Most organizations treat “knowledge management” as a documentation problem: write down what people know, put it in a shared drive, done. Ikujiro Nonaka and Hirotaka Takeuchi’s research, published in their 1995 book The Knowledge-Creating Company, explains why this approach captures so little of what actually matters — and gives change leaders a more useful map of how knowledge really moves.

Two kinds of knowledge

Nonaka and Takeuchi’s starting point is a distinction borrowed from philosopher Michael Polanyi: explicit knowledge is what can be written down, codified, and transferred in a document — a process manual, a spreadsheet formula, a policy. Tacit knowledge is everything else: the judgment a veteran operator uses to tell when a machine sounds wrong before it fails, the read a sales lead has on which objections are real, the feel a project manager has for which risks on the register actually need attention this week. Tacit knowledge resists documentation not because people are careless about writing it down, but because much of it was never consciously articulated in the first place.

Most organizational knowledge is tacit. Most knowledge-management systems only capture explicit knowledge. That mismatch is why so many of those systems end up as large, underused archives.

The SECI model: four ways knowledge moves

Nonaka and Takeuchi mapped four conversion processes that move knowledge between tacit and explicit form, forming a continuous cycle:

Socialization (tacit to tacit). Knowledge passes directly between people through shared experience — an apprentice working alongside a craftsperson, a new hire shadowing a veteran, a team walking a problem together instead of reading about it. No document is produced; the knowledge transfers through observation and practice.

Externalization (tacit to explicit). Someone articulates a previously tacit insight into words, diagrams, or models — the moment an expert’s gut instinct gets written up as a decision rule, or a team’s unspoken way of handling a difficult client gets turned into a documented playbook. This is the hardest conversion in the cycle, and the one most knowledge-management efforts skip straight past.

Combination (explicit to explicit). Existing explicit knowledge gets reorganized, combined, or systematized into something new — merging two departments’ procedures into one, or synthesizing several reports into a single framework. This is the conversion conventional knowledge-management software is actually good at, which is part of why it gets mistaken for the whole picture.

Internalization (explicit to tacit). Documented knowledge gets absorbed through practice until it becomes instinctive — reading the new process manual is Combination; doing the job long enough that you stop needing the manual is Internalization. This is where explicit knowledge earns its keep, by quietly becoming someone’s tacit skill again.

Why this matters mid-change

Change programs are exactly when organizations lose tacit knowledge fastest — through restructuring, role changes, and departures of people who never got around to writing down what they knew because no one asked them to in a form that worked. A reorg that moves people around without first running some version of Externalization loses, quietly and permanently, the judgment those people built up in their old roles.

It also explains why training programs built entirely around manuals and e-learning modules (Combination) often fail to produce the skill they’re meant to build: without Socialization and Internalization alongside them, people can pass a quiz on the material without being able to use it under real conditions.

The diagnostic question

Before a reorganization, a departure, or a major process change, the useful question isn’t “have we documented our processes?” It’s: who holds tacit knowledge this change will scatter or lose, and is there a deliberate Externalization step — direct conversations, shadowing, structured debriefs — built into the transition plan to capture it before it walks out the door?