Hiring for the tenth year, not the first.

The people who've been with us longest don't have much in common on paper. Different cities, different schooling, different reasons they ended up in a kitchen or on a floor. What they share is what you can't interview for: a tolerance for small repeated honesties, a sense of humor about being tired, the ability to say my mistake without it costing them anything internally.

You cannot screen for any of that in forty-five minutes. I used to try. I got pretty good at the performance of trying. Then I noticed that the hires I felt best about walking out of an interview were not always the ones still here a year later, and the hires I felt least sure about were sometimes the people I'd now follow into a fire.

So we stopped pretending the interview was the test. We made the whole thing the test, and we built it slower.

Slow interviews.

Three conversations, minimum. Days apart. Different people each time.

The first conversation is a performance, on both sides. Everyone is their best self for forty-five minutes; that's what those forty-five minutes are for. The second is where something cracks. A contradiction with what they said the first time, a question they hadn't expected, a silence that isn't comfortable. By the third, the person we're talking to has started to resemble the person who would actually show up to shifts.

We also watch whether they ask us hard questions. Someone who's been in hospitality for five years and has no sharp questions about tips, scheduling, sick pay, or how we handle a bad guest is either very easy to please or not paying close attention. I'm not sure which of those is worse.

None of this is a trick. We tell candidates the process is slow and why. The ones who are insulted by it usually do us a favor by leaving on their own.

Generous trials.

Paid. Real shifts. Two or three of them, spaced out. Long enough that the person can't keep performing. Performing is exhausting, and a good trial goes on just past the point where the candidate has to stop.

Interviews ask what you'd do. Trials show what you do. Those are not the same information.

I have hired people who could talk beautifully about service and then froze the first time a four-top sent a risotto back. I have also hired people who were awkward across a table and turned out to run a section like they were born in it. A trial tells you which of those you've got.

What we're actually looking for isn't competence. It's how competence is built. Does this person notice when something goes wrong? Do they name the mistake or hide it? Do they ask for help before or after it becomes a problem? Do they get curious when something surprises them, or defensive? Those habits predict more than skill does, and skill is teachable anyway.

And if it doesn't work out, theirs or ours, we pay for the shifts and part cleanly. A trial that reveals a bad fit is not a failure. It's the whole point of running one.

Honest pay.

Post the number. Pay what you said. Raise it when the work gets harder or the person gets better, without making them ask.

The pitch for honest pay is usually an ethical one, and that's fine, but the practical reason is different and maybe more persuasive. When pay is murky, every other signal gets murky too. People can't tell if you actually value them. You can't tell if they're staying because they want to or because they're stuck. Somebody who's underpaid and resentful looks, from management's angle, a lot like somebody who's comfortable and checked out. You lose the ability to read your own team.

Pay people fairly and the readings clear up. Then the real conversations can finally happen. About the work, about the hours, about where someone wants to go next.

No managers we wouldn't want to work for.

This one is the hardest to hold, because the cost of breaking it is invisible for a long time.

A mediocre line cook affects one station. A mediocre manager affects everyone under them, every shift, for as long as they're there. And not in the ways that show up on a P&L. In the way a twenty-two-year-old decides, after six months, that this industry is full of the same people and maybe it's time to try something else.

The question about every promotion isn't whether this person is ready to manage. It's whether I would take a job reporting to them.

If the honest answer is "mostly" or "probably," that's a no. If I wouldn't work for them, I don't get to ask a line cook who needs the job to.

We promote slowly for this reason. We talk to the people under managers more than we talk to the managers themselves. We make it easy to raise a concern, and boring to raise one. No heroics required. And when we've been wrong about a manager, and we have been, we move. Not cruelly. But we don't sit on it. The cost of sitting on it is paid by people who aren't in the room when we decide.

The tenth year isn't a plan. It's a byproduct.

When interviews are slow, the people who need to hide don't make it through. When trials are generous, fit is decided by work instead of by impression. When pay is honest, nobody stays for a bad reason or leaves for one we could have fixed. When the managers are people we'd want to work for ourselves, staying doesn't require looking the other way.

We're not trying to pick the people who'll still be here in ten years. We can't. Nobody can. We're trying to build the conditions under which, ten years in, staying still makes sense — and under which leaving, when it happens, is a real choice instead of an escape.

That's the whole playbook. We'll keep finding out where it's wrong.