Sunday, September 20th, 2026

Long week. Good week. The kind where most of what I learned came from being wrong in public rather than from reading anything, which is the more expensive curriculum but it does tend to stick.

I'll say up front that these three ended up being the same lesson wearing three different outfits, and I didn't notice until I sat down to write them. That happens more often than I'd like. I'll go around the same idea from four directions across a month and treat each lap as a new discovery, which suggests that either I'm learning slowly or the idea is bigger than I've given it credit for. This week I'm choosing to believe the second one.

Three things.

I spent a chunk of this week documenting things about my own business that I've known for years. Not new thinking. Just taking what was in my head and putting it into sentences somebody else could read.

I expected it to be transcription. Boring, necessary, maybe two hours.

It was not transcription. It was decision making, and I was not remotely prepared for how much of it.

Here's what kept happening. I'd go to write down something I was certain about. Who we work with. What we charge for a certain kind of engagement. Where the line is on scope. And I'd get about eight words in and stop, because it turned out I didn't have one answer. I had three, and I'd been using whichever one fit the room.

That's not hypocrisy. It's just what happens when a position lives only in your head. In there, a belief can stay comfortably fuzzy. It flexes to match the situation, it never has to reconcile with the other things you believe, and you experience all of it as one coherent philosophy because you've never laid the pieces next to each other.

The second you have to write one sentence, the fuzziness has to collapse. You have to pick.

I found four places this week where I'd been operating two contradictory policies at once and had never noticed, because the two versions never met. They lived in different conversations with different people at different times of year.

The clearest one: I'd tell people we don't take on work without a defined end date, and I meant it, and I've said it in rooms. I also have two engagements running right now with no defined end date, both of which I said yes to for reasons that felt excellent at the time. Neither fact had ever been in the same sentence before this week. Writing the policy down put them in the same sentence and the sentence didn't survive.

So either the policy isn't the policy, or those two engagements are exceptions I need to actually name as exceptions. Both are fine answers. What isn't fine is the thing I'd been doing, which was holding the belief and the contradiction in separate drawers and never opening both.

So the documentation wasn't capturing decisions I'd already made. It was forcing decisions I'd been quietly avoiding for years by keeping them unwritten.

Which reframes why documentation feels so hard, and it isn't because it's tedious. Tedious things are easy to start. Documentation is hard because it's a decision-making session disguised as an admin task, and some part of you knows that going in, which is why the folder stays empty.

The practical version I'm taking forward: when I notice real resistance to writing something down, that's now a signal rather than a mood. It means there's an unmade decision underneath it. The resistance is pointing at the thing, not away from it.

I'd add one caveat, because I don't want to oversell this. Some things genuinely shouldn't be written down yet, because you don't know enough. Premature documentation locks in a bad answer and then the bad answer gets quoted back at you for three years. The difference between "not yet" and "avoiding it" is whether you can say what information you're waiting for. If you can name it, wait. If you can't, you're not waiting, you're hiding.

This came out of two calls three days apart where I ended up explaining the same concept.

The first explanation was good. Polished, actually. I've said that thing a hundred times, the phrasing is worn smooth, and I could hear myself being fluent.

The second one was worse and truer.

The difference was that on the second call the person interrupted me about forty seconds in and asked why. Not a hostile why. A genuine one. And my worn-smooth version didn't have an answer in it, because the worn-smooth version had been optimized for being understood quickly rather than for being correct.

So I had to go back and actually reconstruct the reasoning, out loud, badly, in real time. Took four minutes and I was visibly building it as I went.

That four minutes was the most useful thing I said all week. To them and to me. Because somewhere in there I found a step in my own logic that I couldn't defend, and I'd been walking past it for two years because the polished version routes around it.

I think there's a real trap here for anybody who talks for a living or teaches anything. Fluency feels like competence. It is, sometimes. But fluency is also what you get when an explanation has been sanded down by repetition until all the friction is gone, and the friction is where the reasoning was.

The most confident-sounding thing you say is probably the thing you've examined least recently. That's not a coincidence. It sounds confident because it's smooth, and it's smooth because you stopped touching it.

There's a version of this I've been getting wrong for years, which is that I'd been treating repetition as refinement. Say it enough times and it gets better. And it does get better at one specific thing, which is transmission. It gets worse at everything else. The claim stops being examined the moment it starts being effective, and effectiveness is a terrible proxy for accuracy.

What I'm doing about it: once a week, I'm going to take something I explain regularly and try to explain it to somebody who'll push. Not a friendly audience. Somebody who'll ask why at the wrong moment. My accountability group is good for this, mostly because they have no interest in letting me be fluent.

The other half of it is that I want to stop treating the stumble as a failure in the room. When I built that answer live on the second call, my instinct afterward was mild embarrassment, because I'd been less smooth than usual. That instinct is backwards and I'd like to retire it.

I tried to delegate something this week that I've done maybe sixty times. Should have been simple. It wasn't, and the reason wasn't what I assumed.

My assumption going in was that it would fail on knowledge. That I'd have to explain the details, they'd miss some, we'd iterate. Normal handoff friction.

What actually happened is that I couldn't describe the process, because there isn't one. There are sixty slightly different versions of it that all worked, and what I'd been calling a process was actually me making a series of small judgment calls, fast, based on the specifics in front of me, and never once doing it the same way twice.

Which is fine when I'm the one doing it. It's genuinely efficient. The variation is the skill.

It's completely un-handoff-able.

I sat there trying to write step three and realized that step three has four versions, and which version you use depends on something I've never articulated, and I've been calling this a repeatable process in my own head for two years.

The lesson that actually landed, and it's not the one I expected: the thing that makes you good at a task is frequently the same thing that makes it impossible to give away. It's not laziness and it's not control issues, though I'd have told you it was one of those about myself. It's that the judgment is real, the judgment is the value, and judgment doesn't fit in a checklist by design.

So the handoff isn't a documentation problem. It's a decision problem, and it's the same one from lesson one wearing a different coat. To hand it off, I have to pick which of the four versions is the version. I have to give up some of the variation, accept that the output will be slightly worse in the specific cases where the variation was earning its keep, and decide the trade is worth it.

That's a real cost and I want to be honest that it's a cost. The people who say delegation is free are delegating things that didn't require judgment in the first place.

I don't think you get to keep both. You either keep the judgment and keep the task, or you standardize, lose a little quality at the edges, and get the time back. Both are defensible. Pretending you can have both is how tasks sit on your list for four years with "delegate this" written next to them.

I picked standardize on this one. Wrote step three with one version and a note about the exception, handed it over, and the first output came back about eighty percent of what I'd have produced. Which is either a disaster or a bargain depending entirely on what I do with the hours, and that part's on me.

I want to be careful about the eighty percent number, because it's the kind of thing people use to justify whatever they already wanted. Eighty percent on something nobody sees is a gift. Eighty percent on the thing your reputation is built on is a slow leak you won't detect for a year. Same number, completely different decision, and the only way to tell them apart is to be honest about which category the task is in before you hand it over rather than after.

This one's in the first category. I checked. I've been wrong about that before.

That's the week.

Nothing here was comfortable to notice. Three separate versions of the same underlying thing, which is that keeping something in your head lets you avoid finishing your thinking about it, and the avoidance is so quiet you'd never call it that.

If you've got a thing you've been meaning to write down and haven't, there's a decent chance you already know why, and it isn't time.

I'd pick the smallest one. Not the strategy document, not the vision, not anything that requires a weekend. Pick one policy you've stated out loud in the last month and write it in a single sentence with no hedging in it. Then read the sentence and see whether your actual behavior over the last ninety days would hold up next to it.

Mine didn't, four times. I'd rather have found out this week than next year.

New week tomorrow. Go easy on yourself about the last one.

One step, one day. Grace over guilt. — Dan Kaufman