Institutional knowledge vs tribal knowledge, with examples
Institutional knowledge is everything an organisation has accumulated about how it works, including what is written down. Tribal knowledge is the subset that lives only in a group's shared experience, with no record anywhere. The second is a subset of the first, it is the fragile part, and it is the part worth finding first. The term tribal knowledge is also contested, and there are neutral alternatives worth using.
The short answer
Institutional knowledge is the whole body of know-how an organisation has built up over time: how things are done, why decisions were made, who knows what, what has been tried before. Some of it is documented and some of it is not.
Tribal knowledge is the part that exists only inside a group's shared experience, with no record anywhere. It is passed on by working alongside people, and it disappears when they go.
So they are not two competing categories. Tribal knowledge is the fragile subset of institutional knowledge, and that is the whole practical point of keeping the terms apart.
Examples of each
Institutional knowledge, documented: the refund policy, the architecture decision record explaining why the team chose one queue over another, the runbook for month end close, the contract that sets out a client's non-standard terms.
Tribal knowledge, undocumented: that the refund policy is applied differently for accounts acquired in a particular migration, and everyone on the team knows it. That the queue decision was made with an assumption about volume that stopped being true last year. That step four of the month end runbook has to be done before step three, or the reconciliation fails, and the runbook has never been corrected. That one client's non-standard terms exist because of a specific incident in 2023, and renegotiating them would reopen it.
Read those two lists next to each other and the pattern is clear. The documented items describe what is done. The undocumented ones explain why it bends, and the exceptions are where the real operational risk sits.
MindKeepr captures what your team knows and keeps it usable, even after people leave.
Why the distinction changes what you do
If you treat the whole of institutional knowledge as one problem, the obvious response is a documentation programme, and documentation programmes have a poor record. They produce a large volume of material about processes that were already reasonably well understood, because those are the easiest things to write.
If you separate out the tribal portion, the work becomes much narrower and much more valuable. You are no longer trying to write down what the company knows. You are trying to find the specific points where exactly one person holds something, and fix those.
That list is almost always shorter than people expect and more alarming than they expect, which is a useful combination for getting it acted on.
Is tribal knowledge an offensive term?
It is contested, and the question comes up often enough to be worth answering plainly. The phrase borrows from the language of indigenous peoples to mean something informal and unwritten, and a number of organisations have moved away from it for that reason. Others use it without a second thought because it has been standard vocabulary in manufacturing and engineering for decades.
If your company has a style guide, follow it. If it does not, undocumented knowledge, team knowledge, or experiential knowledge all carry the same meaning without the association, and nobody will be confused by them.
The label does not change the problem. Whichever phrase you use, you are describing knowledge with no record and no second holder.
How to find the tribal portion
Asking people to list what only they know does not work, because the knowledge they have stopped noticing is invisible from the inside. The answer has to come from the record rather than from the person.
The practical method is to start from decisions instead of documents. Take a role, list the decisions it owns over the last year, and for each one ask who else could confirm that the reasoning still holds. Where the answer is one name, that is tribal knowledge with a dependency attached. Where it is nobody, it is tribal knowledge with an exposure attached.
This is the problem MindKeepr addresses directly. It reads the work a team has already produced, identifies the decisions and context that only one person has ever touched, and routes a short question on each to a named colleague with a deadline, so the undocumented part becomes documented and confirmed rather than merely discussed.
Other terms you will meet
Tacit knowledge is a related but distinct idea: expertise someone applies without being able to fully explain it, such as an engineer sensing which service is at fault before the dashboard says so. Tribal knowledge can be written down once somebody identifies it. Tacit knowledge resists writing by its nature, and transfers by working alongside people.
Corporate amnesia describes the result of losing either: an organisation repeatedly relearning things it already knew. Institutional memory is used loosely as a synonym for institutional knowledge, usually with more emphasis on history and precedent.
The terms overlap and precision matters less than the action they point to, which is the same in every case: find where the knowledge has one holder, and give it a second one.
- ✓Tribal knowledge is a subset of institutional knowledge, specifically the undocumented part.
- ✓Institutional knowledge survives a departure if it was recorded. Tribal knowledge does not.
- ✓Examples are the fastest way to tell them apart: a policy document versus the reason the policy bends.
- ✓Some organisations avoid the phrase tribal knowledge; undocumented or team knowledge works as well.
- ✓Find the tribal portion by looking for decisions only one person has ever touched.
FAQ
Institutional knowledge is the entire body of know-how an organisation has accumulated, including everything documented. Tribal knowledge is the subset that exists only in a group's shared experience with no written record. Tribal knowledge is therefore a fragile part of institutional knowledge, and the part most likely to be lost when people leave.
A documented refund policy, an architecture decision record explaining why one technology was chosen over another, a month end close runbook, or a contract setting out a client's non-standard terms. Each describes how the organisation works and survives a departure because it was recorded.
Knowing that step four of a runbook must be done before step three or the reconciliation fails, and that the runbook has never been corrected. Or that a refund policy is applied differently to accounts from an old migration. The common feature is that everybody on the team knows it and nothing anywhere records it.
It is contested rather than universally considered offensive. The phrase borrows from the language of indigenous peoples, and some organisations avoid it for that reason while others treat it as long-standing engineering vocabulary. Undocumented knowledge, team knowledge, and experiential knowledge all carry the same meaning if you prefer a neutral alternative.
Institutional memory is the closest common synonym, usually with more emphasis on history and precedent. Organisational knowledge and corporate knowledge are also used. For the undocumented subset specifically, undocumented knowledge or team knowledge are the usual neutral alternatives to tribal knowledge.
Start from decisions rather than documents. List the decisions a role has owned over the past year and ask, for each one, who else could confirm the reasoning still holds. Items with one holder or none are the tribal knowledge worth capturing first, and each should be confirmed by a second named person rather than simply written down.
Start free in minutes, or get a demo on your own tools and team.
Faizan Khan is the co-founder and COO of MindKeepr, the operating memory for governed AI. He has twelve-plus years across enterprise IT and digital marketing and is also the founder and CEO of Cubitrek. At MindKeepr he leads growth, go-to-market, and customer experience.