You have four basic ways to get a message across at work: live or delayed (synchronous or asynchronous), and spoken or written. A call is synchronous and verbal; a recorded video is asynchronous and verbal; a meeting whiteboard is synchronous and written; a memo is asynchronous and written. Almost every channel, Slack, email, a call, a doc, a stand-up, is some blend of those two dials. Choosing well isn't a personality trait. It's a decision you can make on purpose, and the research gives you a rule for it.

The quick version

  • Synchronous means everyone is present at once (a call, a meeting); asynchronous means people respond in their own time (email, a doc, a recorded video). The trade is speed-of-resolution against respect for focus.
  • Written communication is precise, searchable and rereadable but slow to clear up confusion; verbal is fast, rich and good for nuance but vanishes the moment it's said.
  • The useful rule: use richer, live channels to agree on something hard or contentious; use leaner, written, async channels to convey information and let people process it.
  • The two traps are opposite. Defaulting to meetings burns everyone's attention; defaulting to chat creates a constant back-and-forth that's slower and more stressful than the meeting would have been.

The idea in depth: richness, and what it's for

The foundational idea here is media richness theory, set out by Richard Daft and Robert Lengel in 1986. A "rich" medium carries many cues at once, tone, facial expression, instant back-and-forth, body language, while a "lean" one carries few. Face-to-face is the richest channel; a number on a spreadsheet is among the leanest. Their argument, summarised on the media richness theory overview, is that you match richness to equivocality: when a message is ambiguous and could be read several ways, a sensitive decision, a half-formed strategy, a piece of feedback that could sting, you need a rich channel to resolve the ambiguity in real time. When a message is clear and unequivocal, a date, a status, a confirmed number, a lean channel does the job, and a meeting is overkill.

So the move is to ask one question before you pick a channel: could this be misread? If yes, go richer and probably live, get on a call, share your screen, watch the other person's face. If no, write it down and send it; a meeting to "walk through" something unambiguous is a tax on everyone's calendar. The honest limitation is that richness is a tendency, not a law: a calm, well-written message often beats a rushed call, and Daft and Lengel's framework predicts media choice better than it predicts performance. Treat it as a default to reason from, not a verdict.

The theory was sharpened two decades later. In "Media, Tasks, and Communication Processes: A Theory of Media Synchronicity" (MIS Quarterly, 2008), Alan Dennis, Robert Fuller and Joseph Valacich made a distinction that's more useful at the desk than "rich vs lean." Communication, they argue, does two different jobs: conveyance (transmitting a lot of new information for people to digest) and convergence (reaching shared agreement on what it means). For conveyance, lower-synchronicity media work better, people need time alone to read, process and reprocess. For convergence, higher-synchronicity media work better, agreement needs the fast, shared rhythm of a live conversation.

flowchart TD
  A(["What's the job
of this message?"]) --> B{"Convey new info,
or converge on meaning?"}
  B -->|"Convey: lots to absorb"| C(["Go async + written
doc, memo, recorded video"])
  B -->|"Converge: agree something"| D(["Go live + verbal
call, meeting, screen-share"])
  C --> E(["People read, reprocess,
respond in their own time"])
  D --> F(["Shared rhythm resolves
ambiguity fast"])
Media synchronicity theory in one decision: match the channel to whether you're conveying or converging. Leaders Loop

That reframes the whole "should this be a meeting?" debate. A meeting is a convergence tool. If your meeting is really just one person conveying information to a passive audience, it's the wrong channel, that's a document, sent in advance, that people read when their attention is fresh. Save the live time for the part where you actually need to agree.

Why "just message me" is often the expensive option

The seductive belief is that asynchronous chat is always the lighter, kinder choice, you fire off a message, the other person replies when they can, nobody's day is interrupted. Cal Newport, the Georgetown computer scientist, dismantles this in A World Without Email (2021). He calls the default mode of modern work the "hyperactive hive mind": an unstructured stream of back-and-forth messaging that we mistake for efficient. His counter-intuitive point is that asynchronous communication often lengthens the conversation, because a question two people could settle in ninety seconds of live talk instead stretches across a dozen messages sent hours apart, and each ping pulls someone out of focused work. There's a real cost to never paying the "synchrony tax" of just scheduling the quick conversation.

Asynchronous feels lighter because the cost is hidden, paid in interruptions, not in calendar slots.

What this asks of you is to notice when a thread is thrashing. If a Slack channel or email chain has gone three or four rounds without converging, stop typing and switch channels: a five-minute call almost always beats the fifth message. Conversely, and this is the discipline most teams lack, protect the deep, asynchronous work by not expecting instant replies. The honest limitation: Newport writes primarily about knowledge work inside offices, and his examples lean on data he discusses rather than a single controlled trial, so read it as a sharp argument backed by observation, not a settled experiment. But the asymmetry he names is real, and remote-first companies have built whole operating models on it. GitLab's public guide to asynchronous communication codifies the same instinct: document first, transfer the work, and reserve live time for what genuinely needs it.

Written vs verbal: precision against presence

The second dial, written or spoken, trades two different goods. Writing forces clarity (you can't hide a muddled thought in a sentence the way you can in a ramble), it's searchable, and people can reread it. Speaking is faster, carries far more emotional nuance, and lets you course-correct the instant a face falls. The mistake is assuming the spoken word always "lands better" because it's warmer.

That assumption usually traces back to a famous, mangled statistic: that only 7% of communication is the words, 38% is tone, and 55% is body language. This is the most over-claimed number in all of communication training, and it is misused. It comes from Albert Mehrabian's 1960s work, described on his Wikipedia entry, and it applies only to a narrow case: when someone is communicating feelings and attitudes and their words contradict their tone. Mehrabian himself stated that unless a person is talking about feelings or attitudes, the equation does not apply. It is not a law that "words barely matter", for conveying actual information, the words carry the load.

The practical answer is to separate the two jobs. When the content is emotional or relational, difficult feedback, a decision someone will be unhappy about, anything where how it feels matters, go verbal and live, because tone and presence do real work there, and a curt written version reads colder than you mean it. When the content is informational, complex, or needs to be acted on later, a brief, a decision record, a set of instructions, write it down, because precision and a searchable record matter more than warmth. The limitation worth naming: written-only feedback strips out the reassurance a face provides, so use it for praise and logistics freely, but think twice before delivering hard news in text.

quadrantChart
  title Pick a channel by message type
  x-axis "Lean / written" --> "Rich / verbal"
  y-axis "Asynchronous" --> "Synchronous"
  quadrant-1 "Live + verbal: agree, decide, repair"
  quadrant-2 "Live + written: workshop, whiteboard"
  quadrant-3 "Async + written: brief, decide-on-record"
  quadrant-4 "Async + verbal: recorded update, voice note"
  "Hard feedback": 0.85, 0.85
  "Status update": 0.2, 0.25
  "Strategy brief": 0.25, 0.2
  "Crisis decision": 0.8, 0.9
  "Roadmap walkthrough": 0.6, 0.55
The two dials as a map: where a message sits suggests the channel. Positions are illustrative. Leaders Loop

A worked example

(Illustrative throughout, a teaching scenario, not real figures.) A product lead, Priya, needs to do three things this week: tell the team a launch date has slipped, decide with two engineers how to re-sequence the work, and let one engineer know their last piece of work fell short. Her instinct is to call one big meeting and "talk it all through." Run it through the dials instead.

The slipped date is conveyance, clear information, lots of people, no real ambiguity. So it's async and written: a short note in the channel with the new date and the reason, posted once, where everyone can read it when their attention is fresh and search it later. No meeting required. The re-sequencing is convergence, genuinely equivocal, three people, trade-offs to weigh. So it's synchronous and rich: a thirty-minute call with a shared screen, where ambiguity gets resolved in real time and a decision is reached. The hard feedback is emotional and relational. So it's verbal and live, one-to-one, never a written message and never inside the group call, because tone and presence do the work that text can't, and the engineer deserves to read the room as Priya speaks.

One week, three messages, three different channels, and the only live time spent is the half-hour that actually needed agreement, plus a short private conversation. The version where Priya called one big meeting would have bored most of the room with information they could have read, rushed the decision, and forced the feedback into a setting that made it worse. Choosing the channel is the skill.

Frequently asked questions

Should we just default to asynchronous to protect focus?

As a default for conveying information, yes, it respects deep work and leaves a record. But "async-first" curdles into the hyperactive hive mind the moment a thread starts thrashing. The rule is to default async, then switch to a live call the instant a conversation has gone three or four rounds without converging. Protecting focus and clinging to text are not the same thing.

Isn't a meeting always richer and therefore better?

Richer, yes; better, only for the right job. Media synchronicity theory says high-synchronicity channels win for convergence, reaching agreement, and lose for conveyance, where people need quiet time to absorb. A meeting that's really one-way information transfer is a document wearing a calendar invite. Use live time for the part where you genuinely need to decide together.

Is it true that 93% of communication is non-verbal?

No, that's the most misused statistic in the field. Mehrabian's 7-38-55 figures apply only to communicating feelings and attitudes when words and tone conflict; Mehrabian himself said the equation does not apply otherwise. For transmitting actual information, the words carry the meaning. Use rich verbal channels because they help with emotion and ambiguity, not because text barely registers.

When should hard feedback be written versus spoken?

Default to spoken and live for anything difficult or emotional, because tone and presence soften and clarify in ways text can't, and a written version reads colder than intended. Writing is excellent for praise, logistics and creating a record, but delivering bad news in text removes the human cues that make it survivable. If you must follow up in writing, do the conversation first.

How do I stop my team drowning in Slack?

Two moves. First, make conveyance deliberate, important information goes in a durable, searchable place (a doc or a pinned post), not buried in a live channel. Second, normalise switching to a quick call when a thread stalls, and explicitly drop the expectation of instant replies so people can do focused work. The goal isn't less communication; it's putting each message in the channel that fits it.

Related in the Toolkit

Choosing a channel is upstream of almost everything else you do with words: once you've picked async and written, the discipline of structured communication decides whether the reader gets your point in the first line, and how you frame the message for a particular audience (managing up, down & across) shifts with whether you're writing to your boss, your team or a peer.

Where to go next