A flight school adopts a new scheduling system, or moves its ground-school material onto tablets, and within a few weeks the complaints begin. Instructors cannot find slots, students arrive to find a simulator double-booked, and someone suggests the software is simply bad. In many of these cases the software is fine. What failed was a set of unspoken beliefs about how the tool would be used. The remedy is procedural and cheap: write down your assumptions about the rollout, in plain terms, and test them one at a time.
How the trouble first shows up
Nobody describes the symptom accurately at first. People say “the system is unreliable” when they mean “the system does not behave the way I expected.” That gap between expectation and behaviour is where diagnosis should start. Before touching configuration or calling a vendor, collect the expectations that were in play when the decision was made, because the fault often lives there.
Consider a typical case. A school moves briefing packs, checklists and weather material to a shared digital library. The plan assumed that every student would have a charged device, that the library would load in the briefing room, and that instructors would update documents the same way. Each of those is a belief about people, power and connectivity. Not one of them is a statement about the software.
Step one: list what the plan took for granted
Gather the people who designed and used the rollout and ask a single question: what did we need to be true for this to work? Capture every answer as one sentence. Resist the urge to debate during collection; the aim is volume first, quality later.
- Which devices and operating versions would everyone be using?
- Where would the network be strong enough, and where would it not?
- Who would own updates to shared documents, and how would others learn that something changed?
- What would happen on the day the system was unreachable?
- How long would a new user need before working unaided?
Each line becomes a candidate assumption. Some will be so obvious that nobody ever checked them, and those are often the most fragile.
Step two: rate each assumption and attach a check
Beside every statement, note how confident the group is, using only low, medium or high, and give a one-line reason. Then write the cheapest test that could confirm or refute it. A belief about connectivity can be tested by walking the building with a device and noting dead spots. A belief about document ownership can be tested by asking three instructors where they would save a revised checklist; if they give three answers, the assumption was wrong.
This table of belief, confidence and check is the core of the method. It replaces a vague mood of “things are going badly” with a ranked set of questions that can be answered on a Tuesday afternoon.
Step three: diagnose in order of consequence
Do not work through the list top to bottom. Sort it by two criteria: how shaky the belief is, and how much damage follows if it fails. A weak belief with low consequences can wait. A weak belief that touches safety documentation or student schedules goes first.
In aviation settings the consequence dimension deserves extra weight. Anything that affects what a student reads before a flight, or whether the current version of a procedure is in front of them, is not a convenience issue. It belongs at the top of the queue even if everyone feels the software is working well enough. As a reminder of how much routine preparation matters, the school’s earlier piece on how a student pilot can approach weather study shows the kind of material that a well-run digital library should keep current and easy to reach.
Common findings once assumptions are tested
Teams that run this exercise tend to discover the same handful of mismatches.
- Different mental models of ownership. Several people believed someone else maintained the master copy. Nobody did.
- Connectivity beliefs that were never tested. The rollout was demonstrated in one room and assumed to work in all of them.
- Training assumed to be optional. Staff were expected to learn on the job, yet each person learned a slightly different workflow.
- Fallback assumed to exist. The plan said paper copies would be available, but no one had printed them or placed them where they could be found.
None of these require a new tool. They require a conversation, a clarified rule and a short notice to everyone affected.
Step four: fix the assumption, then the system
When a test fails, resist the reflex to patch the software first. Ask whether the belief can be made true by a change in process. If the master-copy question had no owner, name one. If the network is patchy in the briefing room, decide whether the answer is a better access point or a rule that key documents are downloaded before arrival. Only when the process answer is exhausted should you reach for configuration changes or support tickets.
This order matters because process fixes are faster, cheaper and easier to explain. They also leave the school with a clear record of why a rule exists. If a new instructor asks months later why documents must be refreshed before each briefing block, the assumptions page supplies the answer without anyone reconstructing the history from memory.
Keeping the record alive
An assumptions page left untouched becomes decoration. Put a review into the calendar at each natural boundary: a change of cohort, a new aircraft type, an update to the system, a change of premises. At each boundary, reread the list and ask which items are now confirmed, which have been overturned, and which have been added by the new situation. Keep dated notes so the reasoning can be followed later.
Operations teams in IT use the same device when they plan changes to infrastructure, and the parallel is instructive for a technology-minded flight school. The school’s own coverage of practical cloud engineering for faster software delivery touches on that world. For readers curious about learning from it directly, NoOpsSchool is a training website that says it is based in India and lists certifications, courses, consulting and other services for IT practitioners; browse its course and consulting listings if you want to see how such a provider frames its offering, keeping in mind that these are the site’s own descriptions.
A short diagnostic checklist
- Have we written the beliefs behind the rollout, or only the features?
- Is every belief paired with a confidence level and a test?
- Have we ranked by consequence, with safety-relevant material first?
- Did we try a process fix before a technical one?
- Is there a date on the calendar to reread the list?
Closing thought
When a technology change disappoints a flight school, the cause is more often an untested belief than a broken product. Naming those beliefs turns frustration into a short, ordered set of checks. Do the writing once, rank by consequence, test cheaply, and keep the page current. The next rollout will start from a better position than the last one did.

