I’ve Managed People for 20 Years — Here Are the 6 Things I Actually Judge

— To my colleagues who are about to lead teams

Hi everyone,

Last week we talked about how to tell whether someone is worth keeping. You said I have a good eye for people and asked for my secret.

Truth is, there's no secret. I've spent over two decades in management, across tech, finance, and a couple of startups, and I've taken enough wrong turns to boil it down to six angles. I'm putting them in writing for you.

This letter won't talk about KPIs, quarterly targets, or technical skills—you know those better than I do. I'm going to talk about one thing only:

How to see, before someone actually fails, whether they're likely to fail down the road.

1. Look first at what they want, not what they can do

In interviews, almost everyone can answer "What are your strengths?"

Very few can honestly answer "What do you truly want?"

They haven't figured it out themselves. And not figuring it out is a problem.

I once managed a brilliant software engineer named Alex. In his first three months, he fixed two legacy bugs that had haunted the team for over a year. Everyone was impressed. Six months in, things turned weird. He started grabbing projects, fighting over resources, and openly shooting down colleagues' proposals in cross‑functional meetings. I sat down with him and asked what was going on. He said, "I want to own something bigger."

I asked, "What does 'bigger' mean to you?"

He went quiet.

Later I pieced it together: he didn't want "bigger responsibility." He wanted "everyone has to report to me." He eventually left, and quietly took a core module's documentation with him. Not out of malice; he genuinely believed, "I wrote it, so it's mine."

I developed a habit since then: when sizing someone up, I don't ask "What can they do?" first. I observe "What do they become after they get what they want?"

• People driven by money are stable as long as you pay well—but don't expect them to stay through downturns.

• People driven by recognition respond to praise—but they'll fold on tough calls.

• People driven by power need you to stay stronger than them, or they'll use you as a stepping stone.

• People driven by actually getting things done are the hardest to find and the most worth keeping.

Skills can be taught. The deep architecture of desire cannot.

So next time you evaluate someone, don't stare at their résumé. Stare at their "after they get what they want."

2. Don't listen to what they say—watch what they choose

In good times, everyone looks like a saint.

When a project is sailing, anyone can talk strategy, claim ownership, and say "we're all in this together." The real filter happens at three moments:

First, how are spoils divided?

I had a sales manager named Kevin who always volunteered to give up his bonus. Later I realized he only gave up the part that "wasn't really his anyway." But when it came to what he was actually entitled to, he gave up nothing and made sure everyone in the room heard exactly how much he'd contributed.

That's not humility. That's calculating.

Second, who steps up when things go wrong?

I made a mistake once: I promoted a highly articulate consultant, let's call him Brian. A major client project went off the rails—delays, complaints, the works. At the post‑mortem, Brian said three things:

"Requirements weren't aligned upfront."
"Resources lagged in the middle."
"Testing was rushed at the end."

Three sentences, three different subjects. He wasn't wrong; everything he said was true. But his first instinct was "not my fault."

Right there I knew he couldn't lead a team. Not because the facts were inaccurate. Because his default move was to deflect ownership.

Third, what do they do when no one's watching?

It's late, after a long day. Do they just shut their laptop and walk out? Or do they turn off the lights, push in their chair, and pre‑load tomorrow's materials onto the conference room screen?

First is a smart person. Second is a reliable one.

The difference: smart people perform for the boss; reliable people execute to their own standard.

Put these three together, and you'll see a person's "behavioral inertia." No amount of eloquence can mask that.

3. Don't judge whether their answer is right—watch how steady they are when challenged

This is a test I've used for years:

In meetings, I deliberately interrupt someone and point out a flaw in their proposal—sometimes a real one, sometimes a manufactured "straw man." Then I watch their reaction.

Three types:

Type one: immediate pushback. Faster speech, louder voice, pulling out spreadsheets to prove they're right.

Smart, but fiercely defensive. From then on, any feedback you give will be met with fight‑first, think‑later—huge communication cost.

Type two: silence, then explanation. "Why did that happen? Well, because at the time…" Their first move is to find external reasons.

These people aren't defending against "I was wrong." They're defending against "I'm imperfect." You don't need to argue right vs. wrong; you need to help them accept that being wrong is okay. But that process is exhausting.

Type three: pause for one or two seconds, then ask: "What would you suggest I change?"

Rarest of all. Their first instinct isn't "me"—it's "the work."

I later learned a term: cognitive bandwidth—how much mental capacity remains when you're under attack.

• People with narrow bandwidth freeze up when challenged, using all their energy to protect their ego.

• People with wide bandwidth treat the challenge as new information and start processing it.

The real danger isn't someone with low ability. It's someone with narrow bandwidth and above‑average ability—smart but unreachable. They'll cause a blowup sooner or later.

4. Don't look at how long they've done something—watch how they approach something they've never done

You may have noticed: when I staff new initiatives, I rarely pick the candidate with the "most relevant experience."

Experience matters. But it expires.

I recall a junior hire named Sarah, less than a year out of college, thrown into a completely unfamiliar project. The company had never done this before, no external benchmarks, no internal playbook.

Her approach was clumsy: week one, she read everything she could find. Week two, she wrote a list of "what I know for sure I don't know." Week three, she started cold‑calling—peers at other firms, upstream vendors, even people in completely unrelated industries that had analogous problems.

Three months later, she was the most knowledgeable person in the company on that business.

She later told me: "I never studied this. But I did study 'how to learn something.'"

Since then, I added one more rule: give someone something completely out of their wheelhouse, and watch their opening move.

• Some react with "I need training"—student mindset.

• Some react with "Who knows this? I'll find them"—networker mindset.

• Some react with "Let me build a minimum viable test and iterate"—founder mindset.

You don't have to pick the smartest—smart people are often afraid of exposing ignorance. Pick the one who isn't afraid of the unfamiliar.

5. Don't look at how brilliant they've been—look at whether they can overthrow themselves

This is the hardest truth in this letter.

I've seen too many people stop growing after 40.

Not because they can't learn. Because they can't delete.

They've accumulated tons of frameworks, playbooks, and war stories that got them this far. But those have turned into armor—heavy and thick, and they snap at anyone who taps on it.

Their real problem is they're too certain they're right.

I had a VP, Robert, with a sterling track record—three consecutive quarters of over‑achievement. In the fourth quarter, the market flipped and his old playbook stopped working. I suggested a pivot. He showed me three decks of slides, all "proving" that his original approach was still valid.

He wasn't lying to me. He was lying to himself.

Because admitting "I'm no longer right" would mean admitting "I've been wrong for the past two years"—a price too high for him to pay.

I replaced him. The new guy, Mark, came in and within his first week tore down the entire previous plan and rebuilt a new framework from scratch. Someone asked, "Aren't you afraid of being wrong?" He said something I've never forgotten:

"It's better to be wrong and fix it, than to be right and have no one listen."

Whether someone can be kept long‑term doesn't depend on how right they've been. It depends on how quickly they can say "I'll change" once they realize they've been wrong.

6. Don't look at whether they obey—look at whether they're stable

Last one—probably the most important.

Many people think leaders like obedient employees. The truth is: leaders like stable employees.

What does stable mean? It doesn't mean you always say yes. It means:

• If I give you a task, I have a reasonable idea of how you'll handle it.

• If there's a tempting offer, I have a reasonable idea of where you'll draw the line.

• If there's a sudden crisis, I have a reasonable idea that you won't run.

• Even when you oppose me, I know whether it's "logic‑based opposition" or "emotion‑based antagonism."

That's predictability.

A core function of management is reducing systemic uncertainty. There's already enough uncertainty in the business itself. When I select people, I'm desperately looking for certainty:

• This person may not be the strongest, but their emotions are stable.

• This person may not be the most creative, but their delivery is stable.

• This person doesn't have to agree with me always, but their way of disagreeing is stable. No sudden betrayals, no backstabbing, no walking off at a critical moment.

I don't dare use someone with extraordinary ability but huge behavioral variance. Because you never know what they'll do next—on a battlefield, that kind of person is more dangerous than the enemy.

Conversely, a predictable person—even if they occasionally make mistakes—I'll keep giving them chances. Because I know where their boundaries are, and the cost of covering for them is calculable.

Finally, a few honest words

The six rules above—I don't guarantee they're correct. There's no formula for reading people.

But from my own experience over the years:

• Rule 1 helps you filter "how far they can go."

• Rules 2 & 3 help you filter "whether they can go with you."

• Rules 4 & 5 help you filter "how fast they can go."

• Rule 6 helps you judge "whether they'll fall apart along the way."

You can use these as references when you start leading teams.

But there's one more thing—more important than all the above:

Don't use these six rules only on others.

Every time you evaluate someone, turn the six rules back on yourself:

What do you want? Will you change after you get it?
How do you divide rewards when it's your turn?
When something goes wrong, what's your first reaction?
When someone challenges you, do you stay steady?
When handed something unfamiliar, does it scare you?
How long has it been since you last said "I was wrong"?
When others work with you, do they feel secure?

These questions are much harder than KPIs.

But those who can answer them don't need others to "read" them—their path will widen on its own.

Onward together.

— A manager who has taken quite a few detours