Is it a people problem or a process problem?

You've got someone who isn't working out.

Maybe they're slow. Maybe the work comes back wrong and you fix it yourself. Maybe you've had the conversation twice and nothing changed after either one.

So you're weighing it up. Replace them, or accept it. Most owners spend months here, and a lot of them replace the person, and then watch the exact same thing happen to the next one.

That's the expensive version of getting this wrong.

One test

Take the person out of the job in your head. Drop in someone genuinely good. Not perfect, just solid. Somebody you'd hire again.

Does the problem go away?

If it does, you've got a people problem and you should move. If it doesn't, or if you can't actually answer it, the job is broken and the person is just the one standing in it.

Most owners can't answer it, because they've never seen the job done by anyone else. That's not a failure of judgment. You've got one sample.

What it looks like from the outside

Put someone good in the job. The right-hand column doesn't change.

Notice the right-hand column. Every item there survives a replacement. New person, same outcome, because the outcome was built into the job before anybody filled it.

The left column doesn't survive it. Put someone good in and those problems leave with the old person.

What the data says about which one it usually is

For the Platform One Index we ran owner-led companies through the same twenty questions.

46 percent named people as the single thing capping the business, more than marketing and finance put together. Then we scored all five areas separately and ignored what they'd named. People came out worst, at 32.9 out of 100.

So owners point at people constantly, and the people area really does measure worst. Both of those are true.

Here's the part that isn't obvious. "People" in that data almost never means "I hired badly." Read what they actually wrote. A siding contractor between one and five million:

"Build a management system that gives me clear visibility into profitability, performance and accountability without depending on me to personally oversee everything."

That's filed under people. It's a process problem, described precisely, by someone who then put it in the wrong bucket.

A landscaping company, same revenue range, also filed under people:

"I would like to have 6 to 12 months of costs in the bank instead of operating from month to month."

That's cash. It's in the people bucket because the owner feels it as pressure on him, and he is a person, so it goes under people.

Why the two get confused

Because a process problem shows up wearing a person's face.

The handoff between sales and install has no owner. So things fall. The person who most often drops one starts to look careless. You start managing the person, harder, and the person gets worse because being watched makes anyone worse, and now you're certain it's a people problem.

It was a handoff with no owner. It still is. You've just got a demoralized employee on top of it now.

The size it changes at

The mix shifts as you grow, and it shifts the same way for nearly everyone.

Under $250K almost every problem is a people problem, and the person is usually you. There's nobody to hand anything to, so there's no process to be broken.

By five to twenty million it inverts. People drops to 24 percent while operations goes from basically nothing to 15 percent. Same owners, same kind of business, different constraint, because there's finally enough machine for the machine to be the problem.

If you're in that upper range and still certain it's your people, you're running the diagnosis you learned when the company was small.

How to tell for real

The test earlier is a good one and you should use it. It's still a guess.

The measurement is this. Everyone in the business answers the same questions, separately, with no names attached. Then we put your answers next to theirs.

A people problem and a process problem look completely different in that data. If it's people, your account and theirs mostly agree and the scores cluster around specific functions. If it's process, the two accounts split hard on how work moves, and they split in the same place regardless of who you ask.

You can't see that from inside. Not because you're not sharp enough. You've only ever watched this company run one way.

Before you replace anyone

Twenty questions. Ten minutes. Free.

You'll get scored across five areas and see which one your answers point at. If it comes back people, at least you'll be firing on data instead of on the third bad week in a row.