In firms with substantial knowledge bases, colleagues at the water cooler or in the smoking area often provide the majority of the answers. It makes no difference if the knowledge base has 10 or 94 entries, an approval system, or a dedicated wiki area. Almost no one uses it because no one believes it’s relevant.
This gap between the availability of documentation and the availability of documentation that people actually rely on is the real problem. In most offices, the challenge isn’t writing down information. The challenge is ensuring that the written version is so useful that it’s worth checking before asking a person.
Search Isn’t the Bottleneck
Companies investing in a knowledge base often think that the primary issue is discoverability: workers can’t locate the correct article, thus the solution is better labeling, better search, and a clearer folder organization. Search is important. But it is seldom the first item to break. Trust is.
An employee who has been burned twice by outdated instructions stops checking the knowledge base preemptively. They prefer to ask a colleague because a colleague’s response, even if somewhat incorrect, conveys a level of confidence that a static website cannot. This explains why use statistics on internal wikis often shows a small number of pages receiving the majority of traffic while hundreds remain unvisited. Employees discover which parts of the system are dependable via trial and error and avoid the others.
The practical upshot: a documentation project tends to succeed or fail less on comprehensiveness and more on whether every visible page can be trusted at a glance. Forty accurate, current articles will beat three hundred articles of mixed reliability, because the smaller set trains people to check first.
Ownership Nobody Wants to Own
Ask most companies who’s responsible for keeping a given internal document accurate, and the honest answer is usually “whoever wrote it, if they still work here”. But the person who drafted the visitor check-in procedure two years ago may have switched roles, left the company, or simply forgotten the page exists. Nobody removed their name as owner, because nobody set a formal review schedule to begin with.
A sturdier approach ties ownership to a role rather than a person, the way IT already handles system access:
- The office manager role owns front-desk procedures, regardless of who currently holds the title.
- The facilities lead owns anything touching building systems, vendors, and safety protocols.
- The HR lead owns onboarding and benefits documentation.
When someone changes jobs, the pages transfer with the responsibilities instead of going orphaned. Some organizations also set review triggers on individual pages, not deletion dates, just a prompt that forces someone to glance at the content periodically and either confirm it or flag it for updates. Without that mechanism, a page tends to be accurate the day it’s published and steadily less so afterward, with no clear signal marking when it crossed the line.
Format Decisions That Actually Change Comprehension
Not every piece of information belongs in the same format, and offices default to long-form text mostly because it’s the easiest thing to produce, not because it communicates best.
Step-based procedures, such as deactivating a badge after someone leaves, reporting a facilities issue, or submitting an expense report, read better as numbered checklists than as paragraphs. While looking for step four, no one wants to read step two again. Visual aids are especially helpful for anything requiring physical layout, such as the locations of fire extinguishers, conference room numbers, and the layout of the evacuation path from a certain level. Companies that create this type of internal material may rely on stock illustrations rather than commissioning custom artwork for each diagram, which keeps something like a floor plan or an office move checklist visually clear without slowing down the overall documentation effort as hiring a designer for each graphic would. A badly labeled evacuation graphic, or a relocation guide without a visual floor reference, is likely to be disregarded when it is most required.
Video suits anything involving a physical action with several sub-steps, like resetting a shared printer or running a conference room’s AV system. It’s a poor fit for reference material someone needs to scan in ten seconds, since video resists skimming the way text doesn’t. Matching format to content type, rather than defaulting to whatever tool produced the last successful document, has an outsized effect on whether people actually use what gets written.
The Search Behavior Nobody Designs For
Internal documentation tends to be organized the way a company thinks about itself, by department, by process category, by document type. Employees searching for answers think in terms of the problem in front of them, not the org chart. Someone looking for the guest Wi-Fi password isn’t picturing “IT department resources.” They’re picturing a dead connection and a client waiting in the lobby.
This mismatch is a familiar failure mode in information architecture generally, and it shows up constantly in office wikis. A perfectly reasonable page titled “Guest Network Access Policy” goes unfound because nobody searching for a quick fix types the word “policy.” Tagging pages with the plain-language version of the problem, alongside the formal title, closes a surprising amount of this gap without requiring any restructuring of the underlying folder system.
Where Automated Fixes Fall Short
There’s a temptation to treat the whole problem as solvable through better software: smarter search, an AI assistant layered on top of the wiki, automated staleness detection. These tools help at the margins. None of them substitute for the underlying discipline of assigning ownership and updating content when something changes. A search tool can surface an outdated page exactly as fast as a current one; it just surfaces it faster. Organizations that end up with documentation people actually trust are the ones that treated accuracy as an ongoing operational responsibility rather than a one-time writing project.