Days 01–07
Define
Choose one person, one problem, one useful promise, and one honest way to know it arrived.
Gate: A stranger can repeat the promise and the intended change without your explanation.Free tool · 30 days · no signup
Stop building motion. Build proof.
Take one useful promise from definition to delivery, real use, and an evidence-shaped decision. Thirty focused days. One private local workspace. No fake traction required.
The operating principle
A month should end with a decision the evidence earned.
Most 30-day challenges reward visible activity: post daily, ship constantly, keep the streak alive. This one rewards a complete promise reaching a real person.
The finish line is not “launched.” It is a defensible choice to continue, change, or stop—supported by delivery, behaviour, outcome evidence, and the limits beside it.
Four phases. Four gates.
No skipped definitionsDays 08–14
Deliver
Build the smallest complete version and put it in front of real people before polishing the theatre.
Gate: At least one real person can receive the complete promise from beginning to end.Days 15–21
Observe
Watch what people actually do, repair the largest point of friction, and ask for return behaviour.
Gate: Use, confusion, return, and silence are recorded separately instead of being turned into one flattering number.Days 22–30
Verify
Follow up, test the intended change, publish the limits, and decide from evidence what deserves another month.
Gate: The final claim is no stronger than the evidence attached to it.Run the sprint.
Private draft · saved on this device00
Name the test.
One project, one promise, one person. The plan gets weaker when the scope gets wider.
Define · Today
Write the one-sentence promise
Name who this is for, what they receive, and what they should be able to do next.
Evidence: Save the sentence before editing it into brand language.Define
Choose one real person
Describe one specific person with the problem—not a market segment or an audience avatar.
Evidence: Record the situation, constraint, and exact moment the need appears.Define
Name the intended change
Finish: after using this, the person can, believes, chooses, or practises…
Evidence: Save one observable or reportable change.Define
Cut the second promise
Remove every bonus outcome that is not required for the first useful result.
Evidence: Keep a before-and-after version of the promise.Define
Choose the proof ladder
Define what delivery, use, return, and evidenced change mean for this exact work.
Evidence: Write one plain-language signal for each level.Define
Ask five unprompted questions
Show the promise to five relevant people and ask what they think they would receive.
Evidence: Capture their words without correcting them first.Define
Pass the definition gate
Rewrite once using what people misunderstood. Freeze the promise for the next seven days.
Evidence: Date the frozen promise and list its known limits.Deliver
Map the shortest complete path
List only the steps required to move one person from arrival to the promised result.
Evidence: Keep a path with a clear start, finish, and owner.Deliver
Build the rough first half
Make the first half usable without waiting for final copy, animation, or automation.
Evidence: Save a screenshot, file, link, or working artifact.Deliver
Finish the whole promise
Complete the second half. A plain complete experience beats a cinematic fragment.
Evidence: Run the path yourself from a clean starting state.Deliver
Remove one point of friction
Test on a phone and remove the slowest, least clear, or most demanding step.
Evidence: Record what was removed and why.Deliver
Add the smallest feedback door
Give people one obvious way to say what happened, what failed, or what they need next.
Evidence: Verify the feedback path reaches a place you will actually check.Deliver
Deliver it to three people
Invite three relevant people individually. Do not call impressions delivery.
Evidence: Record invitations, completed deliveries, and failures separately.Deliver
Pass the delivery gate
Confirm one person can receive the full promise without you rescuing the experience.
Evidence: Save the path, date, result, and remaining friction.Observe
Watch without explaining
Observe one person using the work. Let confusion become visible before offering help.
Evidence: Note every pause, wrong turn, and question.Observe
Fix the largest break
Repair the one issue most likely to prevent a person from reaching the promised finish.
Evidence: Retest the repaired step from a clean state.Observe
Ask what happened next
Contact the first users and ask what they did after the experience—not whether they liked it.
Evidence: Keep answers in their own words.Observe
Separate attention from use
Write reach, delivery, and meaningful use on different lines.
Evidence: Leave unknown numbers unknown; do not estimate them upward.Observe
Create a reason to return
Add one useful next step that makes a second visit valuable without manufacturing urgency.
Evidence: State what return behaviour would look like.Observe
Invite the second use
Ask prior users to repeat, return, reply, or refer only when the next action serves them.
Evidence: Record the invitation and actual return separately.Observe
Pass the behaviour gate
Review what people did, where they stopped, and whether anyone returned.
Evidence: Publish or save the honest funnel, including every zero.Verify
Choose the follow-up window
Set a window long enough for the intended change to be observable and short enough to complete this sprint.
Evidence: Record the date, cohort, and why the window fits.Verify
Ask an outcome question
Ask what changed, what did not, and what else could explain it. Avoid leading language.
Evidence: Save the exact question and unedited answers.Verify
Test the weak explanation
Look for novelty, politeness, selection bias, outside help, and other reasons for the result.
Evidence: List at least two credible alternative explanations.Verify
Record the limitations
Write what the evidence cannot establish before writing the success sentence.
Evidence: Keep the limitation beside the number, not in a hidden note.Verify
Build the honest scorecard
Enter reach, delivery, use, return, and evidenced change as separate values.
Evidence: Export or print the scorecard with its method attached.Verify
Write the public sentence
Claim only the highest proof level supported. Use zero where zero is the evidence.
Evidence: Include the cohort, window, method, and limitation.Verify
Let someone challenge it
Give the sentence and evidence to one skeptical reader and ask what feels unsupported.
Evidence: Record the challenge and any resulting correction.Verify
Choose continue, change, or stop
Decide from evidence whether to repeat the promise, revise it, or end the experiment.
Evidence: Write the decision and the evidence that earned it.Verify
Publish the record
Share what was attempted, delivered, used, changed, and still unknown. Keep people private.
Evidence: Archive the proof pack and name the next smallest test.The 30-day proof sprint, plainly answered.
What is a 30-day proof sprint?
A 30-day proof sprint is a focused cycle for defining one promise, delivering the smallest complete version, observing real behaviour, and testing whether the intended change has evidence. It is not a 30-day content calendar or a promise of business success.
Do I need an audience before starting?
No. Start with one reachable person who actually has the problem. The sprint treats a completed delivery to one real person as stronger evidence than broad attention with no use.
What counts as completing a day?
Complete a day when the move is done and its evidence exists. Checking a box without the artifact, observation, decision, or dated note is activity, not proof.
Does the planner send my project data anywhere?
No. The working draft is stored only in this browser. It can be printed or exported as JSON, and no account is required.
Can I restore a sprint export on another browser?
Yes. Choose Restore a sprint file, select a version-one SIX33 JSON sprint export, and review its project, completed days, and note count. Export the current draft if you want to keep it, then choose Replace current sprint. Invalid files are rejected before they can replace a draft.
What if the result is zero?
Record zero. A sprint that proves the promise was unclear, unused, or unsupported has produced useful evidence. The next decision should change; the number should not be inflated.