You Wrote the SOPs and Nobody Uses Them

The folder nobody opens

You wrote them.

Clean headings. Screenshots. A table of contents. A folder on the shared drive that everybody has access to and nobody opens.

You know how I know? Because the same question still lands in your inbox on a Tuesday afternoon. Because the same three things still come back wrong. Because when you ask why, somebody says some version of "I didn't know that was the process," and you want to scream, because you wrote the process, you sent the link, and you asked them, nicely, to read it.

Before you write version two, look at the thing you're actually looking at.

The SOP is not the process. The SOP is a receipt. It's proof that you want a company that runs without you. It's also proof you haven't built one yet.

Why nobody uses them

Your people are not lazy. They're efficient. They're doing what any sane person does: taking the shortest path to a right answer.

And in your company, the shortest path to a right answer is you. Ask. He'll know. He'll fix it. He'll tell you the exception. Reading the document and hoping you interpreted it the way he would is the long way around, so nobody takes it.

Every SOP encodes an org chart whether you wrote one down or not. And if the honest answer to "who owns this thinking" is "whoever asks me," then the document is decoration.

I'll go first, because I have to.

I have watched my own team route around my own process, and the reason was not their discipline. It was the culture I built. I skip the step when I'm in a hurry, and they've watched me do it. A company's chronic weakness is usually the owner's personal weakness at scale. So if you want this to change, the first move is not another rule for them. It's you saying the true thing out loud: I skip it too, and that's on me.

Then you still have to build something. And this is where most owners make the exact mistake I made. They decide to review harder.

For months, every feature I was building with Claude Code got a written plan first. I read every one of those plans myself and corrected it before I let it run. That worked right up until it didn't. The plans got complex enough that I could no longer understand what they were saying, and even the ones I did catch and correct still executed with a pile of errors it had to go back and fix.

So I stopped trying to read harder and built a reviewer: a second agent whose only job is to assume every line is wrong and go prove it against the live database. It cut the errors by almost 70%.

That reviewer invents nothing. It doesn't trust the plan, it doesn't trust me, it checks. Which is the part your SOP is missing. Right now your people hand you work, you catch the errors, you send it back, and you call that quality. It isn't quality. It's you standing in for a system that doesn't exist, and it will not scale past your calendar.

What actually makes a company independent

Organizational Independence: building a company that can think without you.

Three questions, in sequence, because you cannot run them out of order.

1. Alignment. Do we know where we're going and what we own? A clear and compelling future. An org chart that says who owns what thinking. The Big 3 that actually matter right now, because everything can't be the priority.

2. Movement. Can they decide and act without me? Values that guide decisions when you're not in the room. Simple frameworks for how to think and act. A culture where the team improves together instead of waiting to be corrected.

3. Measurement. Can we tell if we were right? Start past the end. Define what success looks like, then work backward to what has to be true. And everyone has a KPI, so the signals are visible without you asking.

Then a learning loop brings what the work taught you back into alignment. All of it lives in the Brain, the shared place your company thinks from. All of it stands on trust, because if your people believe you'll overrule them the moment they decide something, none of those three questions get answered honestly.

Now rank it, because you don't have unlimited hours.

If you fix one thing, fix alignment. Movement is the slowest and it's the one that makes an SOP stick. Measurement is the cheapest and it's the one most owners skip completely. And the document itself, the writing, the formatting, the table of contents you're proud of, is the smallest part of this. That's why rewriting it changed nothing.

Alignment: the SOP is downstream of ownership

For every SOP you have, answer one question: whose thinking does this replace? Not who wrote it. Whose judgment is written down in it.

If the answer is "mine, but I still want to approve it," then you haven't delegated thinking. You've delegated typing. And typing is the first thing a busy team drops.

Then look at the Big 3. If your team has more priorities than it can count, the document loses to the week every single time. Not because they don't respect it. Because they're triaging, like you taught them to.

Movement: can they decide without you

When I first connected Athena to Slack for my own use, Stephanie found her there and started issuing her commands. Athena would not answer her. Athena was built for one person and refused everybody else, and for a while that was a running joke about Stephanie not being able to get a word out of her.

The joke is what raised the real question. What if the staff could get access?

That question turned into the painful, still unfinished work of opening Athena to the team responsibly. And it's the reason my staff now bring their questions to her instead of to me.

The refusal was never a bug in Stephanie's behavior. It was a design decision I made, and it cost me every question I answered after that.

So here's the move. When somebody keeps failing to get what they want out of your internal tool, stop explaining the limitation and ask what it would take to give it to them.

Now swap internal tool for SOP.

Your people keep failing to use the document. Stop explaining that the document exists. Start asking what it would take for the document to be the answer they reach for.

Movement is where that answer lives. A framework is just a simple way to think that a person can run without checking with you first. If every judgment call inside your SOP still needs your judgment, you didn't write a process. You wrote a wish.

Measurement: can we tell if we were right

Start past the end. Decide what a good outcome looks like, then work backward to what has to be true for it to happen. And give everyone a KPI, including the person who owns the process.

Here's the reviewer, applied to a procedure. Don't ask "did you read the SOP." Nobody answers that honestly. Ask: take the last several times this work was actually done, where did it diverge from the document, and what did that divergence cost? A checker that assumes the drift is real and goes and proves where. Not compliance theater. Evidence.

Which means the SOP needs a home where it can be checked. Someone owns the number. Someone gets to say "this step is being skipped" and bring the evidence with it, without it being an accusation. That's how you find out whether your problem is the document, the training, or the fact that the real process is different and better than the one you wrote.

The learning loop

Every SOP that survives contact with the work gets changed by the work. Someone owns that change, and it happens on a cadence, not on the day you get angry about it. And the loop has to return all the way into alignment, because if the same step keeps breaking in the same place, the problem may be who owns the decision, not how the document is written.

What this is going to cost you

I'm not going to hand you a process and pretend it's free.

The reason you haven't done this isn't that you don't know how. It's that it costs you something real, four separate times, and each one is a feeling you have to be willing to have before you'll move. Read these and find the one that made your chest tighten. That's your real blocker.

Alignment costs you the chair. The chair is being the approver. Being the exception. Being the person who catches it before it ships. That is not a task on your calendar. It's a position, and it's the one that tells you that you matter. Which is why I keep coming back to what I learned in my own build: my reading was never the quality gate I thought it was. It was the feeling of control. That sentence is easy to read and expensive to accept, because the feeling of control is exactly what you'd be handing over. If you can't say it out loud, you'll build the org chart, put your name on top of it anyway, and be right back in the same inbox in six months.

Movement costs you the joke. When Athena refused Stephanie, I laughed, because laughing was easier than asking why she couldn't get in. The refusal was a design decision I made. Not Stephanie's behavior. And fixing it wasn't a setting I flipped. It was painful, it is still unfinished, and it stays unfinished because opening the door tells you the truth about your team: how much of their decision-making was genuinely delegated, and how much was them checking with me first. Once people can decide without you, you don't get to reach back in when you don't like the call. That's the deal, and it's the part that stings.

Measurement costs you being checked. Every KPI you hand to someone else is a mirror pointed back at you. The process owner gets to stand up and say the step is being skipped, and if it is, that report is about the system, and the system includes you. That only survives if you don't flinch. The first number that makes you look like the bottleneck is the test. Let it stand, or you've just taught everyone that measurement is a weapon you aim at other people.

The learning loop costs you ownership. A document that changes is a document whose ownership can change. You cannot tell your team to improve the process on a cadence and then be surprised when the improvement says the decision belongs to someone else. The loop doesn't just edit a page. It moves authority, and it puts that move in writing, which is why people quietly stop running the loop the moment it starts to work.

Ranking these honestly: the alignment cost is the one that stops most owners, and the learning loop cost is the one nobody sees coming until it's already turning.

So pick the smallest of the four and do it this week. Take one SOP. Answer the ownership question out loud, whose judgment is written in this thing. Then hand the check to someone who isn't you, and let them bring you what they find, and do not fix it yourself. That last part is the whole cost, in one move.

And if you're not ready for that yet, don't rewrite the SOP. Do the harder, freer thing first. Tell your team, out loud, that you skip it too. That's the move that starts all of this, and it's the one most owners never make.