Two operators under the same authority, same aircraft type, comparable flight volume: one receives several interventions during the year, the other just one, and neither has had a notable event. The difference lies neither in size nor in chance. It lies in a file the operator never sees in full — the risk profile the authority keeps current on it — and in a handful of internal behaviours that register in that file very quickly. Knowing how this plan is built means understanding why certain actions cost heavily in inspector-days, and why others, meant to reassure, earn none at all.
“Risk-based” describes a constraint, not a philosophy
An authority has a finite number of inspector-days per year, and that number tracks neither traffic growth nor the growing number of approved organisations. Risk-based oversight is first and foremost the arithmetic answer to that constraint: since not everything can be looked at, priority goes to where a defect would have the greatest consequences and where the organisation has shown the least ability to avoid it. The vocabulary, the scoring grids and the cycles are only the dressing on top of this allocation.
The most widely misunderstood consequence concerns the planning unit. The plan is not written operator by operator, but domain by domain within each operator. The same certificate can, in the same year, carry a domain judged well-controlled and set aside, and a domain judged fragile that absorbs most of the allocated days. The operator who answers “we were audited last year” is reasoning at the scale of the certificate; the authority reasons at the scale of the programme line.
Two other notions are also wrongly conflated. Frequency states at what interval a domain will be reviewed. Depth states how many samples will be taken there, over what period, and whether documentary evidence will suffice or real activity will be observed. A deteriorating profile does not mechanically add visits: more often, it shifts days from one domain to another and moves that domain from a desk review to an on-site observation.
The authority has limited time to inspect, so it focuses on the riskiest areas. It looks at how well an operator handles its risks, not just how many flights it makes. If an operator has problems, the authority may inspect more deeply in that area, even if it means inspecting less in other areas.
An operator with an Airbus A320 fleet flying from Paris to London may receive fewer inspections than a similar operator flying to remote airports with challenging weather. If the second operator has recurring issues with flight planning, the authority may spend more time checking that area, even if it means reducing inspections in other areas like crew training.
Two ingredients: what the operation is worth by nature, what the organisation has demonstrated
The profile is built on two terms that must absolutely be kept separate, because they do not respond to the same lever.
The first is inherent exposure: what the operation implies in itself, regardless of the quality of whoever runs it. Number of fleet types and cockpit heterogeneity, network profile, demanding aerodromes, share of night and low-visibility operations, extended diversion over water, dangerous goods carriage, outsourced continuing airworthiness management, wet-leased aircraft, marked seasonality, growth rate. This share is not negotiable. A regional operator opening a distant base sees its exposure rise the day it signs, before it has even flown.
The second is demonstrated performance, that is, what the organisation has proven it can do with that exposure. It is read in the history of findings and, above all, in how they were closed, in the yield of internal audits, in the flow of occurrence reports, in the stability of the people holding the nominated post-holder functions, in how early changes are notified. It is the only term an internal action can genuinely move.
Hence a simple, and often liberating, geometry: exposure places the operator on the horizontal axis, and it can barely move along it; performance places it on the vertical axis, and it is there, and only there, that its efforts become visible. Trying to look less complex than you are moves nothing. Demonstrating that you master your complexity moves everything.
The data that feed the profile, and the data thought to be neutral
The profile is not fed by an annual declaration. It updates continuously, from sources several of which escape the operator’s direct view. Open findings and their closure dates form its backbone, but it also draws on the flow of occurrence reports — volume, filing delay, proportion of voluntary reports —, the results of ramp inspections undergone abroad, changes notified and those that were not, requests to extend a corrective action deadline, and a set of organisational signals: change of control, rapid growth, turnover among accountable managers, outsourcing of a function previously held in-house.
Several of these elements are read backwards inside organisations. The table below sets the two readings side by side: the one that reassures in a meeting, and the one used for planning.
| What the file shows | Common reading on the operator’s side | What the authority draws from it |
|---|---|---|
| No internal finding for the year | The system is healthy | Audit function with no yield: under-detection hypothesis, to be tested by sampling |
| Very low flow of occurrence reports | Few events occur | Likely under-reporting, or a reporting culture to be checked on the ground |
| Findings closed on time | Clean file | Neutral as long as no dated effectiveness check accompanies the closure |
| Repeated requests to extend deadlines | Administrative formality | Insufficient resources or prioritisation: a signal of organisational strain |
| Same minor discrepancy noted on the ramp across several aircraft | Isolated inspection incidents | Systemic defect in procedure, documentation or training |
| Change notified after it was implemented | The situation has been regularised | Failure of the change-management process, which is a systemic finding |
What these misreadings have in common fits in one sentence: the authority does not measure the absence of problems, it measures the capacity to detect them. An indicator that never moves is not reassuring, it is suspect, because no real system produces a perfectly flat series. An operator whose reporting rate sits markedly below that of comparable operators does not get a pat on the back: it gets a visit whose theme will be the reporting culture.
From profile to calendar: frequency, depth, allocation
The move from profile to programme happens in three steps. First, a target interval is set for each domain, that is, the period at the end of which it must have been reviewed; on the European side, this reference planning cycle is typically counted in years, on the order of two, with the possibility of extension when safety performance is actually demonstrated, and reduction as soon as it deteriorates; the exact duration, its limits and the conditions for extension are set out in the applicable regulation and in the authority’s procedures, which alone are authoritative. Depth is decided next: desk review, on-site audit, or observation of real activity — a simulator session, a line flight, a fuelling operation with passengers on board, a maintenance base visit. Finally, the available days are allocated.
It is this third step that most often escapes operators, and it is the one that produces the most visible effect. At constant budget, strengthening one domain eases another: the days come from somewhere. An operator that accumulates findings on flight time management does not just receive more attention on that theme; it also gets, without knowing it, a loosening elsewhere — until the following cycle, when the authority catches up on what it had set aside. That is how the widespread sense arises of oversight that comes “in waves” whose logic is not understood.
The form of the intervention follows the same logic. A visit announced several weeks in advance tests procedures; an unannounced visit, a night stop, a line observation test practice. When the gap between the manual and practice becomes the inspector’s working hypothesis, the mode of intervention changes before the frequency increases. Suddenly receiving inspectors at unusual hours is a piece of intelligence in itself.
The paradox of the silent operator
A file with no internal finding, no voluntary report and no reported discrepancy produces exactly the opposite of the expected effect. No organisation operating aircraft runs without discrepancies; one that finds none has not proven it has none, it has proven that its detection system has no bite. The inspector then forms the most economical hypothesis, under-detection, and tests it in the most direct way: by re-examining, personally, a domain the internal audit declared compliant, over the same period and with a larger sample.
If the inspector finds something, two findings arise instead of one: the underlying discrepancy, and the failure of the compliance monitoring function that did not see it. The second weighs far more heavily on the profile than the first, because it does not concern a single domain but the measuring instrument itself. A finding on a domain closes one programme line; a finding on the management system reopens them all.
The trap: presenting a blank internal audit register as proof of compliance. This document does not demonstrate the absence of discrepancies, it demonstrates the yield of the function looking for them — and a zero yield is the one result an inspector cannot accept at face value. The inspector will go and check it, with a larger sample than yours, and over the period you declared compliant.
What really moves the profile, and over what timeframe
The movement is asymmetric, and that is the most useful property to know. A deterioration registers immediately: a major finding, a repeated discrepancy, a late notification alter the programme within the following weeks. An improvement, on the other hand, only registers after it has been demonstrated, hence after observation, hence at the earliest at the following cycle. An operator does not recover from a degraded profile by producing a corrective action plan: it recovers by producing, at the following cycle, proof that the plan did what it promised.
What matters, then, is not the closure but its quality. Three elements distinguish it from a mere response: an identified cause explaining why the existing system did not prevent the discrepancy, a corrective measure acting on that cause rather than its symptom, and a dated effectiveness check, carried out afterwards, showing that the discrepancy has not recurred. A closure that amounts to “staff were made aware” closes nothing: it tells the authority that the cause was not sought, and it invites the same finding at the following cycle, this time with a remark on the effectiveness of corrective actions.
The right reflex: date detection, correction and effectiveness check separately, and set the latter to a deadline after the administrative closure. Three distinct dates in a register are worth more, in an oversight file, than ten pages of corrective action plan: they prove someone came back to check.
Change: the only event that reprogrammes everything off-cycle
An operator is practically never caught out by a botched flight. It is caught out by a change. Adding a type, opening a base, extending the network into a new area, a change of capital control, outsourcing continuing airworthiness management, rapid fleet growth, replacing a nominated accountable manager: each of these events changes the inherent exposure, hence the profile, hence the programme — regardless of any fault.
The sensitive point is not the approval itself, it is the moment the authority learns about it. The table below sets side by side the two possible trajectories of the same change, and the gap between the columns is paid for in inspector-days over several cycles.
| Announced before implementation | Discovered after the fact | |
|---|---|---|
| Handling | Integrated into the already planned oversight programme | Off-cycle visit triggered on short notice |
| Burden on the operator | One file, one scoping meeting | One file, plus a response to a finding and a corrective action plan |
| Effect on the profile | Exposure adjusted, performance unchanged | Exposure adjusted and performance degraded |
| Likely finding | None, or a simple observation | Finding on change management, recorded against the management system |
| Duration of the effect | The duration of the change itself | Several planning cycles |
| Room for discussion | Scope, schedule and conditions negotiable | Conditions imposed, schedule endured |
The most dangerous category is that of changes that do not require prior approval, and which the operator, for that reason, treats as non-events. They nonetheless fall under the change-management process, which must identify them, assess their safety effects and keep a record. A change handled properly, correctly assessed and documented, costs the profile nothing. The same change discovered in an amended manual, on a flight plan, or in passing during a ramp inspection almost always opens a systemic finding.
What oversight abroad reports back to your authority
Ramp inspections conducted by other states on your aircraft do not stay where they took place. Their findings are classified by severity category, aggregated, and then reported back to the operator’s state, which turns them into one of the most regular entries in its profile. It is also the entry the operator understands least well, for a structural reason: locally, each finding is seen by a different crew, on a different stop, and looks trivial. Seen from the authority, the same minor discrepancy noted several times over a few months across different registrations is no longer a minor discrepancy: it is the signature of a defect in procedure, documentation or training.
The reflex most often missing is nonetheless simple: these results are accessible to the operator, and almost no one consolidates them. Reconstructing the aggregated view yourself — by type of discrepancy, by aircraft, by month — gives exactly the picture the authority is looking at, with a decisive advantage: the possibility of acting before it does, and of presenting the corrected trend rather than being subjected to it. operator-to-operator experience feedback show that this consolidation remains, in many organisations, the one safety task no one has formally been given.
What to go and look for in your own file
Reconstructing the essentials of the profile as seen from outside requires no tool: three registers are enough, provided they are read with the right questions.

- The findings register: what proportion was found internally rather than by the authority, what median delay separates detection from closure, how many deadline extensions were requested, and how many closures carry a dated effectiveness check.
- The occurrence report register: monthly volume, filing delay, the share of voluntary reports, and above all the list of departments that never report anything — that is the one the inspector will read first.
- The change register: for each change, the gap between the internal decision date and the notification date. A systematically negative gap is immediately visible.
A grey area remains, and it has to be accepted: the weightings, thresholds and scoring grids are not published, and they differ from one authority to another. Trying to optimise a score you cannot see is a waste of time. On the other hand, formally requesting disclosure of the oversight programme concerning you, and of the domains selected for the current cycle, is an ordinary step that many operators have never taken — it also signals, in passing, that the organisation reads its own file.
For anyone preparing for appointment to a nominated post-holder function, the stakes have shifted. The interview no longer concerns only knowledge of the regulations: it concerns the ability to describe the profile of one’s own operator, to name the two most exposed domains, and to say what has been done to correct them. It is also what is trained in courses devoted to the management system and to compliance, where reading a findings register is worth more than a recitation.
The applicable regulations, the decisions and procedures of the competent authority, and the operator’s approved documentation always prevail over this article, which is purely explanatory and does not replace any published procedure.
Questions fréquentes
Can you ask your authority to see the oversight programme concerning you?
Yes, and it is an ordinary step, even though few operators take it. Authorities generally disclose the domains selected for the current cycle and the tentative schedule of interventions, which makes it possible to prepare the right files rather than every file. The internal weightings and the scoring grids that led to this programme, on the other hand, are not published and are not disclosed.
Does an internal audit that surfaces many findings harm the operator?
No, it is the opposite that is worrying. An internal system that regularly finds discrepancies, analyses them and closes them with proof of effectiveness demonstrates that the compliance monitoring function works, which is precisely the point of oversight. A blank register leads the inspector to test a domain declared compliant personally, and a discrepancy found under those conditions counts as two findings instead of one.
How long does it take to come out of a degraded profile?
Longer than it took to enter it, because the mechanism is asymmetric. A major finding or a late notification alters the programme within a few weeks, while an improvement only registers once it has been observed, hence at the earliest at the following planning cycle. Providing a corrective action plan is never enough: it is the dated proof of its effect, found during the next intervention, that reopens the interval.
Do ramp inspections undergone abroad count towards my authority's oversight?
Yes, and they form one of the most regular entries in the profile. Their findings are classified by severity, aggregated and then reported back to the operator’s state, which sees a trend where each crew only saw an isolated incident. Consolidating these results yourself, by type of discrepancy and by aircraft, makes it possible to act before the repetition is interpreted as a systemic defect.
Must a change that does not require prior approval be notified?
It must in any case be handled by the internal change-management process, which identifies it, assesses its safety effects and keeps a record, even when no formal approval is required. The applicable procedure specifies what must be brought to the authority’s attention and within what timeframe. A change discovered after the fact in an amended manual or during an inspection almost always opens a finding concerning the management system itself.
Is rapid fleet growth enough to trigger enhanced oversight?
It mechanically increases inherent exposure, hence the share of the profile the operator cannot reduce, and it draws attention to the functions that rarely keep pace: training, supervision, continuing airworthiness, flight time management. It is not a penalty in itself, but it calls for a demonstration: showing that human resources and procedures were sized before the aircraft arrived moves the operator upward on the performance axis.

Leave a Reply
You must be logged in to post a comment.