People & Leadership Intermediate 4 min read

One-on-Ones: The 30-Minute Habit That Holds a Team

How to run useful one-on-ones with engineers: cadence, agenda template, what to avoid, and the topic rotation that catches problems before they hit retros.

Table of contents
  1. When does the 1-on-1 cadence actually start mattering?
  2. What is the cost of skipping or doing 1-on-1s poorly?
  3. What does the minimal 1-on-1 format look like?
  4. How does cadence scale to multi-team?
  5. What failure modes does the 1-on-1 introduce?
  6. When are 1-on-1s overkill?
  7. Where should you go from here?

The single highest-leverage 30 minutes a tech lead spends each week is the one-on-one. Done well, problems surface here weeks before they hit retro or HR. Done poorly, the meeting becomes a status report and the engineer slowly disengages. This chapter shows the format that works and the topics that matter.

When does the 1-on-1 cadence actually start mattering?

Three signals.

You have direct reports. From the first day someone reports to you, the 1-on-1 is the contract. Skipping signals "I don't prioritise you".

Engineers are quiet in groups but issues exist. Standups look fine; retros are short; team morale feels off. The 1-on-1 is where the silence gets words.

Career growth conversations are coming. Promotion cases, mid-year reviews, growth plans need a paper trail of regular discussion - not invented at review time.

If you have no direct reports and you are a peer tech lead, 1-on-1s are still useful but optional. Use them with engineers who pair with you regularly.

What is the cost of skipping or doing 1-on-1s poorly?

Three failure modes.

Surprise resignations. "I had no idea they were unhappy." The unhappiness was there for months; it never had a 30-minute private surface.

Career stagnation. Engineer wants growth opportunities; never told manager directly. Manager assigned junior work; engineer left in 6 months.

1-on-1 as status report. "What did you ship this week?" The engineer has answered the same question at standup. The 1-on-1 becomes redundant; engineer loses interest in attending.

What does the minimal 1-on-1 format look like?

# 1-on-1: {{ Manager }} <-> {{ Engineer }}

## Standing agenda (rotates focus weekly)

- 0-5 min: Status this week (briefly)
- 5-10 min: Blockers I cannot solve myself
- 10-25 min: Rotating topic (see below)
- 25-30 min: Open / anything else

## Rotating topics (pick one per week)

- Week 1: How is the current project going? What is hardest?
- Week 2: Career goals for this quarter
- Week 3: Feedback for me (the manager)
- Week 4: Team dynamics - what is healthy / unhealthy?

## Running notes

### 2026-06-15
- Project: refund flow on track; concerned about Stripe SDK gap
- Blocker: needs design review on email template - I will ping
- Career: wants to lead the next medium-sized project
- Action: I'll find a project for them in Q3

### 2026-06-08
- ...

The shared running notes are the artifact. Engineer can scroll back to see what we agreed; manager can reference past discussions in performance review.

How does cadence scale to multi-team?

flowchart TB
    EM[EM] --> Direct1[Direct report 1<br/>weekly]
    EM --> Direct2[Direct report 2<br/>weekly]
    EM --> Direct3[Direct report 3<br/>weekly]
    EM --> Skip1[Skip-level 1<br/>monthly]
    EM --> Skip2[Skip-level 2<br/>monthly]
    EM --> Peer[Peer EM<br/>bi-weekly]

A typical EM with 5 direct reports and 5 skip-levels spends ~7-8 hours per week on 1-on-1s. That is the job. If the EM is also coding, the 1-on-1 budget is the first thing to cut - and it shouldn't be. The EM role is people; if there is no time for people, the role is wrong.

What failure modes does the 1-on-1 introduce?

When are 1-on-1s overkill?

Two cases.

Truly senior, autonomous engineer. A staff engineer who only needs you for promotion conversations and crisis decisions might prefer monthly. Ask them.

Cross-functional matrix without authority. If you are a PM with engineers who don't report to you, a weekly 1-on-1 is manager-level overkill. A bi-weekly project-context check is appropriate.

Where should you go from here?

Next chapter: hiring and onboarding - how to add people to the team in a way that increases capacity rather than slowing it. After that, performance and feedback covers the structured conversations that 1-on-1s feed into.

Frequently asked questions

How often should 1-on-1s happen?
Weekly for direct reports, bi-weekly for matrix reports, monthly for skip-levels. Weekly seems frequent until you skip one and lose context for two weeks. The 30-minute weekly is cheaper than the hour-long monthly catch-up. If your calendar can't support weekly, cut something else first.
Who owns the 1-on-1 agenda?
The engineer. The manager schedules and shows up; the engineer drives what's discussed. This signals 'this is your time'. If the engineer never has agenda items, that's data - usually they don't trust the manager yet, or they think the meeting is for the manager. Address it explicitly.
What if the engineer never has anything to talk about?
Two patterns. Either things are genuinely fine (rare for more than a few weeks) or they don't trust the format. Probe with rotating topics: career goals, recent project, team dynamics, manager feedback. Not all in one week - one per week. If 'everything is fine' continues for a quarter, something is wrong with how 1-1s are framed.
Should I take notes?
Yes, and share them. A shared 1-1 doc per engineer (in their personal folder) holds running notes, action items, and goals. The engineer adds agenda items between meetings. Notes are not surveillance - they are context for the next 1-1 and the performance review. Engineers value being remembered.