If you can get the information from a dashboard, you should not be spending thirty minutes of a person's week getting it from their face.
And yet the default 1:1 is a verbal status report. What are you working on, what's blocked, anything else. It's comfortable for both parties. It produces the feeling of management. And it wastes the only recurring slot you have with someone where the agenda is theirs and not yours.
I've run these at Honor, Lyft, Grammarly and Pitch, across organizations from first hire to dozens later, and I've got them wrong in most of the available ways. Here's what I now think.
The slot belongs to them
One change returns more than any other. Say out loud that this is their meeting and you'll take whatever's left.
Most people don't believe you the first time. They'll come with a status update anyway, because that's what every previous manager wanted. It takes about six weeks before someone brings you the thing they actually wanted to talk about. Fill the silence with your own agenda in that window and you never get there.
Six weeks of slightly awkward meetings is the price, and it is worth paying.
What you're actually monitoring
Three things, none of which show up in a project tool:
Whether their picture of their own performance matches yours. That gap is behind nearly every unpleasant surprise in a review cycle. It is also entirely preventable. It exists because somebody found the correcting conversation uncomfortable for two quarters running. I've seen designers arrive from a previous role with a badly distorted sense of their level. It's almost always traceable to a manager who was warm, well-liked and conflict-avoidant.
Whether they're pointed at something they care about. People rarely announce that the work has stopped meaning anything. It shows up as slightly slower responses and slightly less argument. Someone who has stopped pushing back on you is not necessarily agreeing.
What they're not telling you, and why. Every team has information that isn't flowing upward. Your only real tool for it is the 1:1, and it only works if raising a problem has reliably been safe in the past. That's a track record, not a phrase you say.
A team where nobody brings you bad news does not have less bad news.
Recurring culture, and the case against most of it
A related point, because it comes from the same budget.
Recurring meetings pile up. Each one made sense when it started, and almost none are ever looked at again. Within two years a senior person's calendar is mostly things they inherited, and the actual work has moved to the edges of the day.
I've become fairly aggressive about this. Here's the test I use. If this meeting were canceled permanently tomorrow, what specifically would break, and who would notice first? A surprising number have no answer. For some the honest answer is "it's the only time this group talks." That's a real function. But then run it as that, and stop pretending it has an agenda.
One recurring meeting I do think is worth defending almost unconditionally: this one. It's also the first thing people cancel when the week gets heavy, which tells you something about how we rank what is visible against what matters.
Cadence, honestly
Weekly for anyone new, anyone struggling, anyone whose scope just changed. Every two weeks for a settled senior person who'd rather have the time back. And ask them, don't decide for them.
Don't move it. Repeatedly rescheduling someone's 1:1 sends a precise signal, and it is correctly interpreted. If you must move it, move it forward rather than canceling, and say why.
Thirty minutes is usually right, I find. Sixty invites padding. If you consistently need sixty, something is wrong that a longer meeting is compensating for rather than fixing.
Write things down
Keep a running document per person (shared, not secret). What you discussed, what either of you committed to, what you noticed.
Two reasons. First, at review time you have a record rather than a recency bias. Almost everyone's default is to weight the last six weeks as though they were the year. Second, it makes the promotion case writable. Strong people go unpromoted, and it's rarely because they didn't do the work. It's that nobody was accumulating evidence in a form the committee accepts.
Shared is the important part. A private file of notes about someone is surveillance. The same file, open to them, is a working agreement.
The failure I keep repeating
I am consistently too slow with direct feedback in these meetings. Not avoidant, exactly. I'll say it. But I'll wait until I have three examples and a well-formed argument, because I want to be fair and I want to be right.
By which time the person has been doing the thing for two months without anyone telling them. My carefulness cost them the chance to have fixed it in week one. Fairness to the version of them that exists after the conversation is worth more than fairness to my own certainty before it.
Still working on that one.
What the 1:1 cannot carry
Having argued hard for this meeting, I should say where I have seen it do damage, because it is a specific and slightly counterintuitive failure.
A manager who runs excellent 1:1s can build a team that never talks to itself. Everything routes through the center: disagreements between two people arrive separately in two private conversations, the manager arbitrates twice, and neither person ever has the argument with the other. It feels like good management from the inside. What you have built is a hub, and the hub is you, and the team's ability to resolve anything without you has been steadily going down for a year.
It signals that you know things about how two people feel about each other that they have never said to one another. When I notice that now, I stop being the conduit and say so. This is a conversation you two should have. I'll sit in it if that helps.
A related trap: the private slot becomes the only honest channel. If real disagreement only ever happens in 1:1s, meetings become performances, because everyone knows the actual position gets stated later behind a closed door. A team that argues in the open needs to see that it's survivable. They only see that if you let some of it happen in the room.
So the 1:1 is necessary and it is not enough on its own. It fails by being good enough that nobody notices what it is quietly soaking up.
I've led design and product organizations at Honor, Lyft, Grammarly and Pitch, from first designer to dozens later. Reasonable disagreement exists on cadence: some very good managers run monthly 1:1s with senior ICs and use the recovered time for written updates. I haven't made that work, but I've seen it work.
© 2026 Renato Valdés-Olmos