How to Keep a Hard One-on-One From Going Sideways

What each one is actually for

Dialectical Behavior Therapy — Emotion Regulation Skills

The emotion regulation part of DBT is a set of specific, teachable procedures for the situation where a feeling has become too strong to reason your way out of. DBT itself — Dialectical Behaviour Therapy — was built by Marsha Linehan in the late 1980s for people at chronic risk of suicide, and its distinguishing feature is that it does not begin by asking you to think differently. It begins by giving you something to do. The fastest of those is a four-part physical routine DBT calls TIPP: change your body temperature, usually by putting your face in cold water or holding something cold to it; do a short burst of intense movement; breathe out longer than you breathe in; then deliberately tense and release muscle groups. It works on physiology rather than on thought, which is exactly why it is the one that lands when nothing you tell yourself is landing. The other skill worth knowing here is opposite action — you identify what the emotion is urging you to do, then deliberately do its opposite, on the reasoning that acting on an urge is part of what keeps the emotion going. And underneath both sits the simplest one: name what you are feeling, in a word, to yourself.

Effective One-on-Ones

An effective one-on-one is a standing, protected conversation between a manager and one person who reports to them, held on a fixed cadence, and — this is the part that distinguishes it from a status meeting — driven by the report's agenda rather than the manager's. The manager is not there to collect an update; that can be written down. They are there so that whatever is bothering this person has a scheduled, low-stakes place to be said before it becomes the reason they leave. The practical shape is unglamorous: same slot every week or fortnight, rarely cancelled, the report brings the topics, the manager mostly asks and listens, and something gets written down at the end so the next one can start with what happened to it. Its real product is not the topics covered. It is the standing expectation that a difficult thing can be raised here without it becoming an event — which is built slowly, and lost quickly, and lost most often in a single moment when someone said something hard and watched the manager's face change.

What you get out of pairing them

What you get is a report who tells you things while they are still small. That sounds like a soft benefit until you price the alternative, which is finding out in a resignation conversation about a problem that had been running for five months and was fixable for four of them. Here is the mechanism, as I read it. Whether someone brings you a hard thing next month is not decided by your policy on openness or by anything you say about wanting candour. It is decided by what happened the last time they tried — specifically, by the first two seconds, before you had composed anything, when your jaw did something or your tone dropped half a step. People read that instantly and they update on it, and once they have updated they generally do not tell you they have. So the leverage sits in an unusually small window: the seconds between hearing the hard thing and responding to it. That window is governed by how activated your body already was when you walked in, which is governed by the twenty minutes before, which is entirely within your control and which almost nobody uses. The second thing you get, running it the other way, is that the standing meeting gives your own skill practice a place to be reviewed. Learning emotion regulation from a book and then never examining how it went is the reason most people's version fades in a month.

What Dialectical Behavior Therapy — Emotion Regulation Skills rests on

• Affect labeling reduces arousal randomised trials
  Putting an emotion into words reduces the physiological and psychological intensity it would otherwise carry while unexamined.
• Opposite action interrupts emotion behavior loop clinical practice
  Deliberately acting opposite to an emotion's action urge weakens that emotion's behavioral hold over repeated practice.
• Physiological state change precedes cognitive control clinical practice
  Rapidly shifting physiological arousal through techniques such as cold water or paced breathing restores the capacity for cognitive regulation before cognitive strategies become usable.

What Effective One-on-Ones rests on

• Small wins compound randomised trials
  Repeating a small consistent action accumulates over time into a result too large to produce in a single step.
• Early surfacing prevents escalation observational studies
  Regular structured check-ins surface small issues while they are still cheap to address before they compound into larger problems.
• Protected time signals priority observational studies
  Consistently protecting recurring time for a direct report's agenda signals that their concerns matter and increases psychological safety and disclosure.

Where they actually meet

Here is the thing that is easy to miss because the two live in such different buildings. A weekly one-on-one, looked at functionally rather than by job title, is a scheduled appointment for putting feelings into words. Nobody sets it up that way. It is set up as a management practice with an agenda and a doc. But what actually happens in a good one is that someone says out loud, to another person, that they are frustrated with how the handover went, or worried they are not up to the new scope, or angry about a decision — and the saying is doing something distinct from the solving. DBT works the same mechanism from the other end, deliberately and by name: you identify the emotion and put a word on it, on the account that the intensity drops once it is named. One field made it a therapeutic skill practised alone; the other made it a recurring calendar invitation with two people in it and never noticed what it was. That is the meeting point, and it explains a common observation managers make and cannot account for: a report brings a problem, you solve nothing, and they leave visibly lighter. Nothing changed except that it got said to someone who did not flinch.

The principle underneath both

How activated your body is sets the ceiling on what kind of conversation is available, and that ceiling can be moved deliberately in advance. Neither field owns this. DBT states it as a sequencing rule — you cannot use a thinking skill while the physiology is too high, so you change the physiology first and the thinking becomes possible afterwards, in that order and not the other. The management literature arrives at the same place from outside, describing conditions under which people will and will not say a difficult thing, and finding that it hinges on their read of the other person's reaction rather than on their own courage. Put those together and you get something more useful than either alone: in a two-person conversation, there are two ceilings, and the manager's is the one that moves first and constrains the other. A report who is anxious about raising something will calibrate against a manager who is visibly settled, and will abandon it against one who is not. This is why preparation for a hard conversation that consists entirely of rehearsing what you will say tends to underperform. It optimises the words while leaving the state that will actually deliver them untouched.

What that makes the real goal

Detached from both traditions, the goal is: keep the conversation inside the range where both people can still think, and make that range predictable rather than lucky. Not calmness — calm is a state and you do not always get to have it. Predictability is the target, and it is a behaviour: the other person can forecast, before they open their mouth, roughly how you will receive what they are about to say, and their forecast is right. That is what people are actually assessing when they decide whether to raise something, and it is why an inconsistently warm manager is often harder to talk to than a consistently blunt one. Once the objective is stated this way it belongs to nobody in particular, and the question becomes open in a useful way: what else, from any field, reliably keeps two people inside their thinking range during a conversation with stakes in it? Several traditions have worked on precisely that problem, some with better data than either practice here.

Other practices that serve the same goal

Three are worth knowing. The first is Gottman's physiological soothing rule, which came out of a research programme that actually attached heart-rate monitors to arguing couples: when someone's heart rate rises past roughly a hundred beats per minute, their ability to take in new information drops sharply, and the recommendation is to stop for twenty minutes and do something genuinely distracting before continuing. It is the same insight as the DBT sequencing rule, arrived at from measurement rather than from clinical practice, and the twenty-minute figure is empirical rather than invented. The second is Amy Edmondson's work on psychological safety — a body of organisational research, running from studies of hospital teams onward, showing that teams which report more errors are frequently the better-performing ones, because what varies is not the error rate but whether people will say. That is the outcome variable for a one-on-one, measured properly, and it is the strongest evidence on this page for why the meeting matters at all. The third is nonviolent communication, developed by Marshall Rosenberg, which supplies a concrete sentence structure for saying a hard thing — observation, feeling, need, request — and which people find useful precisely because having a form to fall into lowers how much you have to generate under pressure. Its evidence base is thin compared with the other two, and I would not oversell it, but as a scaffold it is genuinely usable.

A tradition that already does this

DBT already built the recurring meeting, and it built it for the therapists rather than the patients. A standard DBT programme includes a mandatory weekly consultation team — clinicians meeting to apply the same skills to themselves, on the explicit reasoning that people delivering this work will otherwise burn out or drift, and that their own regulation is a working part of the treatment rather than a personal matter. Linehan treated it as non-negotiable: a therapist not on a consultation team is not doing DBT. That is a standing, protected, recurring conversation whose stated purpose is keeping the skilled person functional enough to keep doing the work — which is very close to what a one-on-one is for, arrived at inside the clinical lineage, decades before anyone thought of pointing these skills at a manager's calendar. It is a useful precedent because it establishes that the recurring-meeting structure was not bolted on from outside. The tradition that produced the skills also concluded that skills without a review cadence do not survive contact with a hard week.

Using the body-first skills to keep your own face from ending the conversation

The move is entirely preparatory and it takes four minutes. Before a one-on-one you expect to be hard — or before any of them, if you have had a bad morning — do the physical piece first, because the order matters and reversing it wastes it. Splash cold water on your face, or hold a cold can against your cheekbones for thirty seconds; do sixty seconds of something genuinely effortful, stairs are ideal; then breathe with the out-breath clearly longer than the in-breath, roughly four counts in and six or seven out, for about a minute. This is the routine DBT abbreviates as TIPP, and its whole logic is that it operates below the level of thought, which is why it works when you are too wound up for a pep talk to do anything. Then, and only then, do the naming: one sentence to yourself, silently, of what you are actually feeling going in — defensive about the roadmap slip, or worried they are about to resign, or just tired and short. Say the actual word. Vague self-description does not do the job that a specific one does. Now the second skill, which you set up before you go in and use during. Decide in advance what your urge will be when the hard thing lands, and decide the opposite. If the urge is to explain the context, the opposite is a question: "what did that look like from where you were sitting?" If the urge is to fill a silence, the opposite is to let it run four seconds longer than is comfortable. If the urge is to fix it immediately, the opposite is "do you want me to solve this or do you want me to hear it?" — asked genuinely, because the answer is quite often the second one. Write your one opposite action on the top of the doc before you walk in. Under pressure you will not generate it; you will only execute something already decided. When it is working, the tell is small: the report keeps talking after their first point instead of stopping to check your reaction.

Using the standing meeting to keep your own emotion skills from fading

Run it the other direction and the management practice supplies the thing that kills most self-taught emotion regulation, which is that there is no cadence and no review. You read the skill, you use it twice, it works, and then a fortnight later you are back to your old response and you have not noticed, because nothing in your week asks. DBT's own design does not leave this to chance — patients keep a daily record of urges and which skills they used, and a weekly session reviews it. If you are learning these skills outside therapy, you have the skills and not the structure, and the structure is doing more work than people assume. A standing one-on-one can carry it, in either of two ways. If you are the manager, take two minutes at the end of your own one-on-one with your manager, or with a peer you meet regularly, and answer three questions out loud: what set me off this fortnight, what did I actually do, what will I do instead next time. Out loud to a person, not in a notebook — the point is partly the accountability and partly that saying it to someone is itself the skill. If you have no such meeting, make a standing thirty minutes with one peer and take turns. The compounding here is real and slow: one review a fortnight is twenty-six a year, which is more deliberate practice than almost anyone gets on this, and the effect shows up as a shortening of the gap between an emotion arriving and you noticing it. Two cautions. Keep this in a peer meeting or with your own manager, never in the one-on-one with your report — their meeting is for their agenda, and using it to process your reactions inverts the whole thing. And keep the review to what you did, not to what the other person did; the moment it becomes a place to relitigate their behaviour, it has stopped being skills practice.

Tuesday, 2pm, and they say the thing they have been rehearsing since Sunday

It is the usual Tuesday slot. Nine minutes of ordinary stuff, then a pause slightly too long, and they say: "I don't think the reorg was handled well, and I've been thinking about whether this is the right place for me." You are aware of two things arriving at once. Your face wants to do something, and there is a fully formed paragraph in your mouth about why the reorg was constrained by things they do not know about. The paragraph is accurate, and it is the wrong move, and you know both of those things simultaneously. What is helping you here is not willpower. It is that you spent four minutes at 1:52 doing something faintly ridiculous — you went down a flight of stairs and back up, you ran cold water over your wrists in the kitchen, you breathed out slowly six or seven times at your desk — and so the body you are sitting in is not the one you had at 1:45 after the budget call. It gives you about two seconds of room, which is exactly enough. And the top line of your doc says, in four words, ask before you explain. So you say: "okay. Tell me the part you've been thinking about most." And then you do not say anything else for a while. They talk for six minutes. Somewhere in the middle it becomes clear this is not really about the reorg — it is about a specific conversation in March where they found out about a decision affecting them from someone else, and it has been sitting there since. That was never going to come out if your first move had been the paragraph. You do not solve it in the room. You say what you actually think, which is that you did not know that had happened and you would have handled it differently, and you write down the March thing so it is on the doc next time. Afterwards you notice your hands are shaking slightly, which is fine, and which is information you would not have had if you had been talking the whole time.

When this is worth reaching for

Your one-on-ones have gone quiet: they run on time, nothing goes wrong, and nobody brings you anything you didn't already know — which usually means the meeting has become a status update wearing the wrong name. Or you've recently been surprised: someone resigned, or escalated to your manager, or raised something in a survey they'd never raised with you, and in hindsight there were places it could have come up — a direct read on whether the window is open. It also matters when you personally are under real load — a bad quarter, a reorg you didn't choose, something difficult at home — because that's precisely when your baseline activation is high, your two-second window is gone, and the cost lands on people who have no idea what you're carrying. And it matters at the start of a new reporting relationship, in the first eight weeks, because that's when someone is forming their forecast of how you receive things, and the forecast is unusually easy to set and unusually hard to revise afterward. If your reports already bring you difficult things unprompted and you already handle them well, that's a sign you don't have the problem these conditions describe.

Where this does not apply

The hard line first: do not run DBT skills on your report. These skills are for you, on yourself, before and during the meeting. Walking someone through opposite action, or suggesting they try cold water, or naming what you think they are feeling and inviting them to work with it, is a clinical move made by someone with no clinical relationship and considerable power over their income. It is a role violation regardless of how well-intentioned it is, and it is worse when it works, because it makes the relationship one they cannot easily leave. If a report is genuinely struggling, the correct actions are the ordinary managerial ones: ask what would help, adjust load if you can, and point them at the employee assistance programme or occupational health. Second, do not use one-on-ones as the vehicle for feedback that should be its own conversation; a meeting that periodically converts into a performance conversation loses the property that made it useful, because the report can no longer predict which meeting they are in. Third, if you have a personal history of the specific things DBT was built for — chronic self-harm risk, severe emotion dysregulation — then learning these skills from a page rather than from a programme is not the right route, and the full treatment is meaningfully different from the two skills excerpted here. And if there is an active grievance, disciplinary process, or legal matter between you and the report, that conversation belongs somewhere else with someone else present.

What the evidence does and does not support

The two halves here are supported very differently and by different kinds of research. DBT as a whole treatment has substantial randomised evidence, principally for reducing self-harm and suicide attempts in borderline personality disorder, with replications across multiple groups — that is a solid literature. But that evidence is for the full programme, which is a year of individual therapy plus a skills group plus phone coaching plus the therapist consultation team, and it does not transfer to two skills lifted out and used by a manager before a meeting. The individual components have thinner support: putting feelings into words has both brain-imaging work from Lieberman's group and behavioural findings, mostly in fear and exposure contexts, and the effects are real but smaller and more conditional than popular accounts suggest; the physical routine is standard clinical practice with a plausible physiological rationale and little isolated trial data. On the management side there is no randomised trial of one-on-ones at all, and there is unlikely to be one — what exists is observational and case-based, plus the adjacent and much stronger literature on psychological safety, which measures whether people speak up rather than whether a particular meeting format causes it. So: strong evidence that the full therapy works for the population it was designed for, decent evidence that naming an emotion reduces its intensity, good evidence that teams differ in whether people will raise problems, and no evidence whatsoever on the specific proposal here. To test it you would need managers randomised to a pre-meeting routine, with the outcome measured by their reports rather than by them, over at least two quarters. That study does not exist. Run it on yourself and let the quality of what people bring you be the readout.