The review was honest, the findings were correct, and three years later the same failure occurs again with different people.
This is the ordinary outcome, and it is not caused by the findings being wrong or by anybody ignoring them. It is caused by where the findings were stored.
Findings live in people or in structure
A finding held by the people who were present is real for as long as they remain and remember. Both conditions decay. In four years half of them have gone and the rest recall a general lesson rather than the specific mechanism.
A finding encoded into something that operates without anybody remembering it survives indefinitely and requires no vigilance. This is the entire distinction between an organization that learns and one that discusses learning.
The practical question for every finding is therefore not whether people accept it. It is what the finding becomes.
What survives
A step in something that already happens. A line added to an existing checklist, form, or agenda. The strongest form available, because it operates whether or not anyone recalls the reason.
A threshold that triggers automatically. Anything above this figure requires a second signature. This removes the judgment that failed rather than asking it to perform better.
A scheduled event with an owner. The contact list is verified in January and July by a named person. On a calendar, it happens; as an intention, it does not.
A change to what is hired for or trained on. Slow and durable.
What does not
Anything requiring sustained heightened attention. We will all be more careful about this decays within weeks and is not a finding, it is a mood.
A new document nobody has reason to open. A protocol added to a shared drive is stored, not encoded.
A one-off training session. It reaches the people present, none of whom are the people who will be in post in five years.
An understanding that lives with one person. Frequently the most sophisticated learning in the organization, and it leaves when they do.
Encode the reason as well
A structural change with no attached explanation creates a second problem some years later.
The step exists, nobody knows why, and it looks like inefficiency. Eventually somebody competent removes it, and the original failure returns. This is one of the more common ways organizations lose things they had genuinely learned.
One line attached to the change is enough. This check was added after the 2027 notification failure. Anybody proposing to remove it later can then assess whether the reason still applies, which is a legitimate decision, rather than removing something they assume is habit.
Fewer changes survive better
An organization that adopts eleven new procedures after an incident will be operating perhaps two of them within a year, and it will not have chosen which two.
The accumulated weight of unmaintained procedure also does something worse than nothing. It teaches people that documented process is decorative, which erodes compliance with the requirements that do matter.
Three changes, encoded properly, maintained, and checked once, outperform eleven every time. The discipline is choosing which three, and the criterion is which findings would have most changed the outcome.
The ninety-day check
The single step that separates organizations that learn from organizations that hold reviews.
Ninety days after the review, somebody checks whether the three changes exist and are operating, and reports that to the same people who attended. Not whether they were assigned. Whether they are in the form, on the calendar, in the threshold.
Almost nobody does this, which is why almost every organization holds reviews and relearns the same lessons. It takes twenty minutes and it is the whole difference.
The Idaho version of the problem
Small organizations and volunteer-governed bodies carry a sharper form of this, because their institutional memory is almost entirely personal.
A district that handled something difficult in 2019 learned a great deal, and by 2027 the board has turned over twice and the clerk has retired. Nothing was encoded, so nothing transferred, and the next occurrence is met by people starting from nothing.
For these organizations the encoding matters more rather than less, and it is cheaper: a line in the annual calendar, a paragraph in the minutes explaining why a practice exists, a page in the handover file. None of it is sophisticated. It is simply the difference between an institution that accumulates judgment and one that resets every few years.
Edited by Patrick J. Wolf, PhD