The Wikipedia definition of "tacit knowledge" is:
Knowledge that is difficult to extract or articulate and is therefore more difficult to convey to others through verbalization or writing.
The typical examples people use to describe this is like riding a bike or learning how to snowboard. It's illustrative because it's very hard to explain the sensation other than "you'll feel it", but something that we can all intuitively understand. But tacit knowledge in a *business setting* is harder to explain.
In a business, it can be things like:
- How a top salesperson knows when a customer is genuinely ready to buy
- How an experienced scientist weighs conflicting evidence
- How a manager senses that a project is about to go off track
- How a therapeutic area head knows when results are too good to be true
They aren't necessarily working from secret information and they are often looking at the same data that everyone else on the team is. What distinguishes them is how they interpret the information, what pieces they distrust, which parts are considered meaningful and what prior experience has taught them.
Their judgment is the compressed product of everything they have seen before, especially the things that failed.
This is why tacit knowledge is so valuable to an enterprise. It shapes how the organization actually operates, even when it is absent from its formal processes. It determines which evidence is believed, which risks are tolerated, which exceptions are made, and which decisions are escalated.
In mid-2026, there is incredible momentum building to keep tacit knowledge sovereign. The argument is that the enterprise value is in the proprietary knowledge of *how things should work according to the enterprise*. There is an implied requirement of performance or some goal that we are optimizing toward as an enterprise. If this knowledge is so valuable, why hasn't this been written down? What makes it hard to write down and codify? What makes it so valuable?
Let's unpack this using a real-world situation.
I met an HR exec at a major retailer on the night of their promotion celebration. She had a reputation for being the "best at what she does". Her HR experience came from the supply chain logistics industry, so I was curious how that translates into the retail business. The official language used for her unique capability was "fostering alignment," which I've come to understand really meant union busting.
I asked her what made her so effective and the answer she gave struck at the heart of what tacit knowledge means.
If her capability could be reduced to a policy or some internal checklist or standard operating procedure, the company could simply document it and distribute it to every HR leader. Apart from the structural reason of why this shouldn't be documented in the first place, the value is unlikely to reside within any single action.
It was knowing which employee concerns were isolated complaints and which indicated a deeper loss of trust. It was knowing who had credibility within a store, when management intervention would improve a situation, and when that same intervention would make it worse. It was understanding the incentives of employees, managers, and the company and recognizing how those incentives were interacting before the consequences became visible.
The important distinction is that tacit knowledge is not always knowledge that is literally impossible to describe. Often, it is knowledge that cannot be separated from the context in which it is applied. For example:
- This signal matters, but only under these circumstances
- Credible result, but not if produced by this assay
- Signal is likely false positive, given enough measurement of this other result
What makes this tacit is because this type of knowledge only becomes visible at the point of decision. Drug discovery and drug program decision making is filled with this kind of knowledge.
For example, if you ask a therapeutic area head: "What makes a good drug target?" you'll be met with a blank stare, a long pause, and palpable frustration brewing from your question.
The more revealing questions are:
- Why did you distrust this target despite the positive data?
- Why did one contradictory experiment matter more than five supportive ones?
- Why was this safety signal considered manageable in one program but fatal in another?
- What previous failure does this result remind you of?
- What evidence would cause you to reverse your recommendation?
These questions expose the reasoning that exists beneath the formal process.
They also expose something more uncomfortable: expert judgment is not formed in a vacuum.
Scientists operate within budgets, timelines, portfolio priorities, technical constraints, organizational politics, and personal accountability. A discovery scientist, a therapeutic-area head, and an R&D executive may interpret the same evidence differently because they are responsible for different consequences.
That does not necessarily mean one of them is irrational or acting in bad faith. They may be reasoning from different altitudes within the decision-making system.
One person sees the experimental result. Another sees how similar results have behaved across ten previous programs. Another sees the capital allocation decision, the competitive landscape, the regulatory path, and what must be removed from the portfolio if this program advances.
The apparent disagreement may therefore reflect missing context rather than poor reasoning.
This is why we need to narrow what we mean by tacit knowledge.
We shouldn't attempt to capture every opinion, instinct, or unwritten organizational behavior, but instead we should focus on the reasoning used to make consequential decisions:
- Which evidence is relevant
- Which evidence is trustworthy
- Which uncertainties are acceptable
- Which alternatives are genuinely available
- Which assumptions must be true
- What would cause the decision to change
- Who is accountable for the consequences
The decision made at the end is only the visible output.
The real organizational knowledge is contained in how the choice was constructed.
This also explains why asking an expert to "document how they think" usually produces a disappointing result. Experts often cannot reconstruct their reasoning in the abstract. Their knowledge has been compressed into pattern recognition. They may offer principles that sound sensible but fail to capture what they actually do when confronted with a difficult case.
A more effective method is to place a concrete recommendation in front of them.
Ask them to approve it, reject it, or modify it. Then capture the override:
- What did the recommendation get wrong?
- Which evidence was overweighted?
- What important context was missing?
- What pattern did the expert recognize?
- What would have to be different for the original recommendation to become acceptable?
Tacit knowledge is easier to extract through disagreement than explanation.
The objective is not to turn every scientist's intuition into a rigid rule. Drug discovery contains too much uncertainty for that. The objective is to make the reasoning behind consequential decisions observable, contestable, and reusable.
