Safety Performance Monitoring

You cannot manage what you cannot measure. This principle, fundamental to every management discipline, applies with particular force to aviation safety. 14 CFR Part 5 Subpart D requires operators to continuously monitor safety performance, assess whether controls are effective, and drive improvement based on evidence. Safety Performance Monitoring is the mechanism that turns raw safety data into actionable intelligence.

​

Why Measuring Safety Is Hard

Safety is fundamentally about the absence of adverse events. Measuring the absence of something is inherently difficult — you cannot count the accidents that did not happen. This creates a paradox: when your SMS is working well, nothing dramatic happens, and it can be tempting to conclude that safety management is unnecessary overhead. This is precisely why structured measurement matters. Without it, organizations fall into one of two traps: either they assume everything is fine because no accidents have occurred (ignoring the precursors accumulating beneath the surface), or they overreact to isolated events without understanding whether they represent a trend or an anomaly. Safety Performance Monitoring replaces intuition with evidence. It establishes what “normal” looks like for your operation, detects deviations from that baseline, and provides the data needed for informed decision-making about where to focus safety resources.

​

Leading vs. Lagging Indicators

Safety Performance Indicators (SPIs) fall into two categories, each serving a different purpose.

​

Lagging Indicators

Lagging indicators measure events that have already occurred. They tell you what happened. Lagging indicators are important for trend analysis and benchmarking, but they have a critical limitation: by the time a lagging indicator changes, the damage has already occurred. An increase in your accident rate means accidents happened. Lagging indicators are rearview mirrors — essential for understanding where you have been, but insufficient for seeing where you are going.

​

Leading Indicators

Leading indicators measure conditions that predict future events. They tell you what might happen. Leading indicators are predictive. A declining safety report rate may signal deteriorating trust in the reporting system — a precursor to hazards going unidentified. A rising number of overdue CPAs suggests that identified risks are not being addressed. A drop in training compliance means personnel may lack the knowledge to perform their SMS roles effectively.

​

The Monitoring-Assessment-Improvement Cycle

Part 5 Subpart D establishes a three-phase cycle for Safety Assurance:

​

Monitor (Section 5.71)

Continuous monitoring covers seven areas defined by the regulation: operational processes (are people following procedures?), organizational characteristics (are resources and staffing adequate?), external events (industry incidents, regulatory changes, manufacturer bulletins), internal safety reporting (volume, types, trends, and quality), regulatory compliance (are all applicable requirements being met?), environmental factors (physical and organizational environment), and employee safety concerns (are people raising issues through formal and informal channels?). Monitoring is not passive observation. It requires defined metrics (SPIs), data collection processes, and regular review cadences. Without structured monitoring, safety data accumulates without anyone acting on it.

​

Assess (Section 5.73)

Assessment evaluates whether controls are achieving their intended outcomes. Key questions include:

• Are SPI values within their target ranges?
• Are risk controls reducing risk as expected?
• Are corrective actions producing lasting improvements?
• Are there emerging trends that existing controls do not address?
• Are external changes creating new risks?

When assessment reveals that a control is ineffective — an SPI exceeds its alert threshold, or a corrective action did not resolve the underlying issue — the finding triggers corrective action.

​

Improve (Section 5.75)

Improvement closes the loop. Findings from monitoring and assessment drive specific actions: revising procedures, implementing new controls, retraining personnel, or reassessing risks. These improvements are tracked as CPAs with owners, due dates, and verification steps. Continuous improvement is not optional refinement. It is a regulatory requirement. An SMS that does not change over time is not meeting Part 5’s requirements, because operations, personnel, and external conditions are always changing.

​

How PlaneConnection Automates Monitoring

​

Safety Performance Indicators (SPIs)

PlaneConnection’s SPI module lets you define indicators with target values (the performance level you are aiming for), alert thresholds (the level at which a value requires investigation), trend visualization (charts showing SPI values over time, making it easy to spot patterns), and automated data feeds (many SPIs can be calculated automatically from data already in the system, such as report counts, CPA completion rates, and training compliance). You choose which indicators are relevant to your operation. A Part 135 operator flying single-engine piston aircraft into small airports has different risk priorities than one flying large-cabin jets into major hubs. Your SPIs should reflect your specific operation.

​

SmartScore

SmartScore is PlaneConnection’s AI-powered safety scoring system. It aggregates data from multiple sources — safety report trends, CPA timeliness, training compliance, risk register health, investigation thoroughness, and reporting culture indicators — into an overall safety health assessment. SmartScore serves two purposes. As an executive summary, it allows the accountable executive to see at a glance whether the SMS is healthy without reviewing every SPI individually. For trend detection, SmartScore identifies patterns that may not be visible in individual SPIs. A combination of slightly declining reporting rates, slightly increasing CPA overdue counts, and slightly decreasing training compliance might each be within acceptable ranges individually but together signal a concerning trajectory. SmartScore also operates at the pilot level, providing individual safety profiles that aggregate training currency, flight experience, reporting participation, and other relevant factors.

​

AI Insights

PlaneConnection’s AI analysis layer continuously examines your safety data for patterns that human review might miss. Anomaly detection identifies unusual patterns in reporting data, such as unexpected spikes in a particular report category or geographic clustering of events. Similar incident clustering groups related reports that may not have been explicitly linked, revealing recurring issues. Proactive alerts surface trends before they cross alert thresholds, giving you time to investigate and respond. Pattern recognition identifies correlations between operational factors (routes, aircraft, crew combinations, time periods) and safety events that would be difficult to detect manually.

​

Connecting Monitoring to Action

Monitoring is only valuable if it drives action. PlaneConnection connects the monitoring-assessment-improvement cycle through integrated workflows: The goal is to ensure that no monitoring finding disappears into a void. Every deviation from expected performance has a clear path from detection to investigation to resolution to verification.

​

What Good Monitoring Looks Like

A healthy safety performance monitoring practice has several characteristics: Regular cadence. SPIs are reviewed at defined intervals — weekly for high-priority indicators, monthly for trend analysis, quarterly for strategic review. Safety committee meetings should include SPI review as a standing agenda item. Context, not just numbers. An SPI value in isolation is not informative. A report rate of 3.2 per 1,000 flight hours means nothing without knowing what the target is, what the historical trend looks like, and what is driving any changes. Balanced portfolio. Monitoring only lagging indicators creates a false sense of security. Monitoring only leading indicators creates noise without confirming whether safety outcomes are improving. Both types are needed. Actionable thresholds. Alert thresholds should be set at levels where investigation and action are warranted. Thresholds set too sensitively create alert fatigue. Thresholds set too loosely miss genuine signals. Documented responses. When monitoring triggers an investigation or corrective action, the response should be documented. This creates the evidence trail that demonstrates your Safety Assurance processes are functioning — evidence the FAA will look for during surveillance.

​

Related