The file of the candidate for the training manager post has been sitting on the table for three weeks: fifteen years on the line, an active instructor, but a profile that does not tick one of the experience criteria in the published means of compliance. The documentation department says no, operations management replies that no one else will apply this year, and the authority, asked by phone, hints that an alternative means is possible. The word shifts the problem: it is no longer a question of whether the candidate is suitable, but of establishing in writing that the approach chosen achieves the objective the published means was serving. Between the two formulations lies a case file, and the quality of that file decides whether the decision falls within a few weeks or drags across several quarters.
What binds, and what merely proves
A regulatory requirement states an objective and imposes a result. The means of compliance published by the Agency says how to reach it in a way that will be recognised without discussion: it adds no obligation, it creates a presumption. Whoever applies it is deemed to satisfy the requirement, and it would fall to the authority to prove otherwise.
Departing from it is therefore not an infringement. It is a reversal of the burden of proof. The operator who proceeds differently must establish itself that its approach achieves the same objective, and it must establish this before applying it, not at the moment it is asked to. The whole mechanics of the alternative means lies in this shift, and most files that get bogged down have treated it as a negotiation rather than as a demonstration.
Three kinds of text coexist and are not discussed in the same way. The rule is binding as it stands. The acceptable means of compliance is one path among others, holding a privileged status. The guidance material is neither one nor the other: it explains, it proves nothing, and proposing an alternative means to a piece of guidance material makes no sense, since there is nothing to replace. The same confusion produces files that argue against an illustrative example, which never imposed anything.
The costliest misunderstanding is believing that the rule itself can be discussed. If what is troublesome is the requirement itself, no file will succeed, however well made: one then falls under another mechanism, rarer, more tightly framed and considerably less generous.
The rule sets the goal, the published means is one approved way to reach it. If you choose a different path, you must prove it works just as well – before using it, not after. The authority doesn’t have to prove your way is bad, you have to prove your way is good.
An operator wants to use a Boeing 737-800 for ETOPS 180 operations but the published means requires 200 hours of fleet-specific training. They propose 150 hours plus additional simulator sessions. They must prove this achieves the same safety outcome before applying it, not just claim it’s “similar experience.”
Alternative means, exemption, addition: three different doors
Everyday jargon mixes together three things that share neither the same addressee, nor the same duration, nor the same standard of proof. Telling them apart takes five minutes and avoids filing at the wrong door, which does not cost a refusal but a polite silence followed by an invitation to start all over again.
The alternative means replaces a published means while still serving the same requirement. An exemption, by contrast, attacks the requirement itself: it falls under the flexibility provisions, presupposes an urgent and unforeseen operational need, is granted for a limited period, and is decided at a level higher than the office that follows the operator day to day. The addition, finally, is nothing at all in regulatory terms: adding a step to a published means that continues to be applied in full requires no approval.
| Criterion | Alternative means of compliance | Exemption or relief | Addition on top of the published means |
|---|---|---|---|
| What it targets | the path: the published means | the objective: the requirement itself | nothing; the published means remains applied |
| What must be demonstrated | the equivalence of the result obtained | the urgency, the unforeseen nature and the limited character | nothing |
| Who decides | the competent authority, before application | a higher decision level, State concerned | the operator alone |
| Duration | indefinite, as long as the assumptions hold | capped from the moment it is granted | not applicable |
| Downstream follow-up | the authority informs the Agency, which may assess it | information to other States and to the Agency | none |
| Trace in the documentation | amended text plus a register of alternative means | mention and end date | internal procedure, no particular status |
| Effect of early application | non-compliance with the requirement during the period | the same, as the requirement itself remains unmet | not applicable |
The line that decides is the first one. If the difficulty concerns the result required — an objective the operation cannot reach — this is not the alternative means. If it concerns the path — a qualification criterion, a periodicity, the make-up of a team, the form of a record — this is the right mechanism, and the rest is merely a matter of the file.
The addition deserves a note, because that is where undeclared alternative means hide. Adding is free; taking away is not. An internal procedure that reproduces the published means « adapting it » and that, along the way, drops one of its elements is not an addition: it is an alternative means applied without approval. It is discovered at the first serious review of the documentation, and it is discovered all the faster the more carefully the adaptation was written — a clean text compares line by line against the published means.
The file: what an examiner looks for, and in what order
A well-built file reads in one go. Poorly built, it comes back, goes out again, and the decision slips from one quarter to the next. The reading order is remarkably stable from one examiner to another, because it follows the logic of proof, not that of drafting.

- the requirement targeted, quoted in the terms of the binding text;
- the published means concerned, and above all the precise element being replaced;
- the proposed alternative text, drafted exactly as it will appear in the documentation;
- the demonstration of equivalence, element by element;
- the risk assessment and the compensating measures adopted;
- the impact matrix on the rest of the documentation, training and records;
- the exact scope, the monitoring indicators and the review date.
The first item is the one most often botched. Many files open with the published means and its supposed shortcomings, whereas the examiner has no standing to judge a text the Agency wrote. What is judged is an equivalence, and it can only be judged if the objective is stated first. Putting that objective into one’s own words, in a single sentence, is the most worthwhile exercise in the whole file: it sometimes reveals that the alternative means envisaged does not serve the same purpose at all.
The second item is the mapping. A published means is not a single block: it is a stack of conditions, of which usually only one is replaced. A file that announces it is replacing « the published means » condemns itself to demonstrating equivalence for everything, whereas a file that isolates one condition and leaves the others untouched reduces the demonstration to a single point. The two-column table — published condition on the left, proposed provision on the right, with an « unchanged » box everywhere else — is the format that saves the most time, for the examiner as much as for the drafter.
The demonstration of equivalence is not judged by its length. It is judged by its shape: it must link a feature of the proposed means to the effect that the published condition produced. If the published condition aimed to guarantee that a person had witnessed a certain number of situations, saying the candidate is « very experienced » demonstrates nothing; describing what, in his or her case, made those situations witnessed, and how this will be verified, demonstrates something. The practical rule: every sentence of the demonstration must be capable of being contradicted by a fact. A sentence that no fact could contradict is not an argument, it is an opinion.
The risk assessment and the compensating measures close the file. They are too often treated as one more formality, whereas they carry lasting commitments: a compensating measure is a permanent, checkable burden, whose abandonment brings down the equivalence. Enhanced supervision for a period, mentoring, an extra checkpoint in the internal audit programme — each of these measures will have to live, be proven and be reviewed. It is better to propose two that will actually be kept than five that will be forgotten within eighteen months.
The right reflex: write the objective of the requirement and the element-by-element mapping before drafting a single line of the alternative text. Drafted first, the text fixes the solution, and the demonstration ends up being written to justify it. Mapped first, the real gap appears while it is still free to reduce, and it almost always gets reduced.
The delay is not a queue
Asked about the examination time, an authority rarely gives a firm duration, and for good reason: the share that falls to it is often not the longest. The actual delay breaks down into four components, and the heaviest one is the one the operator holds itself, without always realising it.
The first component is admissibility. This is where it is checked that the documents exist and that the request does indeed fall under the mechanism invoked. A file that fails here will not be examined on the merits: it will be sent back, often with a single sentence, and the question of who read it for how long becomes moot.
The second is the technical review proper, and it depends on the specialty concerned: an alternative means touching on training, maintenance or performance does not mobilise the same people, and some of them are few. The third component, invisible in the exchanges, is the time the operator takes to respond to a request for further information, and it is frequently the longest of the four: the response requires reopening work the team thought was finished, often with drafters already moved on to something else. The fourth is administrative: the decision, its notification, and the authority informing the Agency, which follows rather than precedes the use.
A practical consequence follows from this, and it belongs to the operating calendar, not to the documentation department. The date that matters is the date of approval, never the date of filing. Taking up a post, opening a base or starting a campaign planned around the filing date amounts to betting on a decision that is not within one’s control, and the bet is settled by an unpleasant choice: delay, or apply without approval.
The trap: treating the authority’s silence as tacit approval. The alternative means presupposes prior approval; without it, the operator applies neither the published means nor an approved means, and is therefore in breach of the requirement itself throughout the period. Having filed a complete file in good faith does not erase the gap: it merely documents its duration precisely, which is exactly what an auditor will seek to establish.
The precedent: what it gives, what it does not give
« A competitor got it » is the most common argument and the least effective as such. Approval is granted to an applicant, for a scope, in a context: it does not transfer, does not set a precedent in the legal sense, and binds no authority. Two operators answering to two different authorities can receive two different answers, without either being at fault.
Dismissing it as worthless nonetheless costs time, because what it brings is precise: the reasoning behind the review already exists, the objections have already been raised once, the compensating measures have already been discussed, and an examiner who has already settled a neighbouring case spends time on the differences rather than on the principle. A file that frankly sets out the kinship — same requirement targeted, same element replaced, context compared point by point — reads faster than a file that pretends to invent from scratch.
The precedent also plays out at a higher level. An alternative means that several operators request, obtain and apply without incident ends up constituting a signal: this is how published means get revised, and the best fate for an alternative means is to disappear by becoming the published means. This mechanism is slow and cannot be steered; it nonetheless justifies making one’s file readable for someone unfamiliar with the house, since it may be read well beyond the office that reviews it.
The second trap: reusing the text of an alternative means obtained elsewhere. The proposed wording is bound up with a demonstration that rested on the other organisation’s assumptions — fleet size, supervision ratio, network stability, existing oversight arrangements. Transplanted without those assumptions, it produces a file whose demonstration connects to nothing, and the examiner who recognises the text will read the rest with heightened scrutiny. One reuses the template, never the proof.
What becomes of an alternative means once granted
Approval is not the end of the file, it is the start of its upkeep. The alternative means enters the approved documentation with its scope and its conditions, and it thereby becomes an object subject to audit on the same footing as the rest — except that an auditor will pay it more attention, because it is a point where the operator asked to do things differently.
Three events kill it, and none of them sends a warning. The revision of the published means, first: it may absorb the alternative, make it unnecessary, or contradict it, and no one will warn the operator that its file has become obsolete. The change of assumptions, next: the fleet doubles, a base opens, the person whose profile justified the demonstration leaves, and the equivalence demonstrated now concerns a situation that no longer exists. The silent disappearance of a compensating measure, finally, which is the most frequent case: mentoring stopped when the mentor changed post, the extra checkpoint fell out of the audit programme during a revamp, and no one connected it to the approval granted three years earlier.
Hence the most useful and least widespread object: a register of active alternative means. One line per means, with the requirement targeted, the element replaced, the reference and date of the approval act, the scope, the compensating measures and their owner, and a review date. This register has a second virtue: presented spontaneously during a visit, it turns a subject of suspicion into a demonstration of control, whereas a list pieced together under pressure produces exactly the opposite effect.
The right reflex: attach each compensating measure to a function, never to a person, and enter it into the internal audit programme on the very day of approval. A measure that appears in no audit plan is not a measure: it is an intention, and it does not survive the first organisational change.
What to go and check at home
Four checks give an honest picture in one morning. Draw up the list of active alternative means and, for each one, retrieve the dated approval act — a means for which no one can produce the act must be treated as unapproved until proven otherwise. Then check that every compensating measure is still running and that an audit covers it. Reread the internal procedures that « adapt » a published means, looking for the element that was dropped rather than the one that was added. Finally, ask who in the house monitors revisions of the published means touching the areas concerned: if the answer is « no one in particular, » the first alternative means has probably already gone stale.
There remains a grey area that no one settles in writing: the boundary between interpreting a published means and replacing it. A published means often contains terms that call for judgement — proportionate, sufficient, suited to the size of the organisation. Adopting a low-end interpretation is not an alternative means as long as the element is still served; it becomes one once the interpretation empties the element of its effect. The line is drawn case by case, and the only workable safeguard is to write down the interpretation adopted and its justification at the moment it is adopted, while the reasoning is still fresh in memory.
For anyone preparing for a nominated post-holder role or a management interview, three questions do the sorting: how many active alternative means does your organisation have, what compensating measure supports each one, and who holds the approval acts. A candidate able to answer these about their own operator has understood the mechanism better than one who recites the definition. Concrete cases are discussed on the FL330 forums; preparation for management and compliance roles is covered under training.
This article describes a mechanism and does not replace any official source. The text applicable in the State of issue of the approval, the positions published by the competent authority and the Agency, and the operator’s own approved documentation take precedence over what is written here.
Questions fréquentes
Can an alternative means of compliance be applied as soon as the file is submitted, if the authority does not respond?
No. The mechanism relies on prior approval: until it is granted, the operator applies neither the published means nor an approved means, and is in breach of the requirement itself. The authority’s silence does not amount to approval and grants no right to apply. The useful course of action is to follow up in writing, keep a record of that follow-up, and keep the published means in force until the decision is made.
Does an alternative means approved for another operator apply to us?
No, approval is granted to an applicant for a given scope and context, and it does not transfer. The precedent brings something else: the reasoning behind the review already exists, the objections have already been raised once, and the examiner spends time on the differences rather than on the principle. One therefore reuses the file’s template, never the demonstration, which rests on the other organisation’s assumptions.
Must an alternative means be declared when we do more than the published means requires?
No, provided the published means continues to be applied in full: adding a step or a check falls under organisational freedom and requires no approval. The question arises differently as soon as an element of the published means is dropped or replaced, even within a procedure presented as a mere adaptation. The test consists of placing the published means and the internal procedure side by side and looking for the element that has disappeared, not the one that was added.
What happens to an alternative means when the Agency revises the published means it replaced?
It must be reviewed, because the revision may absorb it, make it unnecessary, or contradict it, and no automatic warning will reach the operator. If the new wording of the published means now covers the case, the alternative means is dropped and the documentation reverts to the marked path, which simplifies upkeep. If it contradicts it, a new file is required, and the old approval offers no protection in the meantime.
Can an alternative means apply to a single person or a single aircraft?
Yes, and a narrow scope is often easier to obtain than a general one, because the demonstration of equivalence then concerns a described, verifiable case. The trade-off is that the approval expires with its object: the person’s departure or the aircraft’s exit from the fleet ends the benefit, and a new case requires a new file. The scope must therefore be written unambiguously in the request, failing which it will be interpreted in the most restrictive way.
What is the practical difference between an alternative means and an exemption?
The alternative means targets the path and leaves the requirement intact: it is demonstrated by the equivalence of the result and lasts as long as its assumptions hold. The exemption targets the requirement itself: it falls under the flexibility provisions, presupposes an urgent and unforeseen operational need, is granted for a capped period, and is decided at a higher level. Filing an alternative-means case when the real objective is to escape the requirement always leads to the same outcome: a long review ending in refusal.

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