Timing: Choosing When to Have the Conversation
Right words, wrong moment
You planned it properly. You knew your one observation, you knew your one ask, you'd even practised it.
Then you said it at half eleven at night, when they'd just got in, while their flatmate was in the next room, four minutes before they had to leave. And it went exactly as badly as if you'd planned nothing at all.
This is the most under-rated part of a hard conversation, and it's the cheapest to fix. Everything in SO-09 assumes conditions that mostly aren't there by default — and you don't have to find a perfect moment. You have to avoid four predictably terrible ones.
The structure: You · Them · Where · Deadline
Four checks. The first three are conditions. The fourth stops the first three becoming an excuse.
1. You — are you in a state to do this well?
Three questions, and they're deliberately unglamorous:
- Are you still angry? Not "have you calmed down enough to be polite" — are you still building a case in your head? If yes, you'll deliver the case, not the ask.
- Do you know what you actually want? If you can't finish the sentence "what I want is…", the conversation has no destination and will wander into every other grievance you own.
- Are you tired, hungry, or three hours into something stressful? Judgement under those conditions is measurably worse. Research on decision-making keeps finding that outcomes shift with fatigue and breaks in a way the decider never notices.
The 24-hour rule: if it still matters tomorrow, have it tomorrow. Most things that feel urgent at 11pm are not urgent; they're just loud. The ones that still matter in the morning are the real ones — and now you're the version of yourself that can handle them.
2. Them — are they in a state to hear it?
You don't control this, but you can read it. Bad moments, reliably:
- Mid-task. Someone halfway through something will resent the interruption and hurry you.
- About to leave. A conversation with a hard stop turns into a summary, and hard conversations don't survive being summarised.
- In front of other people. This is the big one. An audience means they're not deciding how to respond to you, they're deciding how to look. You've made it a performance and lost the real conversation.
- Just been hit with something else. Bad day, bad news, bad mark. Your issue joins a queue and gets the response the queue deserves.
The move: ask, don't ambush. "Have you got twenty minutes at some point today?" gives them the chance to pick a moment where they can actually be present — and being asked is itself de-escalating, because it signals this isn't a trap.
3. Where — the room does a surprising amount
- Private. See the audience point. Non-negotiable for anything with stakes.
- Enough time. Twenty minutes minimum, and know what's after it. If either of you has something in fifteen minutes, don't start.
- Side by side beats face to face. Walking, driving, washing up, going somewhere. Sustained eye contact raises the intensity of a hard conversation; being shoulder-to-shoulder lowers it, and gives everyone somewhere neutral to look. This is why difficult things get said in cars.
- Neutral ground, and an exit that isn't dramatic. In someone's own room, one of you is host and one is a guest, and leaving becomes a statement. A kitchen, a walk, a café — either person can disengage without it being an event.
- Not in writing. Text and DMs strip tone, invite re-reading, and let both of you draft. Written is for arranging the conversation, not having it.
4. Deadline — or you'll wait forever
Here's the trap this lesson sets if you stop at check three: there is always a reason today isn't ideal. Waiting for good conditions turns into waiting, and waiting turns into the thing quietly getting worse while you tell yourself you're being thoughtful.
So the fourth check is a date.
"I'm having this conversation by Sunday. If the ideal moment hasn't shown up by then, I'll take an adequate one."
Write the date down. An adequate moment plus an actual conversation beats a perfect moment that never arrives — by an enormous margin.
Why this works
1. It removes the failure modes you can predict. You can't engineer receptiveness. You can absolutely avoid an audience, a hard stop, and 11pm — and those three account for a large share of conversations that go wrong for reasons unrelated to what was said.
2. Asking for the time is itself part of the conversation. "Have you got twenty minutes?" tells them something's coming and gives them agency over when. Both lower the defensive response before a word of the actual content is spoken. It's the Ground move from SO-09, arriving early.
3. Side by side lowers the temperature. Face-to-face with sustained eye contact is a confrontation posture. Shoulder to shoulder is a collaboration posture. The content doesn't change; how hard it is to say does.
4. The deadline converts avoidance into a decision. Without one, "not now" is unfalsifiable — you can always say it again tomorrow. With one, you either have the conversation or you consciously choose not to, and both are better than drifting.
Patterns
Asking for the time, without making it ominous:
| Setting | Line |
|---|---|
| Friend | "Are you around this week? There's something small I want to talk about — nothing dramatic." |
| Family | "Have you got half an hour this weekend? I want to sort something out, not start a row." |
| Work | "Could I get fifteen minutes with you tomorrow? It's about the handover." |
| Partner | "Can we talk properly at some point today? Not right now — when we've both got time." |
| Turning down a bad moment | "I do want to talk about this — but not at eleven at night. Tomorrow?" |
That last one is the most useful line in the lesson. You are allowed to decline a bad moment, including one the other person started. Saying "not now, but genuinely soon, and here's when" isn't dodging.
Don't say: "We need to talk." (Alarm, no information, and now they have hours to construct a defence.) · "Quick question—" for something that isn't quick. · "Can I ask you something?" with no indication of what, which produces the same dread with extra suspense.
When it doesn't work
- When it's genuinely urgent. Safety, a deadline today, something about to be decided without you. Take the imperfect moment.
- When timing has become the excuse. If you've rescheduled three times, the problem isn't the calendar. Look at check one — you probably don't know what you want yet, and that's the thing to fix.
- When there is no good moment, ever. Some households and some workplaces don't contain a private twenty minutes. Then the answer is to create one deliberately — a walk, a lift somewhere, a coffee — rather than to keep waiting for one to occur naturally.
- When they refuse every time you ask. Repeatedly declining to find twenty minutes is itself an answer, and it's information about the relationship rather than about your scheduling.
- When it's small. Not everything needs conditions. Most things can just be said.
The rehearsal protocol
Stage 1 — Written (5 min)
CONVERSATION: .................................................
YOU — still angry? .... know what I want? .... tired? ....
THEM — best likely window this week: ....................
WHERE — private? .... 20+ min? .... side by side? ....
DEADLINE— having it by: ......................................If the "what I want" line is blank, stop. That's not a timing problem and no moment will fix it.
Stage 2 — Rehearse against an AI (10 min)
Full setup: SO-22. Short scenario — you're only rehearsing the ask, not the conversation.
We're going to run a short roleplay rehearsal.
YOU ARE: [who]. It's a normal day, you're mildly busy, nothing is wrong.
I AM: someone about to ask you for a window to talk about something.
RULES
- Stay in character, dialogue only, human-length replies.
- Difficulty 3/5: don't make it easy. Ask "why, what's this about?"
and be slightly wary — as most people are when someone asks to talk.
- Stay in character until I type END REHEARSAL.
WHEN I TYPE "END REHEARSAL":
1. Did my opening sound ominous? Quote it.
2. Did I give them a real choice about when, or a fake one?
3. When you asked "what's this about", did I answer honestly or dodge?
4. Write one better version of my request.
Reply only "ready" and wait.Everyone fails point three the first time. "What's this about?" is the question that decides whether the conversation starts well, and the instinct is to say "nothing, it's fine" — which guarantees they spend the intervening hours assuming the worst.
Stage 3 — Low-stakes real (this week)
Ask for a window for something that barely matters — swapping a shift, splitting a bill, a plan you want to change. Practise the ask, not the argument.
Exercise (25 min, verifiable output)
- Pick a conversation you've been putting off.
- Fill in the four checks from Stage 1. Be honest on "still angry" and "know what I want."
- Look at the last three times you tried to raise something and it went badly. For each, write which check was missing. A pattern will show up — most people fail the same check every time.
- Run the Stage 2 rehearsal. Pay attention to how you answer "what's this about?"
- Write a date. Then actually ask for the window, this week.
✅ Finish check: you have (a) a completed checklist, (b) your recurring failed check named, (c) a date written down, (d) the window actually requested.
Summary card
- You: still angry? know what you want? tired? — 24-hour rule: if it still matters tomorrow, have it tomorrow.
- Them: not mid-task, not about to leave, not in front of other people, not right after bad news.
- Where: private · 20+ minutes · side by side beats face to face · neutral ground · not in writing.
- Deadline: write a date, or you'll wait forever. An adequate moment beats a perfect one that never comes.
- Ask, don't ambush — "have you got twenty minutes?" is de-escalating before you've said anything.
- Never open with "we need to talk."
- You're allowed to decline a bad moment, as long as you name a real one instead.
- If you've rescheduled three times, it isn't the calendar — you don't know what you want yet.
Sources
- Stone, D., Patton, B., Heen, S. — Difficult Conversations, Harvard Negotiation Project, 1999
- Patterson, K. et al. — Crucial Conversations, 2002 (on establishing conditions before content)
- Danziger, S. et al. — Extraneous factors in judicial decisions, 2011 (on fatigue and judgement)
Next lesson: SO-09 — How to Open a Hard Conversation (L2) Related: SO-03 Active Listening · SO-08 Setting Boundaries · SO-22 Rehearsing with AI Path: The Rehearsal Room — prerequisite (required before SO-09)