hollandkevint The 9 Steps
Walk through each step in order. Do not skip to solutions. The ur-problem is to define the problem.
Step 1: Define the Problem
Ask: "Describe the problem in one sentence. Who is affected, and what's the observable impact?"
Push back if the user names a solution disguised as a problem. "We need a dashboard" is a solution. "Regional managers can't compare quarterly performance without emailing the analytics team" is a problem.
Step 2: Is/Is Not Analysis
Narrow the scope. What IS the problem? What is it NOT?
| Dimension | IS | IS NOT |
|---|---|---|
| What | [specific symptoms] | [things that look similar but aren't this] |
| Where | [where it happens] | [where it doesn't happen] |
| When | [when it occurs] | [when it doesn't occur] |
| Who | [who's affected] | [who's not affected] |
The distinctions reveal the boundary conditions. Problems that happen everywhere are different from problems that happen in specific contexts.
Step 3: Diagnose Root Causes
Use Five Whys. Ask "Why?" five times, drilling past symptoms to root causes.
Rules:
- Each "Why?" must point to a cause, not a person
- Stop when you reach something the user can change
- Branch when there are multiple causes at any level
- If "Why?" produces a shrug, that's an open question to investigate
Step 4: Force Field Analysis
What forces push toward solving this problem? What forces resist?
| Driving Forces (push toward change) | Restraining Forces (resist change) |
|---|---|
| [list specific forces with strength 1-3] | [list specific forces with strength 1-3] |
The strategy isn't always to add more driving force. Sometimes reducing a restraining force is cheaper and more effective.
Step 5: Generate Solutions
Now — and only now — generate solutions. At least 3, max 7. For each:
- One sentence describing the solution
- Which root cause it addresses
- Estimated effort (hours/days/weeks)
- Key risk
Include at least one "do nothing" option with explicit consequences.
Step 6: Evaluate Against Criteria
Define 3-5 evaluation criteria BEFORE scoring. Common ones:
- Impact on root cause (not symptoms)
- Feasibility with available resources
- Time to first result
- Reversibility (can you undo it if wrong?)
Score each solution 1-5 on each criterion. Present as a decision matrix.
Step 7: Recommend
Pick the highest-scoring solution. Explain why in 2-3 sentences. Name what would change your recommendation.
Step 8: Define Success Metrics
How will you know the problem is actually solved (not just addressed)?
- Leading indicator: what changes first?
- Lagging indicator: what confirms the problem is gone?
- Timeline: when should you see each indicator?
Step 9: Identify What Could Go Wrong
Pre-mortem: "This solution failed. What happened?"
Name 2-3 failure modes and a mitigation for each. If a failure mode has no mitigation, flag it as an accepted risk.
Output
# Problem Diagnosis: [Problem Name]
Date: [date]
## Problem Statement
[One sentence from Step 1]
## Is/Is Not
[Table from Step 2]
## Root Causes
[Five Whys chain from Step 3]
## Force Field
[Table from Step 4]
## Solutions Considered
[Numbered list with effort/risk from Step 5]
## Decision Matrix
[Scoring table from Step 6]
## Recommendation
[From Step 7]
## Success Metrics
[From Step 8]
## Risks & Mitigations
[From Step 9]
Adapted from BMAD Creative Intelligence Suite problem-solving workflow (Dr. Quinn).
Next step
Diagnosis done → run assumption-mapper to extract what you're betting on, or grill-me to pressure-test the chosen solution.
Part of the ThinkHaven Method Kit by Kevin Holland. Full Board of Directors experience: ThinkHaven