Every conversation in a meeting becomes ambient knowledge that lives in the heads of the people who were in the room. Every written decision becomes durable knowledge that lives forever and scales to everyone who joins later. The compounding advantage of writing things down is the single biggest operating leverage a growing company has.
The math is simple and brutal. A decision explained in a meeting reaches the six people present, decays in their memory within weeks, and has to be re-explained to every future hire, one conversation at a time. The same decision written down reaches everyone who ever joins, verbatim, forever, at zero marginal cost. One mode is linear and lossy; the other compounds.
What to write down
Decisions, with their context and the alternatives considered. Why something was built the way it was. Why something was killed. The reasoning behind a pricing change. The plan for next quarter, with the trade-offs that produced it. Almost anything that took thinking is worth writing.
The highest-value artifact is the decision record, and it needs only five lines: what we decided, why now, what we considered instead, what would change our mind, and who owns it. Five lines, two minutes, written by the person who drove the decision, in the same place every time. Companies that keep this habit stop having the "wait, why did we do it this way?" conversation, which in most teams happens weekly and consumes an hour each time.
What you get
New hires onboard themselves. Decisions stop being re-debated every six months. Async work becomes possible. The company stops depending on the founder's memory. The best leaders compound their thinking through what they write; everyone else has to be in the room to receive it.
There is a second-order effect that matters more than any of these: writing exposes weak reasoning before it ships. An argument that survives being written down is an argument; an argument that only works spoken, with momentum and charisma behind it, is a mood. Teams that write make measurably fewer decisions they later reverse, not because the people are smarter but because the medium filters worse ideas out earlier.
The failure modes
Writing culture dies three ways. First, the wiki graveyard: documents written once, filed in a hierarchy nobody browses, never linked from the place where work happens. The fix is to put writing where the work is (the task, the channel, the pull request), not in a parallel library. Second, the polish trap: documents that take three days because they are being written for an imagined audience of critics. The fix is a hard norm that rough-and-shared beats polished-and-private. Third, write-only culture: everything is documented and nothing is read. The fix is structural, not motivational: every meeting that discusses a topic starts by reading the current document, so reading is where the meeting begins.
How to start this week
Do not announce a documentation initiative; they die of ceremony. Instead: pick the next real decision your team makes and write the five-line record. Link it wherever the decision gets discussed. Do it again for the next decision. After three weeks propose the norm in one sentence: decisions get a record before they count as made. The habit spreads by visible usefulness, not by mandate, and the leaders' own writing sets the ceiling: a founder who decides out loud will have a company that decides out loud.
The shape of good writing
Short enough to be read, complete enough to stand alone, dated, searchable. Not perfect. Done and shared beats polished and private every time.