The Half-Life of Organisational Knowledge

EveryEvery team has a version of the same story. A system has one person who really understands it — not the person listed as the maintainer in the documentation, the person everyone actually asks when something breaks in a way the documentation doesn't cover. Then that person leaves, retires, or moves to another project, and the system keeps running for a while on momentum before the first problem arrives that needs the thing only they knew. The knowledge didn't vanish the day they left. It started decaying the day they stopped being reachable, and the decay is usually invisible until the exact moment it matters.

Storage and Retrieval, Applied to People FoundationalKnowledge that endures for decades — core principles

ThisThis site's own account of individual memory distinguishes storage strength — how durably something is encoded — from retrieval strength — how easily it can actually be accessed right now, following Bjork and Bjork's new theory of disuse¹. The same split applies to an organisation's knowledge of itself, almost unchanged. A departed expert's knowledge isn't instantly deleted from the organisation — it persists in code comments, old tickets, a half-remembered conversation someone else was in the room for. What collapses on their last day isn't storage, it's retrieval: the knowledge is still in there somewhere, but the one fast, reliable index into it just left the building. A system that used to have one-hop access to an answer — ask the expert — now needs a much slower, lossier search through whatever fragments got left behind.

Converting Tacit Into Explicit, on Purpose FoundationalKnowledge that endures for decades — core principles

IkujiroIkujiro Nonaka and Hirotaka Takeuchi's account of organisational knowledge creation treats this conversion — tacit knowledge, held personally and demonstrated rather than stated, turned into explicit knowledge that can be written down, stored, and transferred — as the central mechanism by which a company's knowledge becomes durable rather than person-shaped². This cluster's page on communities of practice already describes how that knowledge got into one person's hands in the first place — absorbed through legitimate peripheral participation, not handed over as a document. Nonaka and Takeuchi's point is that getting it back out again, into a form the organisation can keep after the person who holds it is gone, is not the reverse of the same easy process. Explaining a decision after the fact, in words, to someone who wasn't there for the thousand small cases that built the intuition, routinely produces something thinner than what's actually being used — a documented rule that's true most of the time, standing in for a feel for the exceptions that never quite makes it into the document.

What Resists Conversion, Even When Someone Tries FoundationalKnowledge that endures for decades — core principles

SomeSome of what an expert holds survives being written down perfectly well — a decision rule, a checklist, a diagram of how the pieces fit together. Some of it doesn't, not because the documentation effort was careless, but because a different kind of knowledge was being asked to survive a format built for the wrong one. A seasoned troubleshooter's sense of which three things to check first, in which order, given a particular pattern of symptoms, is built from thousands of small cases whose statistical shape the person themselves may not be able to state — they can demonstrate it reliably without being able to derive it. Writing "check the cache first" captures the rule that's usually right; it doesn't capture the hundred prior cases that taught them when the rule is usually wrong, and a successor working from the written rule alone inherits the first without the second.

Strongest Objection: Capture Infrastructure Already Has a Name on This Site FoundationalKnowledge that endures for decades — core principles

ThisThis site's page on byproduct capture already makes the general case that knowledge surviving past the moment it was produced is the exception, not the default, and names four patterns that make capture happen on purpose rather than by luck: durable instrumentation, oral transmission, researcher or staff mobility, and structured reflection. A departing expert's knowledge is a specific, person-shaped instance of exactly that general problem, not a different one — the question this page adds is narrower: which of those four patterns actually works on tacit knowledge specifically, versus knowledge that was always going to be straightforward to write down. Oral transmission and mobility transfer tacit knowledge reasonably well, because both keep a person — not just a document — in the loop. Durable instrumentation and most forms of written documentation transfer the explicit remainder efficiently and transfer almost none of the tacit part, which is exactly backwards from where the risk of losing it actually sits.

Provisional Conclusion FoundationalKnowledge that endures for decades — core principles

AnAn organisation's knowledge has a half-life measured in how long its retrieval paths survive, not in whether the underlying facts are still true somewhere in storage — and the retrieval path for tacit knowledge is disproportionately one specific person, which is exactly why that knowledge's half-life tends to be set by an HR decision rather than by anything about the knowledge itself. Converting tacit knowledge into something more durable is possible, and Nonaka and Takeuchi's whole research programme is about how organisations do it deliberately, but it's neither free nor complete: some real part of what an expert holds is only ever available while the expert is.

Questions for Further Thought

• If the four capture patterns on this site's byproduct-capture page transfer tacit knowledge unevenly, which combination of them would an organisation actually need to run in parallel before someone's planned departure, rather than scrambling to run one of them after?
• Is there a way to tell, before someone leaves, how much of what they hold is genuinely tacit versus simply undocumented but easy to write down quickly?
• Does a successor who inherits only the explicit remainder of someone's knowledge eventually rebuild the tacit part through their own experience, or does an organisation's tacit knowledge actually reset with every departure, however good its documentation?

Related Topics

• Communities of Practice: Learning by Belonging, Not by Being Told — how the knowledge this page is about gets into one person's hands in the first place.
• Waste Heat or Reactor Fuel — the general capture-infrastructure taxonomy this page applies specifically to a departing expert's tacit knowledge.
• How Sleep Turns Ephemeral Memory Into Foundational Knowledge — the storage-strength/retrieval-strength distinction this page borrows and applies at organisational scale.
• Can an Organisation Forget on Purpose? — the companion question of knowledge an organisation chooses to let go of, rather than loses by accident.

Further Reading

• Nonaka, I., & Takeuchi, H. (1995). The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation. Oxford University Press.
• Polanyi, M. (1966). The Tacit Dimension. University of Chicago Press.
• Bjork, R. A., & Bjork, E. L. (1992). A new theory of disuse and an old theory of stimulus fluctuation. In A. F. Healy, S. M. Kosslyn, & R. M. Shiffrin (Eds.), From Learning Processes to Cognitive Processes: Essays in Honor of William K. Estes (Vol. 2, pp. 35–67). Lawrence Erlbaum Associates.

References