A Cause and Effect Matrix helps DMAIC teams move from a broad process map to a prioritized list of inputs, variables, and potential causes that deserve measurement and analysis.
Definition
A Cause and Effect Matrix is a prioritization tool that connects process inputs, process steps, or potential causes to customer requirements and critical outputs. It is commonly used after a SIPOC or process map and before deeper analysis such as FMEA, data collection, regression, or DOE.
The matrix usually lists customer requirements or CTQs across the top, weights those requirements by importance, lists process inputs down the side, and scores the strength of relationship between each input and each output. The weighted scores help the team identify which inputs deserve attention first.
History
The Cause and Effect Matrix grew from quality function deployment, process mapping, and Six Sigma prioritization practice. Six Sigma teams needed a practical bridge between broad process understanding and focused data collection. The matrix provided a structured way to avoid chasing every possible variable with equal effort.
It remains common in DMAIC because Define and Measure phases often reveal many possible causes. The matrix helps teams use customer importance and process knowledge to narrow the field before investing in detailed measurement and statistical analysis.
When to Use
Use a Cause and Effect Matrix when the team has a process map, several CTQs or output requirements, and many potential inputs. It is useful during Measure and early Analyze work when deciding what to measure, which variables to investigate, which failure modes to review, or which process inputs to include in an FMEA.
Do not use it as proof of root cause. The scores are based on team knowledge and assumptions. High-scoring inputs are priorities for investigation, not verified causes.
Step-by-Step
- Define customer outputs. List CTQs, output measures, customer requirements, safety requirements, or business outcomes.
- Weight output importance. Assign importance weights using VOC, customer impact, risk, cost, or strategic priority.
- List process inputs. Use a process map, SIPOC, Gemba observations, Fishbone diagram, or team knowledge to list inputs and potential causes.
- Score relationships. Rate how strongly each input affects each output using a consistent scale such as 0, 1, 3, and 9.
- Calculate weighted totals. Multiply relationship scores by output weights and sum across each input.
- Rank the inputs. Sort by weighted score and identify the vital few variables for data collection or analysis.
- Challenge the assumptions. Review whether scores reflect evidence, process knowledge, or opinion.
- Translate into next steps. Use the prioritized list to build a data collection plan, FMEA, control plan review, or experimental plan.
Examples
- Manufacturing defect: A team links inputs such as material lot, oven temperature, conveyor speed, fixture condition, and operator setup to CTQs for adhesion, appearance, and dimension.
- Call center accuracy: A service team links script clarity, system prompts, agent experience, queue pressure, and training quality to customer accuracy and call resolution.
- Healthcare discharge: A team links physician signoff, pharmacy readiness, patient education, transport timing, and documentation completeness to discharge time and patient satisfaction.
- Supplier process: Supplier inputs are scored against incoming quality, delivery reliability, and packaging damage to prioritize supplier development work.
Common Pitfalls
- Treating scores as facts. The matrix prioritizes hypotheses. Data still has to verify cause-and-effect relationships.
- Weak CTQ definition. If outputs are vague, scoring becomes subjective and unfocused.
- Too many inputs. A massive matrix becomes hard to score consistently. Group or filter inputs first.
- No customer weighting. Equal output weights can hide what matters most to customers or risk.
- Dominant voices biasing scores. Use facilitation and evidence to prevent hierarchy from driving the ranking.
- No follow-through. The matrix should lead to measurement, FMEA, controls, or experiments.