Six Sigma: Failure Mode and Effects Analysis

In a world where a single product defect can cost millions in recalls or a process error can jeopardize patient safety, waiting for failures to happen is not an option. Failure Mode and Effects Analysis (FMEA) is a proactive, systematic methodology used to identify and eliminate potential failures before they reach the customer. By rigorously assessing what could go wrong, how badly, and how likely it is to be caught, FMEA transforms risk management from reactive firefighting into a disciplined strategic process, which is a cornerstone of Six Sigma and project management excellence.

Understanding the FMEA Framework

At its core, FMEA is a structured, team-based exercise that dissects a process, product, or service to uncover every conceivable way it might fail. A failure mode is defined as the specific manner in which a component, subsystem, or process could potentially not meet its intended function or design requirement. For instance, a failure mode for a car door latch could be "fails to secure closed," and for an invoicing process, it could be "incorrect customer address printed."

The power of FMEA lies in its systematic nature. It moves teams beyond informal brainstorming to a documented, repeatable analysis. Practitioners typically use a standardized spreadsheet or software template to catalog each potential failure mode, its potential effects (the consequences to the customer or next process step), and the hypothesized root causes. This creates a living document that serves as both an engineering and a communication tool, aligning cross-functional teams—from design and manufacturing to quality and service—on potential risks and mitigation plans.

Process FMEA vs. Design FMEA

A critical distinction for effective application is understanding the two primary types of FMEA, which differ in their focus and stage of application. Design FMEA (DFMEA) is conducted during the product development phase. It analyzes potential failures related to the product's design—its materials, geometry, tolerances, and interactions between components. The goal of DFMEA is to prevent design-related failures, ensuring the product is robust and reliable before any physical prototype is built. An example is analyzing a new smartphone battery design for potential failure modes like overheating or rapid capacity loss.

Conversely, Process FMEA (PFMEA) is performed on the manufacturing or assembly process itself. It focuses on how the production process could cause the product to fail, even if the design is sound. PFMEA examines elements like human operators, machinery, methods, measurement systems, and the environment. For example, in assembling that smartphone, a PFMEA might identify a failure mode where "torque on chassis screw is inconsistent," leading to a potential effect of "housing rattle." While DFMEA asks "What could be wrong with the thing we designed?", PFMEA asks "What could go wrong in the way we make or deliver it?"

The RPN: Quantifying Risk with Severity, Occurrence, and Detection

The heart of the FMEA prioritization system is the Risk Priority Number (RPN), a numerical score calculated by multiplying three independent ratings: Severity (S), Occurrence (O), and Detection (D). Each is rated on a standard scale, typically from 1 (Low/Best) to 10 (High/Worst).

• Severity (S) assesses the seriousness of the effect of the failure on the customer. A minor nuisance might rate a 2 or 3, while a failure causing critical injury or system loss would be a 9 or 10. The key question is: "How severe is the consequence to the customer if this failure happens?"
• Occurrence (O) estimates the likelihood or frequency of the cause of the failure happening. This rating is based on historical data, predictive testing, or expert consensus. A failure that happens once in 15 years might be a 2, while one that happens several times per day is a 10. The question is: "How often is the root cause likely to occur?"
• Detection (D) evaluates the effectiveness of current process controls at finding the failure or its cause before it reaches the customer. A failure that is certain to be caught by an automated 100% test scores a 1, while one that is virtually undetectable with current methods scores a 10. The question is: "What is the chance our current controls will detect this failure before it impacts the customer?"

The RPN is then calculated as: $$RPN=S\times O\times D$$

Scores range from 1 to 1000. While there is no universal threshold, organizations typically establish an action priority level (e.g., RPN > 100, or any Severity rating of 9 or 10). This formula forces teams to consider all three dimensions of risk. A high-severity failure with a low occurrence rate might still demand action, especially if it's hard to detect.

Developing and Implementing Recommended Actions

The primary purpose of calculating the RPN is to objectively prioritize which failure modes require preventive action. The FMEA process is not complete until the team develops and implements recommended actions aimed at reducing the highest risks. The strategy for action development targets the three rating factors:

1. Reduce Severity: This is often the most difficult, as it may require a fundamental design change. For example, adding a redundant safety system can reduce the severity of a primary system failure.
2. Reduce Occurrence: This attacks the root cause to make the failure less likely. Actions include design for robustness, mistake-proofing (poka-yoke), improving maintenance schedules, or enhancing operator training.
3. Improve Detection: This involves adding or improving controls to catch the failure earlier. This could mean implementing more frequent inspections, adding sensors, or introducing automated testing.

After actions are implemented, the team must re-rate the S, O, and D scores and calculate a post-action RPN. This quantifies the effectiveness of the intervention and demonstrates risk reduction. The FMEA becomes a closed-loop process: identify, prioritize, act, and verify.

FMEA as a Proactive Prevention Tool

Ultimately, FMEA's greatest value is its shift from a reactive to a proactive quality paradigm. It prevents quality problems before they occur by forcing foresight. In manufacturing, this means preventing scrap, rework, and warranty claims. In service processes—such as loan approval, patient admission, or software deployment—it prevents delays, errors, and customer dissatisfaction.

By conducting FMEA early in the design or process planning phase, organizations can avoid the exponentially higher costs of fixing problems after launch. It is a classic example of the "ounce of prevention" principle, formally embedded into the product and process lifecycle. This proactive stance is a hallmark of a mature, risk-aware organization and is heavily emphasized in professional certifications like PMP and Six Sigma for its role in project risk management.

Common Pitfalls

Even well-intentioned teams can undermine an FMEA's effectiveness. Avoiding these common mistakes is crucial.

1. Treating the RPN as the Sole Decision Metric: Focusing only on the highest RPN can be misleading. A failure with a Severity of 10 (safety-critical) and an Occurrence of 2 (RPN=20 if Detection is 1) is far more important to address than a failure with ratings of S=5, O=5, D=5 (RPN=125). Best practice is to prioritize any high-Severity item first, regardless of the total RPN.
2. Underestimating Occurrence Due to Lack of Data: Teams often rate Occurrence optimistically without historical data. This inflates the Detection rating as the primary "solution," leading to expensive inspection overhauls instead of root cause elimination. Always seek data to ground the Occurrence rating.
3. Confusing Causes with Failure Modes: A statement like "operator error" is a cause, not a failure mode. The failure mode might be "incorrect part installed." Starting with the clear failure mode is essential for a clear analysis. Similarly, the "effect" must be the consequence to the external customer or next process step, not an internal department.
4. Failing to Update the FMEA: An FMEA is not a one-time report to be filed away. It is a living document that should be revisited after failures occur in the field, when processes are changed, or when new technology becomes available. An outdated FMEA provides a false sense of security.

Summary

• FMEA is a proactive, systematic team activity designed to identify and mitigate potential failures in a design or process before they occur.
• DFMEA focuses on product design, while PFMEA focuses on manufacturing and process steps; applying the correct type is essential.
• Risk is prioritized using the Risk Priority Number (RPN), calculated by multiplying ratings for Severity (S), Occurrence (O), and Detection (D).
• Recommended actions should aim to reduce the highest risks by lowering Severity, reducing Occurrence, or improving Detection, with a post-action RPN calculated to verify improvement.
• The methodology's core value is preventing problems, reducing costs, and enhancing safety and customer satisfaction by building knowledge and controls into the system upfront.