The Tacit Knowledge Problem: What Experts Know That They Don't Know They Know

Ask a senior engineer to write down everything a new hire needs to know about the system they own. They will produce something useful and incomplete. Not because they are withholding — they will genuinely try. But because the most valuable parts of their expertise are not accessible through conscious recall. They have automated the judgment calls. They have internalized the patterns. They no longer think about the decision framework they use because it fires automatically after ten thousand repetitions. That expertise is tacit — embedded in practice rather than articulable in documentation.

This is not a character flaw. It is how expertise works. Cognitive science calls it proceduralization: as a skill becomes expert-level, it moves from declarative memory (facts you can state) to procedural memory (patterns that execute automatically). The expert stops being able to explain what they are doing because at the expert level, they are not doing it consciously. They are just seeing the answer.

The practical consequence is that asking experts to self-document their expertise is one of the least effective capture methods available. The documentation they produce reflects what they think a novice needs to know, filtered through the assumption that much of what they know is obvious. It systematically underestimates the depth and specificity of their expertise. A senior customer success manager asked to document the renewal process will write six steps. A knowledge interviewer watching them execute a renewal will observe thirty-seven distinct judgment calls that never appear in the six steps — including the one judgment call that determines whether a renewal closes or churns.

The tacit knowledge problem is not a documentation problem. It is an extraction problem. The knowledge is there. The mechanism for getting it out requires a different approach than asking someone to write.

1. Explicit Knowledge Facts, rules, and procedures the expert can state directly when asked. 'We require three signatures on contracts above $500K.' Explicit knowledge is the easiest to capture and the lowest-value tier — because every competent practitioner in the field has access to the same explicit knowledge. The competitive advantage is in the tacit layer.
2. Tacit Knowledge Judgment calls, pattern recognition, and internalized expertise that the expert applies automatically. 'I can tell from the tone of this email that the customer is considering leaving, even though they haven't said it directly.' This cannot be captured by asking the expert to write it down. It requires extraction through conversation, observation, and case analysis.
3. Collective Knowledge Shared understanding that exists at the team or organizational level and is not held by any single individual. 'Everyone on this team knows that client X needs the report delivered by Tuesday regardless of what the contract says.' Collective knowledge is invisible because no one person feels responsible for documenting it — everyone assumes someone else has. This is often the most operationally critical knowledge in the system.

Collective knowledge is the third type and the one most commonly overlooked in KM programs. It surfaces most clearly when a team is broken up — through a reorg, a product pivot, or mass turnover — and the successor team has none of the shared context that made the original team effective. The successor team is not incompetent. They are operating without the accumulated collective knowledge that took the original team two years to develop. Capturing collective knowledge requires different techniques than individual expert interviews — team workshops, after-action reviews, and systematic documentation of informal conventions and agreements.