Institutional Knowledge: What Walks Out When People Leave

Every company has a person everyone goes to. Ask them how the renewal works, or why that big account left two years ago, and the answer comes back complete, with the context that no document carries. They have been here a long time. They are, without anyone having decided it, the building’s memory.

Then one day they hand in their notice, and a low panic moves through the team. Not about the headcount. About everything in their head that nobody else can see.

That memory has a name. It is institutional knowledge, the accumulated, hard-won sense of how the place works once you strip away the org chart, and it has a habit of leaving through the front door with the people who hold it. Most teams treat capturing it as the whole task: get it written down before they go, breathe out, done. That is half the job, and the easier half. Our position, and the throughline here: capturing the memory matters, but institutional knowledge only pays off when it reaches the next person at the moment of the work, not when it is filed.

What is institutional knowledge?

It is the part of the map that was never printed. A new hire gets the printed half on day one: the org chart, the pricing sheet, the process doc. What they do not get is the worn footpath the veterans walk, the shortcut that avoids the swamp, the bridge that looks solid and is not. That second map exists, and it is the one that keeps people out of trouble. It is held in heads, drawn by experience, and handed over only by working shoulder to shoulder.

The philosopher Michael Polanyi gave this its sharpest line in 1966: “we can know more than we can tell” (Polanyi, The Tacit Dimension). The person everyone asks knows far more about how the company runs than they could ever dictate into a wiki, which is exactly why the wiki, when someone finally writes it, reads thin next to them. This is the organization’s version of tribal knowledge, which is the unwritten motion of one strong performer. Institutional knowledge is that, scaled across years and many people: the whole building’s memory rather than one rep’s.

The cost is not abstract. Panopto’s workplace knowledge study found that employees lose around 5.3 hours every week waiting on knowledge held by a colleague or rebuilding expertise that already existed somewhere (Panopto, Valuing Workplace Knowledge). That is most of a workday, every week, spent fetching what someone already knew. Older research from the McKinsey Global Institute put the searching burden at about a fifth of the average workweek (McKinsey Global Institute, The Social Economy). The drain runs whether or not anyone resigns. A departure makes the leak visible all at once.

Why does capturing institutional knowledge fail to fix anything?

Because writing it down moves the memory one step, and teams mistake that step for the finish. The knowledge goes from a person’s head into a document, which feels like rescue, and in one sense it is: the answer now survives the person. But a document is not a behavior. The captured memory sits in a folder, the rest of the team goes on working the way they always have, and nothing reaches them while the work is in front of them.

This is the same wall every internal knowledge base runs into. The answer exists, it is findable, and the person on a live task still does not use it, because using it means stopping to go look. Picture an attic packed with everything the family ever owned, carefully boxed and labeled. Nothing is lost. Nothing is reachable either, not in the moment you need it, because reaching it means climbing the ladder mid-task. A captured answer in a folder is an attic box. It survived. It is not in your hands.

The folder was never the obstacle.

Forgetting finishes what friction starts. Even the person who reads the captured answer in a training session sheds most of it within days, not from carelessness but from being human. So the memory sits in a folder, the new hire half-remembers a worse version, and real knowledge retention never happens, because retention is not telling someone once. It is the right answer returning when they need it.

When the answer does not get used, the cause is the system, not the people. Nobody is ignoring the archive out of arrogance; the archive never reached them where the work happens. As we argue in knowledge sharing, the fix is to deliver the answer into the flow of the work, not to scold the team about discipline.

How do you keep institutional knowledge in the building?

Capture the memory from the people who hold it, then return it to everyone at the moment of the work. Think of the retiring ship’s captain. You do not keep what is in his head by handing the next captain a transcript of his career. You keep it by making sure that, when the fog rolls in over the same reef, the warning is already on the chart in front of the new hands. The sequence is concrete, and none of it leans on memory or willpower:

• Capture the memory, not the meeting notes. Watch how the work gets done and name the steps. People cannot fully narrate their own judgment, so observe it rather than interview for it.
• Curate it down to the next move. Strip the captured memory to the single next action a person needs, not a fifty-page binder. One clear step beats a complete manual nobody opens.
• Deliver it in the flow of work. Surface that next step inside HubSpot, Salesforce, the inbox, wherever the work already happens, so acting on the memory costs no detour to a separate destination.
• Inspect whether it gets used. Measure adherence to the captured memory, then coach the gap. You cannot spread a standard you never inspect, and inspection is what turns one veteran’s knowledge into a team habit.

That last step is where most retention efforts die. A memo posted to a wiki and never inspected changes nothing; an answer delivered in the moment and measured for adherence becomes how the team works. The difference is not the quality of the document. It is whether anyone built the path from the archive back into the work.

The payoff shows in the data. Our State of Sales Enablement found that teams whose guidance lives in the flow of the work hit quota at 49 percent against 15 percent for teams whose knowledge sits in a separate destination (The State of Sales Enablement). The knowledge was the same; what changed was whether it reached the person where they stood. When a half-century of tacit-knowledge research and our own field data agree that delivery at the moment of need is the deciding variable, that is the case worth betting on.

What we recommend

There are two ways to treat the memory locked in your most experienced people. You can leave it tacit and hope it rubs off, then absorb the slow ramps and the lost context as a cost of doing business. Or you can capture the hard-won memory, deliver it back to everyone in the flow, and inspect whether it gets used, so the knowledge outlives any one person.

We recommend the second, and not as insurance. The evidence runs one way: expertise that stays in heads is fragile, leaks hours every week, and disappears with the people who hold it, while an answer delivered at the moment of need more than triples quota attainment over knowledge parked in a separate tool. Writing it down was never the finish line. Getting the next person to act on it, while the work is in motion, is.

So find the memory that already runs your best work, capture it from the people who hold it, and put it back in everyone’s hands at the moment they need it. Start with where that captured knowledge should live in the internal knowledge base, the motion of one strong performer in tribal knowledge, and the broader pattern in knowledge sharing.