Home / Software FMEA / Likelihood
FMEA Step 6 — Likelihood Tab
Likelihood
Requs AI Predict auto-calculates likelihood from your Survey tab answers. Likelihood for software is not predicted from failure rates — failure rate is the time to the next failure of any mode, not a specific one, and software has no hardware-style “wear-out.” Wear-out and failure are not synonymous.
Inputs
The likelihood columns
The Likelihood tab shows each failure mode row, ten to a page.
Column
Description
Read/write
CDE
Unique identifier for a failure mode that caused at least one world failure.
Read only
Applicability
Depending on survey responses, some CDEs may not apply to your system — see Effects.
Read only
Preventive measures
A design control that prevents the defect from manifesting as a failure. Example: data logging risks filling the disk; a “rollover” that auto-harvests files before the drive fills is a preventive measure.
Read/Write
Existence likelihood
How likely this failure mode is present in the software, given it’s relevant. Based on: unit-test coverage by developers, black-box system test coverage, and prior history of similar failure modes.
Read only
Manifestation likelihood
How likely the failure mode is observed given it’s relevant and exists. Based on whether it’s a single-point or multiple-point failure.
Read only
Control likelihood
How controlled the failure mode is, given it’s relevant and exists. Based on design controls, specification controls, code controls, and UI design controls (if applicable).
Read only
Detectability likelihood
How obvious the failure mode is in a test environment, based on the specific tests conducted prior to deployment.
Read only
Average likelihood
Numerical average of existence, manifestation, control, and detectability likelihood.
Read only
RPN
Risk Priority Number — severity multiplied by average likelihood.
Read only
Overrides
You can override any likelihood assessment.
Read/Write
Justification for overrides
Required if you override an assessment. Base overrides on tangible evidence, not guesswork — likelihood is estimated from controls, test procedures/results, and multiple-point failure likelihood.
Read/Write
Compensating provisions
Actions from a user, or environment changes, that reduce likelihood. Example: alongside an auto-rollover preventive measure, a compensating provision is the user regularly harvesting log files, or fitting a larger drive.
Read/Write
Coming in V2.12
The Likelihood tab will add an option for “Controls Category” and “SwCI” per Mil-Std-882E.
Outputs
Auto-calculated from your survey
Likelihood is auto-calculated from the inputs you provided in the Survey tab. The worksheet can be exported using the Critical to Quality List tab.
Continue the FMEA workflow.
Next, define how each failure mode gets mitigated.