Three Weeks Without Slack

In December, I worked with a small team on a project where we agreed, as an experiment, to remove Slack from the working environment for the duration. Three weeks. Email, video calls when genuinely necessary, and a shared document for asynchronous status updates. No real-time channel.

What changed immediately: the number of times I was interrupted. In the first three days, I noticed that my capacity to stay with a problem for more than thirty minutes had increased noticeably. This was not a placebo effect — the interruptions had stopped, and so had the attention fragmentation that accompanied them. This part worked exactly as the theory would predict.

What did not change: the volume of communication required to coordinate the work. The messages did not disappear. They arrived as emails instead of Slack messages. What changed was the pace at which they arrived and the pace at which I was expected to respond. Email carries a different social contract than a real-time channel. The expectation of near-immediate response that Slack imposes — partly by design, partly by social convention — largely dissolved. This was the more significant change.

What became apparent by week two: a significant proportion of the conversations that had been happening in Slack were not, in retrospect, conversations. They were status broadcasts — pieces of information someone wanted to share, which required acknowledgment from others, but did not actually require real-time dialogue. These migrated cleanly to the shared document and became more legible for having been written in a linear format rather than scattered across threads.

What I rebuilt instead: a practice of writing a brief end-of-day summary — three to five sentences about what I had worked on, what was blocked, what I expected to do next. This replaced the ambient awareness that a real-time channel provides. The delay was acceptable. The improvement in the quality of my own working conditions was significant enough that I have continued the practice on subsequent projects, even when Slack is nominally available.

One note of honesty: this experiment was possible because the team was small, the project was time-bounded, and everyone agreed to it in advance. These conditions are not universal. The observation holds within its scope.