Institutional knowledge as a feature

Companies are always trying to solve the “institutional knowledge” problem. What if we treated it as a feature instead of a bug?

Everyone loves being in on a great secret. It’s juicy, it’s contextual, it’s fun.

What if institutional knowledge worked the same way?

AI has made it obvious that context is everything, and companies like Glean have built entire products around solving the context problem. One of the big use cases is consolidating institutional knowledge before it walks out the door.

But what if institutional knowledge is what makes companies special in the first place?

There are the obvious risks: it doesn’t scale, it creates personnel risk if someone leaves, it creates bottlenecks.

Maybe that’s a feature, not a bug.

Smaller companies move faster on institutional knowledge

When the real reasons behind your success are hard to pin down (baked into instinct and relationships instead of a doc somewhere) competitors can’t copy you even if they try.

You can steal a playbook. You can’t steal five years of shared ~vibes~.

That’s the actual case for tribal knowledge. The stuff that never makes it into a wiki is often the stuff nobody else can reverse-engineer.

The common experience at startups is that early stage is absolute chaos, but absolutely fun. Everyone is sprinting at problems, iterating, and speed is the only advantage you have.

These organizations move fast because you can just iterate over the knowledge one person holds. Let’s say at your five-person startup, one person owns demand gen. By default, that person is also doing product marketing, branding, and creative. The roles collapse into each other because there’s no one else to hand it off to.

The result is faster learning, because that one person can act on what they’re learning in real time instead of routing it through three other teams first.

The magic is that everyone else in the company picks it up too, just by osmosis. You’re working intimately with each other, roles overlap constantly, and to actually get anything off the ground your marketer has to sit in the room with the founders and engineers.

Those collisions are what spread the knowledge, like a virus (shoutout to the last wedding I attended where everyone go the flu).

How does this work at large companies?

The spread of knowledge across an org is the result of organic conversations, collaboration, and culture.

It’s a big part of why companies are pushing return to office, even if nobody says it that plainly.

In a remote setting, how do you build a culture of curiosity and natural collisions?

The knowledge doesn’t disappear at scale. The collisions do. Slack threads aren’t a replacement for a hallway conversation, no matter how many channels you make.

So the fix isn’t more documentation. It’s more collisions.

The key is 1:1 conversations and relationships outside your functional area. That’s where the creativity actually lands.

You want people naturally building relationships and having conversations. You need it. That’s what makes work fun, that’s what makes you stay at a company, and that’s how institutional knowledge actually spreads.

Building natural relationships unfortunately means icebreakers and forced fun.

At LeagueSide, we used Donut to force people into 15-minute coffee chats with someone in a different department once a week. Non-negotiable, no rescheduling.

The goal was to manufacture, on purpose, the kind of friction that happens for free on a five-person team and quietly disappears the second you hit forty. If you’re a stickler for structure, throw in some icebreakers, personal and professional, to get things rolling.

The honest counterpoint

Companies like Stripe and Amazon built real advantages the opposite way: relentless internal documentation, six-page memos, wikis people actually read. That’s not nothing. It gets people ramped faster, it protects you when someone quits, and it stops three teams from independently rediscovering the same mistake.

So this isn’t “documentation bad, vibes good.” It’s that documentation and tribal knowledge do two different jobs.

One protects you from the bus factor. The other protects you from being copied, and gives people a reason to stick around. Most companies over-invest in the first and don’t even realize the second is an asset instead of just a mess to clean up.

This doesn’t work without the right culture

None of this works unless you’ve built the right group of people. The kind who actually want to spend time together.

People assume building culture means hiring a bunch of similar people. It doesn’t. Build a company of people with shared values but wildly different experiences and backgrounds.

When you get that mix right, great conversations happen even when they’re forced.

Secret sauce

Maybe your secret sauce as a company is your secrets. Maybe that’s what makes it special. Maybe the insight was never in the facts. It’s in how people interpret them, argue about them, and pass them along, kind of like a long game of corporate telephone.

Every human group that’s ever existed has run on gossip, because gossip is one of the ways we figure out who’s in and who’s out. Companies aren’t any different.

Institutional knowledge isn’t just information. It’s the thing that makes people feel like insiders.