Why Company Knowledge Becomes Impossible to Find After 20 Employees
Why Company Knowledge Becomes Impossible to Find After 20 Employees
Below ten people, knowledge management is not really a problem. Everyone is in most of the same conversations. The team is small enough that you know who to ask for anything. Information flows through proximity.
Above twenty people, this breaks. And it breaks in ways that are expensive and frustrating, but hard to point to in a spreadsheet.
Where Knowledge Ends Up
By the time a business reaches twenty or thirty people, its operational knowledge is distributed across four or five locations that do not connect.
Some of it is in email threads. An important client requirement was established three years ago in an exchange between two people who are still at the company, buried in a thread that only they know exists. The information is accurate. It is not findable.
Some of it is in people's heads. The operations manager knows why the pricing exception was made for the Hartley account. The finance assistant knows the quirk in how the payroll system handles overtime for part-time staff. The sales director knows the informal agreement with the Birmingham supplier. This knowledge is accurate and current. When those people leave, it leaves with them.
Some of it is in shared drives, in folders that made sense to whoever created them three years ago and are now navigated by habit rather than logic. The document is there. Finding it takes ten minutes.
Some of it has been written down somewhere, a policy document, a process guide, an onboarding note, but the version on the shared drive is not the current one, and nobody is sure what the current one is or where it lives.
The Failure Modes
The most visible failure mode is new staff asking the same questions repeatedly. They cannot find the answer in any document, so they ask a colleague. The colleague answers from memory. This happens three times a week, across multiple new starters, all asking slightly different questions that would all be resolved by the same document if it existed and were findable.
The second failure mode is inconsistent answers given to clients or customers. Two account managers give different answers about the same policy because they each remembered it differently. The client gets confused. The business looks disorganised.
The third failure mode is the knowledge departure. A long-serving employee leaves. In their exit interview, they mention a few things people should know. Two months later, the team discovers that the things they mentioned were one tenth of what was actually in their head. The information is gone.
The fourth failure mode is the manager bottleneck. When nobody can find the answer in a document and the relevant colleague is not available, the question escalates to a manager. The manager becomes the human search engine for the business, answering operational questions that should be self-service because the knowledge infrastructure does not exist to support self-service.
This Is Not an IT Problem
The instinctive response to knowledge management problems is to buy a tool. An intranet, a wiki, a document management system. The tools exist. They are often already in place.
The problem is not tooling. It is that the tools require someone to maintain them, and in a busy SME, maintenance is the first thing that stops when people are under pressure. The wiki is accurate in the first quarter after it is set up. By year two, half the articles are out of date and nobody trusts it.
The maintenance burden is the structural problem. A knowledge system that requires continuous manual curation will decay. A system that surfaces existing documents and keeps them findable without requiring perfect organisation from the team is a different category of tool.
What a Private Knowledge System Changes
A private knowledge system indexes the documents your business already has, policies, guides, SOPs, contracts, email archives, meeting notes, and makes them searchable by meaning rather than filename.
A staff member can ask "what is the policy on client payment terms for professional services projects" and get an accurate, cited answer from the relevant document, without knowing the document exists or where it is stored. The answer comes with a reference so they can read the source if they need the full context.
This changes the new starter experience, the manager bottleneck problem, and the inconsistent-answers problem simultaneously. It does not solve the tribal knowledge that lives entirely in people's heads. But it addresses the large proportion of that knowledge that was written down at some point and simply became unfindable.
If you are wondering how a private knowledge system compares to tools like SharePoint search or a shared drive, the private AI system vs SaaS comparison explains the meaningful differences.
If your business is at the stage where finding the answer to an operational question regularly takes longer than it should, or where you are aware that key knowledge lives in specific people rather than in the business itself, request a system review and I can show you what a practical knowledge system looks like for your scale.