The Intel Brief · Transmission 006
Why Avionics Quotes Fail
· 4 min read
Almost every avionics retrofit that goes sideways went sideways at the desk, weeks before anyone opened a panel. The parts were wrong, the integration was never checked, and the hours came from memory. The shop finds out with the aircraft already torn down.
The quote is written before anyone looks at the architecture
A customer sends a wish list. Somebody prices the boxes, adds a labor number that felt right on the last job, and sends it. What did not happen is the part that matters: nobody drew what feeds what, over which bus, with which adapters, and nobody checked whether the combination is legal and complete.
That work still gets done eventually. It just gets done later, with the panel open and the clock running, which is the most expensive possible time to discover it.
The gaps nobody scoped
The pattern repeats across shops. A magnetometer that the display needs and nobody listed. A backup battery the STC requires. An altitude source, an audio path, a position source, a standby. None of these show up on a customer wish list, because customers ask for the box on the panel, not the parts behind it.
Mission-level requirements are worse, because no single component declares them. An aircraft configured for a specific operating environment may need equipment that nothing else in the scope calls out. Catching that requires looking at the panel as a system rather than a parts list.
The conflicts a parts list cannot show
Two units can each be perfect for the job and still refuse to talk. A legacy radio with no compatible output for a modern display. A device that only speaks CAN in an install with no bridge for it. Two boxes that share no data bus in common, so the data path everyone assumed exists does not.
These are not exotic failures. They are ordinary, they are documented in the install manuals, and they are invisible until somebody cross-references those manuals against each other. Doing that by hand for a full panel is hours of work, so under schedule pressure it gets skipped.
Hours from memory, and quotes that cannot be defended
Labor is usually the softest number on the page. It comes from what the last similar job took, adjusted by feel. The trouble arrives when the customer asks why the number moved, or why it is higher than the shop down the road, and there is no benchmark to point at. Then the shop either eats the difference or spends its credibility defending a guess.
The Aircraft Electronics Association publishes regional labor rate data every year for exactly this reason. Quotes checked against a published range hold up in conversation. Quotes built on recollection do not.
Move the discovery earlier
None of this is a skill problem. Shops that get burned are staffed by people who know the equipment cold. It is a sequencing problem: the verification happens after the commitment instead of before it.
Everything above is checkable at the desk, on paper, before a quote leaves the building. Gap analysis against mission requirements. Conflict detection across the data buses. An architecture drawing that shows the real paths. Labor benchmarked against published rates. Done in that order, the surprises move from the hangar floor to the planning stage, where they cost minutes instead of days.
That sequence is what we built AV8MX Studio Pro to run.