responsibilities: Family commitments make weekday evenings unpredictable.
- Available windows and total time for this cycle: Two 20-minute lunches and one 20-minute weekend session per week; 240 minutes across four weeks.
- Time for preparation, execution, and feedback: 10 / 35 / 15 minutes per week.
- Buffer window: At most one extra 10-minute window each week. It may not be available and is excluded from the basic budget.
- Time or resources I will not reallocate: Agreed family time; no staying up late to make up missed sessions.
- Conditions requiring agreement or still uncertain: Agree on the weekend recording window in advance; check whether a feedback partner is available.

## Starting Point
- After lunch on Tuesday and Thursday, if the break is still available:
- At my desk, open notes/project-intro.md:
- First do: Write one point I want to explain today.
- Main obstacle and the one change for this cycle: Materials are scattered; use this single outline.
- Contacts or alerts that must remain available: Work on-call messages and urgent family contact.
- Next step to save before finishing: The unclear point and the example to use next time.

## Three Versions
| Version | When to use it | Action and stopping condition | What to save |
| --- | --- | --- | --- |
| Minimum | Only 5 minutes available | Write three explanatory sentences and flag one uncertain expression | Three sentences and one question |
| Standard | 20 minutes available | Revise the outline at lunch; on the weekend, record, listen, and correct one point | Outline or recording, with a revision note |
| Extended | Weekly tasks are done and the buffer is available | Answer one follow-up using different wording; stop at 10 minutes | One question-and-answer record |

- Practice or validation the minimum version cannot replace: Speaking aloud, responding to spontaneous questions, and feedback from an actual listener.
- Maximum time per session: 20 minutes, with a separate 10-minute limit for the extended version.
- Conditions for reducing the workload: Unexpected work takes one window; retain only one key point.
- Conditions for pausing: No available windows this week; save the state for next week's review.

## Restart After an Interruption
- State file and most recent work locations: notes/learning-state.md; recordings/.
- How to reassess ability at the first restart: Speak for 30 seconds without the outline, compare with the baseline, and record gaps.
- First bounded task after the break: Explain one use of the tool, save the recording, and stop.
- Next reassessment date or condition: The next Sunday, or after the work roster is confirmed.
- How to adjust scope or negotiate if a deadline is affected: Use a shorter introduction and agree with the team in advance; do not claim the full practice plan is complete.

## Cycle Log
| Date | Planned action | Actual output and location | Obstacle/cost | Next step |
| --- | --- | --- | --- | --- |
| 2026-09-22 | Write points about purpose and users | Three sentences in notes/project-intro.md | An unexpected message left 5 minutes; used the minimum version | Check one expression next time |
| 2026-09-26 | Make the weekend recording | recordings/week-1.m4a, 60 seconds | The limitation was unclear; stayed within budget | First write one specific limitation |

## Cycle Review
- What changed from the baseline, with evidence: The first week's recording includes purpose and users; listener understanding has not yet been checked.
- What I still cannot independently explain or complete: The tool's limitation; the two spontaneous follow-ups have not been practiced.
- Whether the arrangement exceeded the budget or displaced responsibilities: Within budget in week one, but Tuesday's window lasted only 5 minutes.
- Conditions for continuing: The goal remains useful, windows remain available, and recordings or feedback show specific improvements.
- Conditions for changing the method: If practice remains memorized recitation, add explanations without notes and feedback.
- Conditions for pausing or stopping: Pause if windows disappear; reassess the purpose if the team discussion is canceled.
- Decision for this cycle and supporting reasons: Continue into the next week and address the limitation first. Current records cannot establish the four-week outcome.
- The one change for the next cycle: Ask a feedback partner to restate the limitation in one sentence. If they are unavailable, record that feedback is missing.

The first week’s records show an incomplete result: the goal has not yet been met, but the next step is clear. The final review still needs four weeks of evidence. The plan itself is not evidence of its effectiveness.

Ask AI to Organize Evidence, Not Invent Progress

Below are my consistency plan and actual records for this cycle. Check for contradictions in the time budget, whether acceptance criteria are observable, and whether I am treating the minimum version as a substitute for complete practice. Identify evidence gaps. Do not invent completed tasks or feedback. Respect the existing responsibilities and suggest only one adjustment for the next cycle. If key facts are missing, clearly identify what I need to verify.

[Paste the plan and records]

Before accepting a suggestion, check your calendar, actual work, and other people’s availability. Revising the plan to fit reality is part of using this template.