Observation, not explanation
“The coating detached during humidity testing” is an observation. “Poor surface preparation caused adhesion failure” is a causal hypothesis and should not be embedded in the problem statement.
Special Processes Institute
sp-i.org
Problem Definition Record
Is / Is Not Problem Definition Tool
This app is designed to help a user arrive at a simple, accurate and commonly understood problem statement before beginning root cause analysis. It guides the user through the facts, requirement-versus-actual condition, evidence, problem boundaries, IS / IS NOT comparisons, extent, timing, affected population and known unknowns. The objective is to ensure the problem is understood as completely as the available evidence allows — without prematurely deciding why it happened. The final output should either be a clear, evidence-based problem statement that can be transferred into a separate 8D, RCCA, 5 Why, Fishbone or other analysis method, or a clear instruction that further understanding, evidence or data is required before reliable root cause analysis can begin.
Define before analysing Evidence over assumption Do not ask “why” too early
Not assessed
RCA readiness
0%
Definition completeness
0
Confirmed facts
0
Known unknowns
0
Readiness gaps
Why this tool exists
A reliable root cause analysis begins with a reliable problem definition. If the problem is vague, contains assumptions, combines several issues, confuses detection with occurrence, or omits basic evidence and boundaries, even sophisticated RCCA methods can produce convincing but unreliable conclusions.
ComplaintCapture what was reported
ObserveWhat actually happened?
VerifyWhat evidence proves it?
BoundWhat is / is not affected?
ChallengeRemove assumptions and gaps
StateSimple RCA-ready description
Expected versus actual
Define what should have happened and what objectively happened instead. A problem is fundamentally a demonstrated deviation from an expected condition.
One problem at a time
If cracking, dimensional error and coating failure are all present, do not automatically combine them. Separate them unless evidence demonstrates that they are manifestations of one problem.
Detection is not occurrence
Where a problem was discovered is not automatically where it originated. Record the detection point and the earliest confirmed point at which the condition existed separately.
Boundaries create discrimination
The best comparison is often something nearly identical that does not have the problem. IS / IS NOT analysis helps define those boundaries without prematurely seeking cause.
Stopping is a valid result
If the evidence is insufficient, the correct outcome is not a weak problem statement. The correct outcome is a defined list of missing data and further investigation needed before RCCA begins.
What a strong problem statement should contain
Simplicity is a test of understanding. If the problem cannot be articulated simply and clearly, it may not yet be sufficiently understood. Use the detail gathered in this tool to establish the evidence and boundaries, but keep the final problem statement concise: describe what is wrong, where/when relevant, and the essential extent without turning the statement into the whole investigation record.
| Element | Purpose | Good example content |
|---|---|---|
| Object / subject | What is affected? | Specific part, process, feature, output, service, batch, interface or characteristic. |
| Requirement | What should happen? | Drawing limit, specification criterion, process requirement, approved standard or defined expected condition. |
| Actual deviation | What happened instead? | Measured, observed or otherwise objectively demonstrated departure. |
| Location | Where is it seen? | Product feature, process step, customer, test stage, equipment, cavity, line, shift or interface. |
| Timing | When is it seen? | Date, cycle, shift, lot, first-known occurrence, event or operating condition. |
| Extent | How much? | X of Y, magnitude, frequency, trend, severity or population. |
| IS / IS NOT | What are the boundaries? | Affected versus comparable unaffected populations. |
| Evidence | How do we know? | Measurement, inspection result, test report, photo reference, log, sample, record or repeatable observation. |
Prefer
Example: “8 of 126 brackets manufactured 4–7 August had internal-bore coating thickness of 11–18 μm, below the 20 μm minimum requirement.”
Avoid
Example: “The plating process is bad because the bath chemistry was wrong.”
This combines judgement, an assumed cause and an undefined problem.
This combines judgement, an assumed cause and an undefined problem.
Guided Problem Definition
1. Initial problem0% complete
Problem Statement Challenge
The tool checks for common indicators of vague wording, causal assumptions and insufficient specificity. This is guidance, not linguistic proof.
Enter a statement and select Challenge Statement.
Analysis Readiness
Complete the guided definition and select Assess Readiness.
Problem Definition Output
A handover record for use before 8D / RCCA.
Complete the guided definition and assess readiness to generate an output.