Project Risk Log: Name the Early Signals Before Work Starts

project planning often looks like a single task. In practice, the outcome depends on a chain of small choices, and the earliest unchecked choice usually creates the most expensive correction. A useful productivity system converts an important result into a small observable behavior that survives a normal week.

Use this guide when a concrete project risk log: name the early signals before work starts decision is already on the table. It focuses on the sequence, evidence, and recovery path—not on claiming that one answer fits every material, person, location, organization, or appetite.

1. Define three observable outcomes. 2. Use recent evidence to estimate effort. 3. Protect focused work on the calendar. 4. Assign communication and interruption rules. 5. Capture completion evidence. 6. Review the week and change one design variable.

Read the list once before acting. Mark the checkpoint with the weakest evidence. That is where the plan needs attention; later refinement cannot rescue a decision built on an unresolved early constraint.

Describe the result in observable terms. Include the person, object, or business process affected; the real environment; the acceptable range; and the point at which the work must stop. For project planning, an unacceptable outcome includes a system that adds tracking but not progress, depends on ideal motivation, or hides overloaded commitments.

Separate hard constraints from preferences. A hard constraint can disqualify the method even when it is faster or cheaper. Write assumptions as assumptions, attach an owner, and give high-consequence unknowns a deadline for resolution.

Build the evidence packet around calendar history, completed-work evidence, workload constraints, baseline behavior, and a clearly defined result. Keep it small enough to use during the work. Label each source with its date and scope, and separate a controlling requirement from a preference.

Set up the workspace and communication path before the demanding step. Define three observable outcomes; then confirm that use recent evidence to estimate effort. Make the stop authority explicit. The person who notices a problem should not need to negotiate permission while the exposure or failure is growing.

1. Define three observable outcomes.

Do not treat “define three observable outcomes” as a box to tick. Explain what the step protects and what evidence will prove it worked. Capture cue, action, duration, completion evidence, interruption, result, and one design adjustment, then compare the observation with the stated result. Continue only when the evidence supports the next checkpoint.

Checkpoint: before moving to “use recent evidence to estimate effort,” write one sentence describing what passed, what did not, and who owns the unresolved item.

2. Use recent evidence to estimate effort.

Treat this as the handoff checkpoint: use recent evidence to estimate effort. The person receiving the work should be able to state the result, the remaining risk, and the next review date. If the handoff requires hidden context, the project planning instruction is not finished.

Checkpoint: before moving to “protect focused work on the calendar,” write one sentence describing what passed, what did not, and who owns the unresolved item.

3. Protect focused work on the calendar.

Start by turning “protect focused work on the calendar” into a fact someone can verify. Use calendar history, completed-work evidence, workload constraints, baseline behavior, and a clearly defined result. Write the source and date beside the conclusion; otherwise the team cannot distinguish evidence from memory. For project planning, this checkpoint is complete only when the next operator knows what is confirmed and what remains unknown.

Checkpoint: before moving to “assign communication and interruption rules,” write one sentence describing what passed, what did not, and who owns the unresolved item.

4. Assign communication and interruption rules.

Start by turning “assign communication and interruption rules” into a fact someone can verify. Use calendar history, completed-work evidence, workload constraints, baseline behavior, and a clearly defined result. Write the source and date beside the conclusion; otherwise the team cannot distinguish evidence from memory. For project planning, this checkpoint is complete only when the next operator knows what is confirmed and what remains unknown.

Checkpoint: before moving to “capture completion evidence,” write one sentence describing what passed, what did not, and who owns the unresolved item.

5. Capture completion evidence.

Start by turning “capture completion evidence” into a fact someone can verify. Use calendar history, completed-work evidence, workload constraints, baseline behavior, and a clearly defined result. Write the source and date beside the conclusion; otherwise the team cannot distinguish evidence from memory. For project planning, this checkpoint is complete only when the next operator knows what is confirmed and what remains unknown.

Checkpoint: before moving to “review the week and change one design variable,” write one sentence describing what passed, what did not, and who owns the unresolved item.

6. Review the week and change one design variable.

Use a two-person check when the consequence is meaningful. One person should review the week and change one design variable; the other should compare the action with calendar history, completed-work evidence, workload constraints, baseline behavior, and a clearly defined result. The second check is not bureaucracy—it catches a mismatch while the work is still reversible.

Checkpoint: before moving to “schedule the next inspection or review,” write one sentence describing what passed, what did not, and who owns the unresolved item.

Use one week or one deliberately bounded work cycle and change only one meaningful variable. Define the expected result and the stopping signal before beginning. If a system that adds tracking but not progress, depends on ideal motivation, or hides overloaded commitments appears, end the test and return to the last acceptable condition.

Keep the test honest. Do not add help, favorable conditions, or expert intervention that will be absent during normal use. If the difficult case cannot be tested responsibly, escalate it to the qualified person or authority who can evaluate it.

- Choosing a tool, product, setting, contract form, or template before the project planning requirement is defined. - Testing only the easiest condition and assuming the result represents normal project planning use. - Changing several variables together, which hides the cause of success or failure. - Continuing after a system that adds tracking but not progress, depends on ideal motivation, or hides overloaded commitments because time or money has already been invested. - Finishing the visible task without recording cue, action, duration, completion evidence, interruption, result, and one design adjustment.

When a mistake appears, stabilize first. Protect the person, material, rights, equipment, food, environment, or client experience involved. Return to the first checkpoint contradicted by the evidence, revise one variable, and create a new stop rule before trying again.

This is an implementation framework, not medical or mental-health treatment. Reduce the workload when the system creates sustained distress, and seek qualified support when health or disability affects execution.

Authoritative starting points:

- CDC shared-goals worksheet - NIH behavior-change overview[1] [2]

Confirm that a source applies to the exact model, jurisdiction, land manager, product category, transaction, clinical situation, or activity. Save the access date and pair general guidance with current manufacturer instructions or individualized professional advice when appropriate.

Review cue, action, duration, completion evidence, interruption, result, and one design adjustment. Compare the observation with the result statement, not with the effort invested. Decide to adopt, adjust, obtain qualified help, or stop.

Turn the final note into a short checklist for the next person. Include the six checkpoints, the approved range, a photograph or example where useful, the stop rule, and the escalation contact. A workflow is not delegated until another person can recognize both a good result and a reason to stop.

Write the desired result and the unacceptable outcome. Complete checkpoint one using a current source or direct observation. Then prepare one week or one deliberately bounded work cycle. If the critical evidence is missing, use the time to send one precise question instead of improvising.

The goal of this short session is not to finish project planning. It is to reach the first defensible action with the stop rule already in place.

- 14-day productivity experiment - weekly review template - results resource library

What should be verified first?.

Verify the fact that could disqualify the entire approach. In this workflow that usually means define three observable outcomes, followed by a check that you can use recent evidence to estimate effort under real conditions.

How detailed should the written plan be?.

Detailed enough that another capable person can perform the next checkpoint and recognize a system that adds tracking but not progress, depends on ideal motivation, or hides overloaded commitments. For most situations, one page plus the controlling sources and cue, action, duration, completion evidence, interruption, result, and one design adjustment is more useful than a long narrative.

When is a small test not appropriate?.

Skip informal testing when a recall, emergency, legal restriction, clinical concern, structural question, food-safety uncertainty, unknown hazardous material, or manufacturer prohibition requires an authoritative response first.

What evidence should be saved afterward?.

Save cue, action, duration, completion evidence, interruption, result, and one design adjustment. Add the source date, the person who approved the result, and the date or trigger for the next review.

What if the first attempt fails?.

Stop, protect the affected people and property, and preserve the evidence. Identify the earliest failed checkpoint, change one variable, and decide whether a second bounded test or qualified professional review is the responsible next step.