The dirt path: why nobody documents anything

Every documentation initiative dies the same death: an organisational benefit paid for by an individual cost. The escape route is capturing memory as a by-product of work.

Find the newest building on any university campus and look at the lawn. There's a paved path, curving in a generous arc from the entrance. Beside it, cutting diagonally across the grass, a worn dirt line.

Facilities sees vandalism. Students see a shortcut. Neither party is being irrational; they're optimising different things. The designer optimises for neatness and drainage. The person walking optimises for getting to class.

Every organisation has dirt paths: the ways work actually gets done, worn in by the people doing it. Keep that image in mind, because it explains why your last documentation initiative died.

TL;DR

  • Almost nobody wakes up wanting to document processes, and the reason is economics rather than laziness.

  • Documentation is a future organisational benefit funded by a present individual cost. That's a structurally hard sell.

  • The fix: capture memory as a by-product of work. Create less, verify more.

  • The dirt path always wins. Build on it instead of beside it.

The economics of forgetting

Say it plainly: operational memory is a public good, and effort is a private cost.

The team gets the benefits: faster onboarding, cleaner handovers, fewer repeat mistakes. The individual pays the price: the clicks, the writing, the attention, all now, for someone else's future benefit. Calling that a character flaw misses the point. It is correct personal economics, and no amount of "documentation culture" memos changes it.

The cruel part: the gap widens exactly when it matters most. Stressed teams take the shortest path. The teams feeling this problem most are already drowning. Asking a drowning team to document its processes is asking the crew to catalogue every valve while the ship sinks.

I wrote about this in Why Growing Teams Break at 8–20 Employees, and the diagnosis stands:

"The real problem is not documentation… Your team has no operational memory."

Why the usual fixes fail

Look at the standard remedies and they share one flaw:

  • SOPs are written once, then stale by the second process change. (We invented version control for documents decades ago and still don't use it for the ones that matter.)

  • Wikis fill up. Everything's in Notion, and nobody follows it.

  • Hiring an ops person to own documentation just moves the drowning somewhere else.

Every one of these asks people to create: to produce the perfect artifact in advance, for a hypothetical reader. None of them asks anyone to verify: to confirm what actually happened, once, while it's cheap.

That's the whole trick. Turn creation into verification, and the resistance disappears.

The shift: capture as exhaust

You've already lived through this shift; you just didn't call it knowledge management.

Nobody adopted Slack to build an organisational memory, and yet the record of decisions, exceptions and workarounds appeared anyway, as a by-product. Same with Git: the history of your codebase emerged from the act of writing code. Same with the CRM: the history of your deals emerged from selling.

Memory emerged from the work itself.

So the design principle for operational memory is simple: observe the work, a session at a time, and generate the record from it. The only human act left is a one-tap decision: is this right, or not?

The user's effort changes from Create to Verify. Documentation becomes exhaust. The memory is the product.

One boundary, said out loud because it matters: this isn't surveillance. Capture is session-scoped and started by the person doing the work; it's a record you chose to make, rather than a camera watching you.

What this looks like in practice

Concretely, for a workflow worth keeping:

  • Capture the real browser path, the one with the detour and the exception, rather than the diagram of what it's supposed to be.

  • Screenshots before and after, so the record shows the state of the world instead of your description of it.

  • Exceptions annotated in the moment: a word or a tap, not a paragraph.

  • Versioned, so it stops going stale silently and dies honestly when superseded.

It can look small in practice. In the sim project I wrote about, a team of two hit the same scaling problems as much larger companies, and handled them with the same shape: session cards, evidence-backed handovers, decisions filed while still fresh. The principle is scale-free. (The Human Is The Override)

MimeOps

This is the problem MimeOps exists to solve. MimeOps captures real workflows as they happen, the browser path, the screenshots and the small decisions, and turns them into procedures your team can trust, update and reuse. No more lost context, and no repeated explanations. mimeops.com

Stop asking people to document their work. Start letting their work document itself.


The Recovery Advantage series. Previous: "Recovery metrics: the numbers that actually predict survival". Next: "Govern before you sell: everyone wants your archive" (publish URLs pending).

If your team is growing past the point where everyone just knows things, get in touch.

Written By

Pin It on Pinterest