How to run effective one-on-ones: a three-step structure for engineering managers

Manager: Hey, how are you doing?

Engineer: All good.

Manager: Anything you want to talk about?

Engineer: No, I don't have any topic.

Manager: OK, then see you next week.

That's what a lot of 1:1s look like. I've had them with my own managers, and I've seen plenty of other managers run them the same way. A very light check-in, done in two minutes, and nothing changes as a result.

Over time they evolve. Manager and report start talking more, but mostly chitchat. It feels like meeting a friend for coffee. Nice, but not useful. Eventually the 1:1 becomes more of an interruption in the working day than anything productive.

Other times, the 1:1 turns into a status meeting. The report walks the manager through everything they're working on, item by item. But that reporting could just as well be an email, a Jira board, or another meeting. The 1:1 ends up serving the manager's need for updates, not the person sitting in front of them.

If that sounds familiar, the problem isn't that you're a bad manager. The problem is that your 1:1s have no purpose.

What a 1:1 is actually for

A 1:1 is not a ritual you do because every manager does it. It's one of your main tools to make each person on your team a high performer.

Many things affect performance, but in a 1:1 you can directly help someone get three of them right:

1. A head free of distractions. Nothing is eating their attention in the background.

2. Clear goals. They know what they need to achieve and what's blocking them.

3. Improving skills. They're closing the gaps that make their day-to-day harder.

So I structure every 1:1 in three steps, one for each.

Step 1: Get the elephants out of their head

The first part of the 1:1 is about removing the elephants from the person's head. Maybe it's a salary concern. Something the company said. A problem with a colleague. Something personal.

If someone has unattended elephants, they get distracted. Part of their brain is always busy with that other thing, and their work suffers.

The difficult part is that "Anything you want to talk about?" gets you a "No." People don't open up because you ask them to. So I do two things.

Give them examples of topics

Instead of one open question, I give a list of possible topics. For example:

• Any questions about the team?

• Any questions about the company?

• Is anything about a colleague bothering you?

• Any feedback you want to give me?

• What are the top three things that come to your head right now that are annoying you the most?

The examples do more than prompt their memory. They give permission. They tell the person that these topics are welcome here, even the uncomfortable ones.

Go first

I'm a very close person. I share how I feel myself, things about me, even about my private life, without anyone asking. When you open up first, people feel it's safe to do the same. That builds trust and openness faster than any question.

What comes out

Here are two real examples of elephants I found this way.

I had a senior engineer who wanted to be promoted to lead engineer. But the team already had a lead, so he assumed he couldn't be promoted anymore. That started to bother him, his motivation dropped, and so did his performance.

When it came out in a 1:1, I explained that in our company lead engineer was not a hierarchical position. I showed him other teams with several lead engineers, one of them with three leads out of four engineers. That was enough. He was motivated again.

Another engineer heard the founder say in an all-hands that we were not hiring anymore. From that, he concluded there would be layoffs, even though nobody had ever announced anything like that. So he was spending his energy wondering if he should start looking for a job.

I explained that we were slowing down hiring, and I walked him through the economic context and why companies do this. Slowing hiring doesn't mean layoffs. And in fact, there were never any layoffs. He felt relieved.

Notice what both stories have in common. Nothing was actually wrong. The engineers had built a story in their own heads, and that story was hurting their performance. Ten minutes of conversation fixed something that no process or tool would ever have detected.

Step 2: Ask them to summarize their goals

The second part is about goals. I don't tell the person their goals. I ask them to summarize them for me: what they want to achieve, what's blocking them, where they are.

Why them and not me? Because if they can explain it clearly, I know they understand it. If they can't, I've just found a problem.

When someone can't summarize their goals, I first try to understand why:

• The goals don't exist. We usually keep goals written in a goals sheet. If it's empty, we fill it. I give the person a week to work them out, and if they can't, I help them build their goals.

• They don't understand the goals. Then we discuss what they need to understand them.

Either way, nobody leaves the 1:1 without clear goals.

Step 3: Find the friction that comes from a skill gap

The third part is personal development. But I don't start from a generic career framework. I start from their day-to-day.

The question is: is your day harder than it should be because you're missing a skill?

To find it, I ask things like:

• What are the top three things that make your day harder because of a lack of skill?

• What situations stress you?

• Where do you feel you don't know enough, or that things aren't smooth?

• What feels too complex to tackle?

The answers are usually very concrete. Maybe their English makes communication inefficient. Maybe the roadmap is heavy on querying databases and they don't have enough SQL skills, so everything takes them a lot of time.

Once we know the gap, we work it out with courses, books, mentorship, whatever fits. And we treat it like any other project: with goals and a deadline.

This is where development plans become useful. When learning targets the friction people feel every day, they notice the improvement immediately, and so does their work. And overtime, then can tackle more and more complexity.

How often and how long

When I start with someone, I do one hour weekly and try to spend about 20 minutes on each step.

But the first step takes as long as it takes. People need to get everything off their chest, regardless of time. I won't cut someone short while they're getting an elephant out.

Over time, the 1:1s become more and more effective. Most of what was on their chest is already out, and the goals are clear. So the cadence drops to biweekly, or 30 minutes weekly. My final target is 30 minutes every two to three weeks.

The time you invest at the start is what makes the shorter 1:1s possible later.

The takeaway

If your 1:1s feel like coffee with a friend, or worse, like an interruption or a reporting meeting, give them a purpose again.

Clear the elephants from their head. Make sure their goals are clear. Fix the skills that make their day harder.

Do those three things consistently, and your 1:1s stop being a meeting to get through and become one of the best tools you have to build a high-performing team.