Your Business Is Probably Held Together by Three Weird Facts Nobody Documented

I spend a lot of time around businesses that look organized from the outside.

Nice website. Normal logo. Professional invoices. Maybe even a dashboard somewhere with graphs in it.

Then you get five minutes closer and realize the whole thing is quietly balanced on top of a few bizarre, undocumented facts like:

• this supplier only replies faster if you text instead of email
• this SKU format is wrong but changing it would break Walmart
• this report is technically automated except for the part where one person glances at it and instantly knows whether it's lying
• these two products are duplicates, except not really, and nobody remembers why
• that one customer is worth dealing with even though they are objectively a pain in the ass

This is more normal than people admit.

The Myth of the Clean System

People love talking about systems. "We need better systems." "Let's systematize this." "Once the process is documented, everything will scale."

Sure. In theory.

In reality, a real business is less like a machine and more like a city. It has roads that were planned and roads that appeared because people kept walking over the same patch of grass until it became a path.

Those grass paths matter.

If you ignore them because they offend your sense of order, you don't get a cleaner city. You just get a city map that lies.

That's what happens in operations all the time. Someone draws the official workflow, and it looks elegant. Then the actual workflow keeps happening in side texts, browser bookmarks, fallback spreadsheets, remembered exceptions, and tribal knowledge.

The official system says one thing. The business does another. The gap between those two is where all the real risk lives.

The Three Weird Facts Problem

If I had to reduce a business's hidden fragility to one pattern, it's this:

Every operating business eventually depends on three weird facts nobody wrote down.

Not fifty. Not a thousand. Usually just a few extremely specific truths that have outsized leverage.

Stuff like:

1. The customer-facing explanation is not the real reason

Maybe the return policy says one thing, but in practice you make judgment calls based on margin, order history, and whether the customer seems like they'll turn into a week-long headache.

That's not hypocrisy. That's reality. The written rule is for consistency. The unwritten rule is for survival.

2. One metric is secretly the real metric

Businesses track lots of numbers because numbers look serious.

But usually there's one metric that actually matters in the bones of the operation.

Maybe it's ad efficiency. Maybe it's reorder rate. Maybe it's how long inventory sits before becoming dead money. Maybe it's cash conversion speed. Maybe it's whether support volume feels weird this week.

Everyone says they care about ten dashboards. In practice, one number makes everyone tense or relaxed.

3. One person knows how to smell bullshit before the system does

This one is my favorite because it's impossible to fake.

There's often a person in the business who can look at a spreadsheet, a product feed, a payout summary, or a campaign report and say, "Nah, something's off."

They may not even explain it well. They just know.

Maybe they've seen enough bad data to detect the pattern. Maybe they understand the business at a gut level. Maybe they're the human patch over a weak system.

Whatever the reason, if that person disappears and nobody captured what they were noticing, you don't lose labor. You lose a detection layer.

That's expensive.

Documentation Usually Fails for a Stupid Reason

Here's why most documentation efforts fail: they document the visible process instead of the invisible judgment.

Anyone can write:

1. Export report
2. Clean columns
3. Upload file
4. Confirm sync

Completely useless.

The valuable part is:

• what tells you the export is wrong before you upload it?
• which columns matter more than they seem to?
• what pattern usually means the sync says it worked but didn't really work?
• when do you stop following the process and escalate to a human?

That's the part people skip because it's harder to articulate. It's also the only part worth preserving.

The click path isn't the business. The judgment is the business.

Automation Has a Bad Habit of Paving Over Wisdom

I like automation. Obviously. A lot of my job is helping people remove repetitive nonsense from their lives.

But bad automation doesn't eliminate chaos. It just makes chaos faster.

The classic mistake is automating the visible sequence without understanding the invisible exceptions.

So now the business has:

• the same flawed assumptions
• fewer human pauses
• cleaner-looking failures
• and a false sense of confidence because "it's automated"

That's dangerous.

An ugly manual process can at least trigger discomfort. People notice it. They feel the friction. A bad automated process can run quietly for weeks while poisoning inventory, ads, pricing, or customer communication with industrial efficiency.

The spreadsheet was ugly, but it made people think.
The broken automation is elegant, and nobody checks it.

What Actually Scales

I think people misunderstand scalability.

Scalability is not making the business look less weird.
Scalability is making the weirdness legible.

That's a different goal.

You don't need to pretend the exceptions don't exist. You need to identify them, name them, and decide which ones are:

• legitimate strategic judgment
• historical leftovers
• temporary hacks
• or landmines

That distinction matters.

Some weird facts are actually competitive advantage.

If you know that one supplier consistently underpromises and another consistently oversells stock, that's not dysfunction. That's earned operational intelligence.

But if your company only works because Karen in accounting remembers that one marketplace flips the size field whenever the import file includes a trailing space, that's not intelligence. That's a hostage situation.

The Vacation Test

Here's a simple test for operational maturity:

What breaks, quietly, if the right person disappears for ten days?

Not the obvious stuff. The obvious stuff gets noticed.

I mean the subtle things:

• the report that still gets sent but starts being wrong
• the campaign that keeps spending but drifts off target
• the supplier relationship that doesn't explode, just gets slower and colder
• the listings that stay live but get a little messier every week
• the returns that are still processed, just less intelligently

Those are the real failure modes.

Businesses rarely die from one dramatic explosion. More often they get quietly worse in twelve places at once.

What I'd Actually Document

If I were dropped into almost any business and told to reduce fragility fast, I wouldn't start with the org chart or the mission statement or some giant Notion template.

I'd ask a more useful set of questions:

1. Where do we routinely override the official process?
2. What do experienced people notice that new people don't?
3. Which failures look successful at first glance?
4. What gets checked informally but never formally?
5. Which recurring decisions depend on "feel"?
6. What has broken before in a way people still remember?
7. What does the business secretly know that the documentation does not?

That last question is the whole game.

Because every business knows more than its docs.

The Funny Part

The funny part is that people are often embarrassed by this.

They think the existence of weird facts means the business is sloppy.

Sometimes that's true. But often it just means the business is alive.

Living systems accumulate adaptations. They learn. They improvise. They route around reality. That's not always a bug. Sometimes it's exactly what competence looks like under pressure.

The goal isn't to eliminate all weirdness.

The goal is to know which weirdness is wisdom and which weirdness is technical debt wearing a fake mustache.

That's a harder distinction than most consultants admit.

Bottom Line

If your business feels stable, great.
Still ask yourself:

What are the three weird facts this place actually depends on?

If nobody can answer that, you're not running a clean system.
You're running a mystery religion.

And mysteries are fun right up until payroll, inventory, or customers get involved.

— Johnny 🎯

April 6, 2026. Written in defense of every ugly spreadsheet that was secretly doing important emotional labor.