What exactly disappears
The procedures stay. The drawings stay. What disappears is the layer around them: why that one step is in there, which sound does not belong, what you do when the reading is fine and it still does not feel right.
That layer is not anecdote. It is the reason an experienced operator finds a fault in ten minutes where a new colleague takes a day, and sometimes the reason an incident does not happen.
Organisations only notice the loss afterwards. Somebody leaves, and six months later it turns out an operation that always ran smoothly now goes wrong regularly. Nobody makes the connection, because the knowledge did not disappear all at once but with one person.
Why it is written down nowhere
There are two reasons, and neither is unwillingness.
The first is that the knowledge has largely become unconscious. Ask an experienced specialist how he does something and he tells you the procedure. Ask what goes wrong if you do it slightly differently, and the interesting part arrives. He knows it perfectly well, but it does not occur to him to say it, because to him it is obvious.
The second is that there is never a moment for it. Knowledge transfer sits in everybody’s annual plan and in nobody’s week. It slips until the final month, by which point it no longer fits.
How you get it out
What works in practice is not an interview but a series of focused sessions, and the way you ask decides almost everything.
Do not ask how something is done, because that is in the procedure. Ask what goes wrong. Ask where he has seen others get it wrong. Ask about the time it nearly went badly. Ask why that one step was put in; there is usually a story behind it that makes its importance clear in one go.
Two people at a time works better than one. They complete each other, they contradict each other, and the contradiction itself carries information: evidently there are two ways of doing it and not everybody agrees which is better. That is exactly the kind of thing that is recorded nowhere.
Record it. Not to publish the recording, but because you will want to go back to how he said it.
Starting in time
The practical mistake made most often is starting too late. Three months before departure, the diary is full of handover, farewells and remaining leave.
Two years is realistic. That leaves room to do it calmly and, more importantly, the expert is still there to check what has been made of it. Material he approved himself carries a standing that material assembled after he left never gets.
The wisest opening question is not who is leaving soon, but this: which operation would become a problem if this one person fell ill tomorrow? That list is shorter than you think, and it is exactly the list to start with.
For many organisations that list is the starting point of a broader training programme. How to build on from there is in the guide on setting up or renewing a corporate academy. Our approach is described under capturing expert knowledge.