In short. Mapping critical knowledge means starting from processes and their tasks, linking each task to the knowledge it requires, establishing who holds it, then computing vulnerability × relevance. The red zone of the map holds a few dozen items. On one site, allow four to eight weeks for a first defensible map.
A map is not an inventory. Listing five hundred pieces of know-how is useless: nobody can act on five hundred lines. A useful map is judged by what it allows you to decide — therefore by its ability to rank, and to reduce the list to what deserves a budget and a schedule.
Step 1 — Start from processes, never from people
Describe your processes and their tasks, then link to each task the skills and knowledge it requires. It is long the first time, and it is the only work that cannot be skipped: without that link no measurement means anything, because you cannot say what a piece of knowledge is for.
Step 2 — Know who holds what, without spending a year on it
Three sources combine, from the fastest to the most reliable: the versatility matrix you already keep in Excel, which imports as is; peer appointment, where everyone answers “who do you call?” for their own tasks; and mastery tests, which measure instead of declaring. A measured level always prevails over a declared one.
Step 3 — The calculation: vulnerability × relevance
Vulnerability answers: how threatened is this knowledge? It accounts for the number of holders, their actual level, their age and the departures already scheduled, and the difficulty of relearning. Relevance answers: what does it cost if it disappears? It accounts for the processes served, their criticality and the impact of a stoppage. The product of the two gives the criticality, on a scale a board can read.
The result reads like a map: every knowledge item is a dot, placed along those two axes. Top right, the red zone — a few dozen items, never hundreds. That is your work plan for the next twelve months.
Step 4 — Attach the documents, and see the holes
Once the map is drawn, connect your document systems. Every page attached to a knowledge item becomes its evidence. And the most useful information then appears on its own: the red-zone knowledge items with no reference document at all. Those live only in people's heads, and the map tells you whose.
How long for the first map?
On a reasonable perimeter — one site, two or three processes, fifty to two hundred people — count four to eight weeks for a first defensible map, most of which goes into describing processes and assigning people to tasks. The rest computes itself, and updates at every test, every departure, every new assignment.
To situate the tools: Wiki, document management, knowledge management.
Try it free for fourteen days — the trial space already contains five years of simulated data, so you see the map at once.
See also: Knowledge criticality: how it is computed · Retirement: how to stop losing know-how · Your first hour: a first real diagnosis · Skills mapping: method and software · Key-person risk: knowledge held by one person · How to build a knowledge management system from scratch
Frequently asked questions
Where should a knowledge map start?
With processes, never with people: describe the processes and their tasks, then link each task to the skills and knowledge it requires.
Can we import our Excel versatility matrix?
Yes, it imports as is. It combines with peer appointment and mastery tests; a measured level always prevails over a declared one.
How many knowledge items end up in the red zone?
A few dozen, never hundreds. That is your work plan for the next twelve months.