Documentation Is a Power Problem

People don’t fail to document because they’re busy or undisciplined. They fail to document because institutional knowledge is leverage, and documentation surrenders it. This is the third time¹ I’ve written about incentive structures hiding behind individual behavior. The pattern keeps showing up because organizations keep misdiagnosing it.

The standard framing, and why it’s wrong

Most advice treats documentation as a hygiene problem. Build the habit, pick the right tool, schedule the time. This assumes people agree documentation is valuable and just need a nudge.

But the pattern is too consistent and too universal to be explained by laziness. Engineers, PMs, designers, leads, across industries, across decades, all under-document. That’s not a discipline failure. That’s an incentive structure working exactly as designed.

Sometimes it genuinely is a discipline problem. Someone is overwhelmed, context-switching constantly, and documentation falls off the list. That’s real. But it doesn’t explain why the most senior, most organized, most capable people in an organization are consistently the worst documenters. Discipline explains the junior engineer who forgets to update the README. It doesn’t explain the VP who has never once written down the rationale behind a major platform decision.

Knowledge as positional advantage

The person who holds undocumented context is the person you have to ask. That makes them essential.

The architect who never writes down system rationale. The PM who keeps stakeholder relationships and decision history in their head. The ops lead whose deployment process exists only as muscle memory. These aren’t oversights. They’re leverage preservation strategies, sometimes conscious, often not.

Writing it down means anyone can do what only you could do before.

“Ask Sarah” isn’t documentation. It’s a single point of failure with a Slack handle.

Why senior people document least

Not because they’re busiest. Because their organizational value is partly constructed from being irreplaceable.

The more senior the role, the more the knowledge is contextual, relational, and political. Exactly the kind that’s hardest to extract and most valuable to hoard. A junior engineer’s work is visible in code. A senior leader’s work lives in “I know why we made that call in Q3 and who was in the room.”

Why documentation initiatives die

Every organization has tried: wikis, Notion, Confluence, decision logs, ADRs. They launch with energy and decay within weeks.

The standard explanation is “no one maintains them.” The power explanation is simpler: the people with the most valuable knowledge to document have the most to lose from documenting it. The people with the least valuable knowledge are the ones maintaining the wiki. The incentive gradient runs exactly opposite to what the initiative requires.

Why the “bus factor” argument doesn’t work

Telling someone to document as if they leave tomorrow is asking them to preemptively reduce their own organizational value. No rational actor does this voluntarily unless the incentive structure explicitly rewards transferability. And it almost never does.

Performance reviews reward impact, ownership, being the go-to person. Not making yourself replaceable. The bus-factor framing treats documentation as insurance. But you’re asking people to pay the premium on a policy that only benefits others.

What knowledge-hoarding actually costs the hoarder

The leverage strategy works until it calcifies into a trap.

The moment the hoarder loses negotiating power, the knowledge debt comes due all at once. Role transitions, reorganizations, layoffs, compliance audits. These aren’t hypothetical. They’re recurring organizational events, and in every one of them, the person who hoarded context is suddenly expected to transfer it under pressure, on a timeline they don’t control, with no process in place because they never built one.

The hoarder who looked indispensable in stable times looks like a liability in unstable ones. They can’t move to a new role without a months-long extraction process. They can’t take leave without Slack pings. They get promoted into positions where the old knowledge is irrelevant, but they’ve built no practice of making their new knowledge transferable either. The thing that looked like job security becomes a constraint on their own career mobility.

The irony is precise: hoarding knowledge to stay essential makes you essential to exactly one role, forever.

The organizational design implication

If documentation is a power problem, the fix isn’t better templates or documentation sprints. It’s making knowledge-hoarding visibly costly and knowledge-sharing structurally rewarded.

Start with the mechanisms that directly change incentive gradients:

Tie promotion criteria to team capability, not personal indispensability. This is where most organizations fail, because it requires rewriting what “impact” means. If your promotion rubric rewards “was the critical decision-maker on X,” you’re rewarding hoarding. If it rewards “built the team’s ability to make decisions like X without a single point of failure,” you’re rewarding transferability. The language shift is small. The behavioral shift is enormous.

Make “could someone else run this in your absence?” a standing question in performance reviews. Not as a checkbox. As a genuine evaluation criterion with consequences. The uncomfortable version of this conversation happens when X is a high performer whose manager doesn’t want to create friction. That’s exactly when the question matters most. If you only flag single-point-of-failure risk on low performers, you’re treating it as a performance issue, not a structural one.

Treat repeated “only X knows this” situations as a risk flag, not a compliment to X. This requires leadership to stop rewarding the hero dynamic that makes knowledge-hoarding attractive in the first place.

Fund documentation time the way you fund tech debt: as infrastructure, not charity. If it’s not in the capacity plan, it’s not real. Every organization that tells people to “find time to document” without allocating capacity is signaling that documentation is optional. People read signals, not memos.

None of these mechanisms work in isolation. The incentive structure has to change at multiple points simultaneously, or the rational actor simply optimizes around whichever lever you pulled.

The uncomfortable question

What do you know right now that no one else on your team could reconstruct from your artifacts?

That’s not your value. That’s your organization’s vulnerability, and your own trap.