What Makes a Decision Worth Recording Formally
Most technical decisions never get written down because they don't feel significant at the moment they're made. By the time anyone realises a decision mattered, the reasoning behind it has usually already been forgotten.
Most technical decisions never get written down. They don’t feel like decisions at the time. They feel like the obvious next step, and recording something that seems self-evident feels like unnecessary overhead. By the time anyone realises a decision actually mattered, the reasoning behind it is usually gone, along with the person who made it or their memory of why.
How invisible reasoning produces two failures from one cause
There is a specific pattern worth tracing closely, because it explains something that otherwise looks like two unrelated problems.
A chair or senior figure makes a call. The reasoning behind it is never made visible, only the conclusion is stated. The room reads this as a personal preference rather than a technical position, because that is what an unexplained call looks like. Nobody formally challenges it. Challenging what appears to be someone’s personal preference feels like a confrontation. Challenging a stated technical position, with visible reasoning attached, does not. So the moment passes unchallenged, not because the decision was necessarily right, but because it was never framed as something open to technical challenge in the first place.
The same invisible reasoning then produces a second failure, later, when nobody asks whether the decision was significant enough to record. The test for significance, whether reversing it would be costly, whether it depends on assumptions that could later prove wrong, only gets applied to things recognised as decisions. A call that read as personal preference was never recognised that way. It skipped the test entirely, not through anyone’s deliberate choice, but because significance is a question asked about positions, not about preferences.
These are not two separate problems. They are one failure, invisible reasoning, producing two different downstream effects. Silence in the room when the decision is made. Silence in the record afterward.
The danger of treating decisions lightly during design
This failure compounds when engineers, under the ordinary pressure of a live design programme, start treating consequential calls as routine to avoid confronting them properly. A choice that genuinely deserves scrutiny gets waved through because raising it would slow things down, or because it resembles a dozen other calls that really were routine. The engineer making it is not being careless in any obvious way. They are responding to real pressure, and the decision simply does not announce itself as one that needed more attention than it got.
This is where the earlier pattern becomes dangerous rather than just inefficient. A decision made under this kind of pressure, without visible reasoning attached, is exactly the decision most likely to slip past both challenge and record for the same reason already described. It never enters the pipeline where anyone tests it for significance, because it was never surfaced as something requiring that test.
Why waiting is already too late
By the time anyone recognises that a decision mattered, the moment for capturing it properly has usually already passed. The people who made it have often moved to other work. The reasoning that seemed obvious at the time is no longer reconstructable with any confidence. What gets written down after the fact is a reconstruction, not a record, and a reconstruction written to justify a decision already made carries a different kind of risk than one written honestly at the moment the decision was actually made.
What a decision record actually needs to contain
A useful record does not simply state what was decided. It states the reasoning behind the decision clearly enough that a future reader can see it was derived from evidence, not asserted by whoever happened to be in the chair. It names the alternatives that were considered and why they were set aside. It states the conditions under which the decision remains valid, so that if those conditions change later, someone knows to revisit it.
Writing the reasoning down this clearly does more than create a record. It forces the decision to declare itself as a technical position at the moment it is made, which is precisely the thing that was missing in the failure pattern described above. A decision with visible reasoning cannot easily be mistaken for personal preference, and a decision that has declared itself a position naturally invites the significance test that an unexplained call slips past entirely.