Enterprise knowledge management is a large literature written for organisations with a dedicated function, and most of it does not transfer. Three parts do, and they are the parts that get skipped because they are the least interesting.
Transfers: named ownership of every artefact
The enterprise version has a governance board and an accountability matrix. The small-company version is a name on an article. Both exist to answer the same question, which is who gets asked whether this is still true, and only one of them takes a month to set up.
Transfers: a review cycle with a date
Formal records regimes settle this by statute rather than by preference. Under 44 U.S.C. 3507(g) the Director may not approve a collection of information for a period in excess of 3 years, so a federal form has to come back and prove it is still needed on a clock. A company with no regulator has to set its own clock, and that is the whole of what the enterprise practice contributes here.
Transfers: a taxonomy derived from questions
Enterprise KM calls it information architecture and spends real money on it. The transferable insight is smaller and free: categories should come from the questions people ask, not from the org chart, because readers navigate by problem and not by department.
Does not transfer: the maturity model
Five-level maturity assessments are a way of describing a large organisation to itself and of arguing for next year's budget. At two hundred people the output is a document, the document is not the work, and the exercise consumes exactly the attention that ownership and review dates needed.
Does not transfer: the community of practice
It depends on having enough people doing the same job to form one, and on their having time that is not already committed. Below a certain size it is three people in a channel, which is fine, works, and does not need a name, a charter or a meeting series attached to it.
Transfers: the records schedule, as a habit rather than a regime
Federal agencies keep records under schedules that say how long each series is kept and when it is disposed of, and the National Archives publishes the general ones. The bureaucracy does not transfer; the idea underneath it does. Every article has a lifespan, and most of them have one shorter than the base's. Writing a retirement rule beside the review cadence, unconfirmed for two cycles means archived, gives a base a way to shrink without a meeting, which is the capability a growing base most lacks. The schedule is one sentence in the one-page strategy and one status in the record, and it is the difference between a base that is pruned and one that is abandoned.
Does not transfer: the knowledge platform procurement
The enterprise version selects a platform through a requirements document, a vendor shortlist and a pilot, over a year, and the platform arrives with a programme attached. At two hundred people the base lives in the tool the team already writes in, and the selection question is which of the tools already paid for will hold an owner and a confirmation date per page. That is a week of trials, not a year, and the money the enterprise process would have spent on the platform is better spent on the hours the owners will need to write the first forty articles.
Questions people ask about enterprise knowledge management
Is knowledge management a job at our size?
Not a full one. It is a part of somebody's job, and the part that matters is chasing reviews. Everything else in the discipline is downstream of that.
Do we need a KM strategy document?
Only if somebody will act on it. A one-page answer to who owns what and how often it is reviewed beats a twelve-page strategy, and it is the same content without the framing.
What about capturing tacit knowledge?
It is the real problem and there is no software answer to it. The practical version is noticing when the same question has been answered twice in chat and turning it into an article, which is a habit rather than a system.
What is the smallest KM programme that works?
A one-page strategy naming the bases and their owners, a review cadence with its monthly cost stated, and somebody whose job includes chasing the reviews that fall due. That is the whole programme at two hundred people, and it is more than most companies of two thousand run.