Async Communication: How Distributed Companies Replace Real-Time Dependency

What It Is

Async communication means sending a message without expecting an immediate response. The recipient responds when they are ready — not when you sent the message.

At its simplest, this is email. But modern async-first companies have turned it into a full operating system: replacing status meetings with structured written check-ins, replacing chat interruptions with threaded discussions, replacing verbal decision-making with documented deliberation.

Async is not the absence of communication. It is the deliberate design of communication so that it does not require two people to be available at the same moment.

Why It Works

Interruptions can increase stress even when people compensate by working faster. In a laboratory experiment, Mark, Gudith, and Klocke found faster completion of interrupted tasks, with no significant quality difference, but greater stress, frustration, time pressure, and effort. The study does not establish a universal 23-minute recovery cost. (Source)

Writing can expose gaps in an explanation. 37signals argues for long-form written context and preserving what is communicated. [Inference — writing is a mechanism to examine; this source does not measure a general improvement in decision quality.] (Source)

Async can reduce dependence on overlapping hours. [Inference — design implication of the documented GitLab and Buffer guidance.] Preserve context for colleagues who are offline while retaining live discussion when the task benefits from it. These sources do not establish that all globally distributed teams must use one communication model. (GitLab, Buffer)

Preserved context can extend participation beyond attendance. GitLab asks people to document conclusions from offline conversations. A meeting can also leave notes or a recording; written communication only remains useful when it is retained, discoverable, and current. [Inference — implementation conditions, not a measured outcome.] (Source)

Where It Fails

When ownership and response expectations are unclear. [Inference — implementation risk.] A written thread can stall when nobody knows who must decide or when to escalate. GitLab’s guidance supplies one explicit escalation rule; it does not establish that every async rollout becomes slower. (Source)

When the work needs immediate interaction. Buffer names urgent or complex work and sensitive topics as reasons for live discussion; GitLab also allows deliberate synchronous starts. [Inference — keep a live route appropriate to the situation rather than treating every conflict as requiring one format.] (Buffer, GitLab)

When the expected format creates a barrier. [Inference — accessibility and implementation question, not a measured hiring or promotion effect.] Strong written expectations can require training, feedback, or another format. The sources here do not provide hiring-funnel or promotion comparisons by communication skill. (37signals, Buffer)

Notification norms require active management. “Async” with Slack’s default notification settings is not async — it is just intermittent stress. Buffer explicitly published their Slack notification agreements for this reason. (Source) The tools default to urgency; the norms have to actively override that default.

Best Examples

37signals / Basecamp — Its handbook describes daily and weekly automatic check-in prompts, plus manual cycle Kickoffs and Heartbeats. The daily question requires an answer at least twice a week; the weekly question is required weekly. These publish work context without requiring everyone to respond together. They do not establish how many meetings were eliminated. (Handbook, communication principles)

Todoist (formerly Doist) — Async communication alongside recurring support. Its published accountability system describes regular manager 1:1s; the current onboarding guide includes weekly mentor and team-head check-ins. These are deliberate meetings within an async working model. (Accountability, onboarding)

GitLab — Async is the starting point, with deliberate synchronous discussion for strategy, disagreements, trust-building and hard-to-express context. The three-exchange guideline helps identify stalled conversations; it is not a prohibition on starting with a call. (Source)

Buffer — Its November 2023 account explicitly says Buffer is remote and not async-first. It describes live calls for urgent or complex work and relationship-building. The named Threads tool is a historical example. (Source)

Implementation Guide

[Inference — implementation questions from the cited company examples; response windows and outcomes require local evaluation.]

1. Specify response expectations. State which requests can wait and who owns escalation when waiting is costly.
2. Replace a status meeting only when its function is covered. Test whether a written update supplies the context and decisions people need; counting a removed meeting does not measure productivity.
3. Attach discussion to the work. Link the task, document, or decision record rather than scattering its context across unconnected channels.
4. Set notification expectations explicitly. Define when a message requires a live response rather than assuming a tool makes that choice.
5. Include context in a request. Explain the goal, what has been tried, and the decision or help required.
6. Define live starts and escalation paths. Retain synchronous discussion for work that benefits from immediate interaction.

(37signals, GitLab, Buffer, Zapier)

Common Mistakes

[Inference — failure modes to examine, not measured cross-company frequencies.]

• Removing a meeting without preserving its function. Check whether context, ownership, and a decision route remain available.
• Equating fewer interruptions with a demonstrated company outcome. Evaluate delivery and coordination as well as individual experience.
• Treating a tool as the operating practice. State the response, notification, and escalation norms.
• Leaving writing support implicit. Provide examples and feedback, and allow another format when it serves the work or the person better.

(GitLab, 37signals, Buffer)

Related Frameworks

• Sync vs Async Decision Matrix — choose a communication mode for the work
• The 20–30 Person Fracture Point — a working hypothesis about coordination needs during growth

Sources

Additional direct references for this review:

• https://zapier.com/blog/how-to-work-asynchronously/

Other retained links below include historical/background context; dated accounts do not establish a current universal policy.

1. The 37signals Guide to Internal Communication: https://basecamp.com/guides/how-we-communicate
2. 37signals Employee Handbook: How We Work: https://basecamp.com/handbook/how-we-work
3. How Doist Works Remote: https://doist.com/how-we-work/how-doist-works-remote
4. GitLab: How to Embrace Async Communication: https://handbook.gitlab.com/handbook/communication/#asynchronous-communication
5. How to Work Asynchronously — Zapier: https://zapier.com/blog/how-to-work-asynchronously
6. Buffer is Remote but not Async-First: https://buffer.com/resources/remote-not-async-first/
7. Gloria Mark, Gudith, Klocke — The Cost of Interrupted Work: More Speed and Stress, CHI 2008 (laboratory study of interruption, speed and stress): https://ics.uci.edu/~gmark/chi08-mark.pdf

Todoist accountability (September 2022): https://async.twist.com/accountability-remote-workers/

Todoist onboarding (accessed September 2026): https://handbook.todoist.net/s/6ca6f66e-2d02-4e05-b629-b1c772f5ff23

GitLab current communication guidance: https://handbook.gitlab.com/handbook/communication/

Inferences

• GitLab, 37signals, and Buffer document mechanisms for choosing channels, preserving context, and deciding when to meet. These cases suggest examining those mechanisms alongside a company’s async label; they do not establish a performance ranking or a six-month failure timetable.
• Async-first works best when the organizational culture also values written communication as a form of thinking — not just record-keeping. When writing is treated as intellectual work, the quality of async communication rises. When writing is treated as bureaucratic overhead, people do it reluctantly and poorly.

Work with Alex

If your team is drowning in meetings and you want to design a coordination system that actually gives people back their time, Alex helps leadership teams build the operating infrastructure for async-first distributed work.