Problem → Hypothesis → Measurement Template

An interactive one-page template that links problem statements to a clear hypothesis, leading and lagging signals, success criteria, required data, and a recommended method. Saveable responses let teams capture, iterate, and re-use experiment-ready plans.

{ "Title": "Problem → Hypothesis → Measurement Template", "IntroductionHtml": "

This interactive template helps teams move from a loosely stated problem to a testable hypothesis and a measurable plan. Use it to clarify who will act on the result, what success looks like, what data you need, and the method you'll use. Save responses so plans can be revisited, copied, and connected to experiment tracking.

Filled example (Customer churn)

Problem: Monthly active users who contact support are churning at a higher rate than average.
Decision owner: Head of Customer Success
Desired outcome: Reduce 90-day churn for contacted users from 12% to 8% within 3 months.
Candidate hypothesis: If we proactively offer a dedicated onboarding session to contacted users, then their 90-day churn rate will decrease.
Leading signals: Signup for onboarding session, reduction in support repeat contacts within 30 days.
Lagging signals: 90-day churn rate, net revenue retention.
Success criteria: 90-day churn drops to <= 8% and effect is statistically significant (p < 0.05) with minimum detectable effect of 3 percentage points.
Required datasets: user_id, contact_date, support_contact_reason, onboarding_session_attendance, churn_flag (90-day), revenue; daily export, owned by CS analytics.
Suggested method: Randomized experiment (A/B test).
Estimated effort: ~120 person-hours.
Risks: Selection bias if attendance is voluntary; mitigate with incentivized attendance and intention-to-treat analysis.

", "SubmitLabel": "Save Template", "SuccessMessage": "Template saved. You can return to edit, copy to create a new plan, or link this plan to an experiment tracker.", "DataType": "problem-hypothesis-measurement-template", "SchemaVersion": "1.0", "Fields": [ { "Key": "problem_statement", "FieldType": "textarea", "Label": "Problem statement", "HelpText": "Describe the operational or business problem you're trying to address. One concise sentence is ideal.", "Required": true }, { "Key": "decision_owner", "FieldType": "text", "Label": "Decision owner", "HelpText": "Who will use this result to make a decision? Provide a role or person.", "Required": true }, { "Key": "desired_outcome", "FieldType": "textarea", "Label": "Desired outcome", "HelpText": "What measurable change will indicate the decision owner achieved their goal? Include time bounds where appropriate.", "Required": true }, { "Key": "candidate_hypothesis", "FieldType": "textarea", "Label": "Candidate hypothesis (If... then...)", "HelpText": "State the hypothesis clearly in an if/then format so it maps directly to an intervention and an expected outcome.", "Required": true }, { "Key": "leading_signals", "FieldType": "textarea", "Label": "Leading signals", "HelpText": "Early indicators that the intervention is moving in the desired direction. Useful for quick checks and iteration.", "Required": false }, { "Key": "lagging_signals", "FieldType": "textarea", "Label": "Lagging signals", "HelpText": "Outcome measures that confirm success but may take longer to observe.", "Required": false }, { "Key": "success_criteria", "FieldType": "textarea", "Label": "Success criteria", "HelpText": "Pre-specified numeric and/or statistical criteria for declaring success. Be concrete (metrics, thresholds, significance where appropriate).", "Required": true }, { "Key": "required_datasets", "FieldType": "textarea", "Label": "Required datasets", "HelpText": "List datasets, specific fields, refresh cadence, owners, and any access notes or limitations.", "Required": true }, { "Key": "suggested_method", "FieldType": "select", "Label": "Suggested method", "HelpText": "Choose the primary method you plan to use. This guides analysis approach and pre-analysis planning.", "Required": true, "Options": [ { "Value": "eda", "Label": "Exploratory Data Analysis (EDA)" }, { "Value": "experiment", "Label": "Randomized experiment / A/B test" }, { "Value": "quasi-experiment", "Label": "Quasi-experimental (diff-in-diff, regression discontinuity)" }, { "Value": "predictive", "Label": "Predictive model / forecasting" }, { "Value": "qualitative", "Label": "Qualitative or mixed methods" } ] }, { "Key": "estimated_effort_hours", "FieldType": "number", "Label": "Estimated effort (hours)", "HelpText": "Approximate person-hours to design, run, analyze, and operationalize.", "Required": false }, { "Key": "risks", "FieldType": "textarea", "Label": "Risks and mitigation", "HelpText": "Known risks, biases, ethical or safety concerns, and planned mitigations.", "Required": false }, { "Key": "notes_next_steps", "FieldType": "textarea", "Label": "Notes / next steps", "HelpText": "Any additional context, stakeholders to notify, or immediate next actions.", "Required": false } ] }

Discussion

Comments and conversation will live here.