FMEA Occurrence Ratings: Linking Failure Rates and Cpk to the 1-10 Scale

Assigning an occurrence rating in FMEA should be straightforward: how likely is this cause to happen? In practice, it is one of the most contested ratings in every FMEA session. Teams without historical failure data guess. Teams with data argue about how to interpret it. And nearly everyone struggles with the question: does an occurrence rating of 4 mean the same thing across different products and processes?

This guide walks through how to determine FMEA occurrence ratings using a structured approach—from the ideal case (you have process capability data) to the common case (you are estimating from experience).

What the FMEA Occurrence Rating Measures

Occurrence rates the likelihood that a specific cause will result in the associated failure mode during the product or process lifetime. It is rated on a 1–10 scale, where 1 means the cause is nearly impossible and 10 means it is persistent or inevitable.

Critical distinctions practitioners need to internalize:

Occurrence is rated on the cause, not the failure mode itself. One failure mode can have multiple causes, each with a different occurrence rating.
Occurrence is assessed without regard to detection. Do not lower the occurrence rating because you have a good inspection system. Detection has its own rating column.
Occurrence is reduced through prevention controls—error-proofing (poka-yoke), process capability improvement, and design margin. Not through inspection.

Tip The AIAG-VDA handbook emphasizes that the occurrence rating has a relative meaning, not an absolute value. The failure rate ranges in the table below are guidelines, not thresholds. Your team should calibrate occurrence across your specific product family.

FMEA Occurrence Rating Scale: 1–10 With Failure Rate Guidelines

Rating	Likelihood	Estimated Failure Rate	Cpk Equivalent	Manufacturing Example
1	Extremely unlikely	≤ 1 in 1,000,000	≥ 2.00	Failure eliminated by proven error-proofing (e.g., physical poka-yoke prevents wrong orientation)
2	Remote	1 in 500,000	≥ 1.67	Mature process with established capability; similar designs have near-zero field failures
3	Very low	1 in 100,000	≥ 1.33	Well-controlled process with SPC monitoring; Cpk consistently above 1.33
4	Low	1 in 10,000	≥ 1.17	Capable process with occasional drift; minor tool wear causes infrequent defects
5	Moderate	1 in 2,000	≥ 1.00	Process is capable (Cpk ~1.0) but variation is visible; periodic adjustments needed
6	Moderately high	1 in 500	≥ 0.83	Process shows consistent but manageable variation; SPC charts show patterns near control limits
7	High	1 in 100	≥ 0.67	Process is marginally capable; defects occur regularly and require sorting or rework
8	Very high	1 in 20	≥ 0.51	Process is not capable; frequent defects require 100% inspection or containment
9	Extremely high	1 in 10	≥ 0.33	Failure is almost certain without intervention; known chronic issue
10	Persistent	≥ 1 in 5	< 0.33	Cause is inherent to the current design or process; failure occurs on nearly every unit

How to Determine the Occurrence Rating: Three Methods

Method 1: Process Capability Data (Best)

If you have Cpk data from a capability study on the process step associated with the failure cause, you can map it directly to the occurrence scale. This is the most defensible approach for IATF 16949 audits because it ties the rating to measured data.

Worked Example

Your machining process for a shaft diameter has a Cpk of 1.45 based on the last capability study (30 consecutive parts, stable process). The specification is 25.00 ± 0.05 mm.

Mapping: Cpk 1.45 falls between 1.33 and 1.67, which corresponds to Occurrence = 3 (very low likelihood). The process is well-controlled with SPC monitoring.

If the Cpk drops to 0.95 at the next study, the occurrence rating should be updated to Occurrence = 5 (moderate). This is what it means for FMEA to be a living document—new data changes the ratings.

Common Mistake Using Cpk from the initial capability study (PPAP) as a permanent occurrence rating. Cpk changes over time as tools wear, materials vary, and process conditions drift. Re-evaluate occurrence ratings when capability studies are updated—at minimum during annual FMEA reviews.

Method 2: Historical Failure Data (Good)

When Cpk data is not available for the specific failure cause, use historical records:

Warranty/field failure data: PPM rates from warranty claims on the same or similar product.
Internal scrap/rework data: Defect rates from production records, sorted by failure mode.
8D/CAPA records: Past corrective actions for the same failure type across similar products.
Supplier quality data: Incoming inspection reject rates for purchased components.

Map the historical failure rate to the occurrence scale in the table above. If your scrap data shows a particular defect occurs at roughly 1 in 500 parts, that maps to Occurrence = 6.

Method 3: Engineering Judgment (Acceptable When Documented)

For new products or processes with no historical data, the team must estimate occurrence based on engineering experience. This is inherently subjective, which is why it generates the most debate in FMEA sessions.

Techniques to improve estimation quality:

Analogous process comparison: Identify the most similar existing process and start from its known occurrence rates. Adjust up or down based on complexity, material, and tooling differences.
Prevention control assessment: Evaluate what prevention controls are already designed into the process. Proven poka-yoke devices justify lower occurrence; manual-dependent steps justify higher occurrence.
Blind voting: Have each team member independently assign a rating before group discussion. This prevents anchoring bias where the first number spoken dominates the consensus.

Tip When using engineering judgment, document the rationale in the FMEA. “Occurrence = 4 based on similar process at Plant B with no recorded failures in 24 months of production” is far more defensible in an audit than an unexplained rating.

How Occurrence Interacts With Action Priority

Under the AIAG-VDA Action Priority system, occurrence is the second factor evaluated after severity. For failure modes with severity ratings in the 5–8 range, occurrence becomes the deciding factor between High, Medium, and Low action priority.

The practical implication: reducing occurrence is the primary lever for reducing risk on non-safety-critical failure modes. This is why the AIAG-VDA handbook prioritizes prevention over detection—prevention controls directly reduce occurrence, while detection controls only affect the Detection rating.

Use the RPN & Action Priority calculator to see how changing occurrence affects both RPN and AP for a given severity/detection combination.

Reducing Occurrence: The Prevention Control Hierarchy

Occurrence ratings should decrease over time as the team implements prevention controls. The hierarchy, from most effective to least:

Error-proofing (poka-yoke): Physical or system-level mechanisms that make the failure cause impossible. Example: a fixture that physically prevents a part from being loaded in the wrong orientation. Justifies Occurrence = 1–2.
Process capability improvement: Reducing variation through better tooling, tighter material specs, or improved process parameters. Moves occurrence down as Cpk improves.
Preventive maintenance: Scheduled replacement of wear items (cutters, seals, bearings) before they degrade to a failure-producing state. Reduces occurrence for wear-related causes.
Training and standardized work: For operator-dependent processes. Least reliable prevention control because it depends on human consistency.

Notice that inspection and testing are not on this list. Inspection is a detection control, not a prevention control. Adding a 100% visual inspection does not change the occurrence of the defect—it changes the likelihood of catching it. This is the single most common confusion in FMEA rating sessions.

Common Pitfalls When Rating Occurrence

Conflating occurrence with detection: “We inspect 100%, so occurrence is low”—wrong. Occurrence is about how often the defect is created, not how often it escapes.
Using design life as the time horizon: Occurrence should reflect the likelihood during the manufacturing process or the product usage period, depending on PFMEA vs. DFMEA. Clarify the time frame before rating.
Rating all new processes at 5: Teams without data default to the middle of the scale. If the process is genuinely new with no analogy, a 5 or 6 may be appropriate—but document why, and commit to updating when data becomes available.
Never updating occurrence after launch: Once production data is available, occurrence ratings should be revised to reflect actual failure rates. A process FMEA that still shows estimated occurrence ratings two years into production is not a living document.

Key Takeaways

Occurrence rates the likelihood of a specific cause, not the failure mode itself, on a 1–10 scale.
Use Cpk data when available (most defensible), historical failure rates when Cpk is not available, and documented engineering judgment as a last resort.
Occurrence is reduced by prevention controls (poka-yoke, capability improvement), not by detection controls (inspection, testing).
Under AIAG-VDA Action Priority, occurrence is the deciding factor for non-safety-critical failure modes with severity 5–8.
Update occurrence ratings when new capability data or production failure rates become available.