opening
the reading
the reading
runningChoosing a company to open on
Every reason carries the conditions that would change it, written before the filing arrives rather than after.
Under every reason in a reading there is a short list. It names the conditions that would change that reason — a metric turning negative, a disclosure disappearing, a ratio crossing a threshold.
Those conditions are set when the reason is derived, not when the filing arrives. That ordering is the entire point.
It is easy to say what a result means after seeing it. Give a person a number and they will produce an account of why it was predictable. Give them the opposite number and they will produce a different account, equally fluent.
This is not a failure of effort. It is how interpretation works when the evidence arrives before the standard. The standard adjusts to fit.
Writing the condition first removes the adjustment. If a reason rests on costs growing more slowly than revenue, and the condition says the reason comes under pressure when the efficiency ratio rises above a stated level, then the next filing either crosses that line or it does not. There is nothing to decide at the moment of reading, because the deciding already happened.
Not every statement of concern qualifies. A condition has to be the kind of thing a filing can answer.
That means it refers to something disclosed, not something inferred. It names a direction or a threshold rather than a quality. And it is specific enough that two people reading the same filing would agree on whether it was met.
Revenue growth turning negative is a condition. Competitive pressure increasing is not — no filing settles it, so it would sit under the reason forever, looking like vigilance and functioning as decoration.
The discipline is uncomfortable in the same way the reasons themselves are. Some things that would genuinely worry you about a business cannot be written as a condition a document will answer. Those stay out, and the reason is weaker for their absence, which is accurate.
A condition says what would change the argument. It does not say what is likely, and it does not say what to do.
A condition being met does not mean a reason is finished. It means the thing that was identified in advance as the test has happened, and the reason now has to be looked at rather than assumed. What follows from that is the reader's judgment, operating on better information than it would otherwise have.
Most conditions are never met. That is the expected case, and a reason sitting intact for eight quarters with its conditions unmet is a reason that has been tested eight times rather than one that has been ignored.
The conditions appear on the free reading, next to the reason, before anything has happened to either.
They are published because a test nobody can see is not a test. If the conditions stayed internal, every result the engine produced would arrive with its own justification attached, and there would be no way to check whether the standard had moved. Putting them in public, dated, before the filing, means the standard is fixed by the time the evidence shows up.
It also means we can be visibly wrong about what mattered. A reason can weaken through something none of its conditions anticipated, which says the conditions were poorly chosen. That stays on the record too, because a list of tests that quietly improves after each result is not a list of tests.
The useful part of this does not require the product.
For anything you hold, the exercise is to write down, in advance, the specific disclosed fact that would make you reconsider. Not a feeling about the business. A line in a document, with a direction attached.
Most people find this harder than expected, and the difficulty is the finding. A position you cannot write a condition for is one you will hold through anything, because nothing you have specified can contradict it.