Most feedback is not actually urgent. The mistake already happened. Waiting 48 hours to address it doesn't materially change anything. It just changes whether anyone can hear you when you do.
Delivering difficult feedback to someone who's already at capacity doesn't make you direct. It mostly means they'll remember how it felt more than what you said. That's not them being sensitive. That's how human memory works under stress. You can be annoyed about it or you can factor it in. Those are the options.
The other thing managers believe is that delaying feedback is somehow dishonest or avoidant. Sometimes that's true. Often it is not. Knowing the difference is, unfortunately, the whole job.
How to read that someone is not in a good headspace
The signals aren't subtle. You just have to be looking.
- Shorter responses than usual, especially in writing. One-word Slack replies from someone who normally sends paragraphs.
- Absence of their usual energy in a meeting: quieter, less engaged, slightly glazed.
- Visible distraction: eye contact dropping, delays before answering.
- Something they said in passing: "yeah it's been a lot lately," "not the best week," "I'm just trying to get through today."
People usually flag their own state if you're paying attention. The difficulty is that in a busy workday, we are often not paying that kind of attention. (This is fine to admit. It is less fine to never adjust for it.)
The case for delaying non-urgent feedback
A useful question to ask yourself before opening your mouth: "If I deliver this right now, what's the best-case outcome?"
If the answer is "they get defensive, half-listen, and leave feeling worse," then waiting is better for everyone. Including for the outcome you were trying to achieve in the first place.
"Non-urgent" covers a lot of ground: most process feedback, stylistic feedback, performance patterns, things you want to address from a meeting that happened yesterday. The category of feedback that genuinely cannot wait is smaller than it usually feels in the moment. Much smaller, if we're being honest about it.
How to acknowledge state without making it weird
You do not need to conduct a full emotional excavation. A light touch is usually better.
Something like: "Hey, I want to chat about something, but I want to check in first. How are you doing? You seemed a bit flat in the standup." Then actually listen to the answer, rather than treating it as a formality before the real conversation starts.
If they say they're fine, take that at face value and use your judgement. If they say they're having a rough time, you now have the information you need. The call you make might be: "Let's leave this for another day." Which you can say simply. "I wanted to talk through something, but it can wait. Let's find a better time this week."
That's not soft. That's reading the situation correctly. (These are not in conflict.)
When the feedback genuinely can't wait
Sometimes it can't. Someone's behaviour in a meeting was serious enough to address today. A mistake is about to compound. A client conversation is happening tomorrow and they need the input first.
In those cases, name the context before you start. Don't pretend you haven't noticed the state they're in.
Try: "I can see today's been a lot. I want to flag something that I don't think can wait, but I'll keep it short and we can talk it through more properly when things are calmer."
Or: "I know the timing isn't great. I just want to make sure we're aligned on [specific thing] before [specific deadline]. This won't take long."
This does a few things. It signals that you've registered their state, which itself lowers defensiveness. It sets an expectation of brevity, which reduces the dread of a long difficult conversation. And it frames the feedback as practical rather than evaluative: you're solving a problem together, not issuing a verdict.
After you've said the essential thing, stop. Don't tack on every other piece of related feedback because you've got them in the room. Do the minimum, leave the door open for a fuller conversation, and actually follow up when the timing is better.
The follow-up matters
If you delay feedback, write a reminder somewhere so it actually happens. "We'll talk about this later" said to someone who's struggling, and then never revisited, is its own kind of message. Not a good one.
Two days later, bring it back: "I wanted to circle back to what I mentioned on Tuesday."
That follow-through, more than almost anything else, is what separates feedback that builds trust from feedback that quietly erodes it. The delay only works if you close the loop. Otherwise you've just postponed an avoidance, which is a different thing entirely.