You leave the call certain you came across as cold. You weren't. You were warm and engaged, and you laughed in all the right places. But there was that moment when you both started speaking at once, both stopped, both said "no, you go," and then both started again. After three rounds of that, the rhythm was gone. Everyone's smile got slightly stiffer. You spent the rest of the meeting feeling like the one person in the room who couldn't hold a conversation.
You can hold a conversation. Your timing was just working against a machine that was never built to respect it.
The gap you've never noticed
In face-to-face talk, people switch speakers with startling precision. Stivers and colleagues compared conversations across ten languages in a 2009 study in PNAS. They found that in every language, the typical gap between one person finishing and the next starting was around 200 milliseconds. That is about the length of a blink. Languages differed slightly, but all clustered near the same tiny window.
That speed is odd, because producing a sentence takes much longer. Researchers like Stephen Levinson have argued that planning even a short reply takes well over half a second. The only way to answer in 200 milliseconds is to start preparing your response while the other person is still talking. You predict where their sentence will end, using grammar, intonation, and the rhythm of their speech. Then you launch on cue, like a sprinter reading the starter's shoulders.
Nobody teaches this. Children pick it up as toddlers. It runs underneath conscious thought, which is why you only notice it when it fails.
What a delay does to that machinery
Video calls add latency. Often it's only a few hundred milliseconds, but that is enough to wreck a system tuned to 200.
Suppose the delay is a third of a second each way. You finish your sentence, and your colleague hears the end of it late. They respond on their usual schedule, and their answer reaches you late again. From your side, the silence after you stopped was far longer than 200 milliseconds. Your brain does what it always does with a long gap. It assumes something is wrong, so you start talking again, just as their reply arrives. Now you've collided, though neither of you did anything impolite.
The pattern gets worse because the cues you use to predict turn endings are weakened. Compressed audio flattens intonation. A frozen frame hides the small breath or eyebrow movement that signals "I'm about to speak." Your predictions get worse, so your timing gets worse.
Research on video conversation points the same way. A 2021 study in the Journal of Experimental Psychology: General by Boland and colleagues compared conversations over a Zoom-like connection with in-person ones. They found that the video conversations had longer gaps between turns and fewer of the quick, overlapping bits of feedback that keep talk flowing. Other work, including Templeton and colleagues in 2022, has found that slower responses are read as signs of lower social connection. People treat a long gap as information about the relationship, not just about the wiring.
That is the uncomfortable part. Your listener's brain does not file the delay under "bad internet." It files it under "this person seems less interested in me."
Why it feels personal
We interpret response time as a social signal because, for all of human history, it was one. A fast reply says "I was listening and I'm with you." A pause says "I'm thinking," "I disagree," or "I'd rather not." Your brain reads these cues automatically and without asking whether a router is to blame.
So on a laggy call, both parties are quietly misreading each other. You think they hesitated before answering your question. They think you seemed distracted. Neither is true. Both are the same 400 milliseconds of network, interpreted by two brains running ancient software.
This also explains why talking over each other on video feels worse than the same overlap in a room. In a room, an overlap resolves in a fraction of a second, because you both hear it live and one of you yields without thinking. On video, audio systems often cut one speaker off, so you each hear a swallowed half-word and a sudden silence. That is jarring in a way a normal interruption isn't, and it leaves both people feeling slightly rude.
Slow down on purpose
The fix is not to try harder. It is to swap the instinctive system for a deliberate one, because the instinctive one is miscalibrated for this medium.
Start by understanding that you now have to wait slightly longer than feels natural before speaking. If the habit you carry is "respond at 200 milliseconds," give it more like a full second on video. It will feel like an eternity, and to the other person it will feel normal, because they are living on the same delayed clock.
Next, make the end of your turn unmistakable. In a room, a trailing intonation or a glance does the job. On a call, those cues are muffled, so use words. "That's my thinking, what do you see?" is clumsy in person but works well on video, because it removes the guesswork that lag creates.
Finally, treat a collision as a shared glitch and not a failure of manners. The best repair is a quick, light one. A half-smile and "go ahead, I think we're lagging" lets both people off the hook, and it reframes the awkwardness as the technology's fault, where it belongs.
Your next moves
- Run a one-minute latency test today. Before your next call, ask a colleague to count aloud with you, alternating numbers ("one," "two," "three"). If you stumble over each other, you now know the lag is real and you can stop blaming yourself.
- Add a one-second pause before your first reply. On your next video call, silently count "one-Mississippi" after someone finishes speaking before you begin. Notice how many collisions never happen.
- End your turns with an explicit handoff. Practice closing a long comment with a clear signal: "That's all from me," or by naming the next speaker. Do it at least three times in your next meeting.
- Name the lag once, early. If a call is clearly delayed, say so in the first five minutes: "We're going to talk over each other a bit, so let's leave a beat before replying." It sets the norm and takes the sting out of every overlap that follows.
- Fix what you control. Switch from Wi-Fi to a wired connection if you can, close bandwidth-heavy apps, and use a headset so your own audio doesn't echo. Lower latency on your end shrinks the gap for everyone.
The part that stays with you
After a stiff, stammering call, it's tempting to conclude something about yourself: that you're not charismatic enough, or that you misjudge rooms. Almost always, the truth is humbler. You were running a beautifully tuned timing system in an environment that scrambled it. What you can do is notice the pattern afterward, separate "that was a lag problem" from "that was a me problem," and adjust one small habit at a time.
That after-the-fact noticing is what MeetingMortem is built for. It gives you a short space to look back at a meeting, record how the moments really felt, and spot the patterns you'd otherwise carry around as vague self-doubt. If you'd like a calmer way to review your calls, you can try it at meetingmortem.lumenlabs.works.