Workplace Interruptions: How to Measure Them and Actually Reduce Them
Every knowledge worker knows the feeling of a day with no interruptions, mostly because it happens so rarely. Interruptions are the largest single tax on focused work in most teams, and they are almost never managed, because they arrive one at a time, each one reasonable, from people who are also being interrupted. This guide treats interruptions as a team problem with a team fix: where they come from, why they cost more than they appear to, how to measure the load, and which agreements reduce it.
Why Interruptions Cost More Than the Minutes They Take
A two-minute question does not cost two minutes. It costs the two minutes, plus the time to re-find where you were, plus the error rate increase in the work that follows, plus, for work that needs sustained thought, sometimes the whole session, because the thread cannot be recovered. Research on task switching has consistently found that the resumption cost is many times the length of the interruption itself, and that it rises with the complexity of the interrupted work.
There is a second cost that is easier to miss. A person who expects to be interrupted stops attempting work that needs uninterrupted time. They do the shallow work that survives interruption and defer the deep work indefinitely, and the deferral is invisible because nothing is visibly failing. Our guide on context switching covers the cognitive side; this guide is about the sources and the fixes.
The costs are also unevenly distributed. The people most interrupted are usually the most knowledgeable, because they are the ones colleagues go to, which means interruptions systematically tax the team’s highest-value time.
Where Interruptions Come From
Interruptions are usually discussed as if they were all the same thing. They are not, and the source determines the fix.
Chat and messaging. The largest source in most remote and hybrid teams. Direct messages carry an implicit expectation of a fast reply, group channels generate notifications for everything, and the tools default to interrupting. Meetings. Not interruptions in the usual sense, but a calendar with meetings every ninety minutes makes uninterrupted work structurally impossible; our guide to meeting overload covers this source. Colleagues in person. The desk visit, the tap on the shoulder. Declining in remote teams, still dominant in offices and open-plan layouts. Email. Less interruptive than chat because the reply expectation is slower, but still checked compulsively. Self-interruption. The person’s own switching to check something, which is often a response to the expectation of external interruption. Tools and systems. Alerts, build notifications, monitoring dashboards, ticket updates.
Ask a team to estimate the split and they will blame meetings and email. Measure it and chat usually dominates, followed by self-interruption. That gap is why measurement comes before fixes.
Measuring Interruption Load
Interruption load can be measured without anyone counting taps on the shoulder, because interruptions leave a trace in how work sessions are shaped. Three measures cover most of it.
Focus block length. The distribution of uninterrupted sessions on a single piece of work. A team whose median focus block is twelve minutes has an interruption problem regardless of what anyone reports. Switches per hour. Application and context changes per hour of nominally focused time, with chat and email switches counted separately. Interrupted-block share. The proportion of planned focus blocks, such as calendar-marked deep-work time, in which the person was pulled into chat or a meeting.
All three are available from activity data, and all three are most useful at the team level and over weeks. Individual figures belong to the individual: the point is to give each person their own picture of how fragmented their day is, not to rank people by how interruptible they are. Interruption load is mostly a property of the environment, not of the person, and measuring it per person and then blaming them for it misses the point entirely.
Interruption Load, Last 4 Weeks
Median focus block length by week
Interruption sources
▲ Median focus block more than doubled after the team adopted a two-hour daily no-chat window in W2.
Illustrative eMonitor dashboard.
The Agreements That Reduce Interruptions
Personal fixes, blockers and headphones, help the individual and shift the load to colleagues. The fixes that work are agreements the team makes together, because most interruptions come from within the team.
A daily no-interruption window. Two hours, same time every day, in which chat is muted and nobody is expected to reply. This is the single highest-return agreement and the one teams most resist, because it feels unresponsive. Measured, it typically doubles median focus block length without any detectable effect on response times outside the window.
Explicit response-time expectations. Direct messages answered within two hours, not two minutes; channel messages by end of day; anything genuinely urgent uses a named urgent path. Most interruption pressure comes from an assumed expectation that nobody actually holds.
Batching. Questions collected and asked once, at an agreed time, rather than as they arise. This works particularly well for the most-interrupted experts: a daily fifteen-minute open slot replaces twenty scattered questions.
Notification defaults. Channels muted by default, keyword alerts only, tool notifications routed to a digest. This is a five-minute configuration change per person that most teams never make.
Meeting-free time. A half-day a week with no internal meetings, so that at least one long block exists for everyone. Our guide to deep work covers how to protect it.
Handling the Interruptions That Remain
Some interruptions are the job. Support teams, on-call engineers and managers exist partly to be interrupted, and a policy that pretends otherwise fails. For these roles the fix is not fewer interruptions but structured ones: rotation, so that one person absorbs the interruptions for a day while others focus; a triage step, so that what arrives is sorted before it lands on a person; and protected recovery time after an interruptive shift.
For everyone else, the useful discipline is to make interruptions cheaper rather than only rarer. A short written note of where you were before answering, a habit of finishing the sentence before looking up, and a rule that the interrupter states the urgency first all reduce the resumption cost even when the interruption count stays the same.
Measure the Load Before You Change the Rules
eMonitor shows focus block length, switches per hour and interrupted-block share per person and team, so no-chat windows and response-time agreements can be judged by their effect.
Running an Interruption Reduction as a Team
Four weeks is enough to see whether an agreement works. Week one: measure the baseline and share it with the team, sources included. Week two: agree one change, not five, and the no-interruption window is the usual first choice. Weeks three and four: run it, and re-measure. Then decide together whether to keep it, adjust it, or add the next agreement.
The team-level numbers are what get shared; individual figures stay with individuals. Presenting the baseline as “our median focus block is fourteen minutes” produces a shared problem. Presenting it as a per-person league table produces defensiveness and a quiet decision by the most interrupted people to stop helping colleagues, which is worse than the interruptions were.
Our guide to improving team productivity puts interruption reduction alongside the other structural changes, and the guide to time management tools covers the notification and focus tooling that supports the agreements.
What Success Looks Like
Not zero interruptions; that is neither possible nor desirable in a team that helps each other. Success is a median focus block long enough for the work the team is paid to do, a known and respected urgent path for the things that genuinely cannot wait, and a team that can say what its interruption load is and what it did about it last quarter. Teams that get there rarely go back, because the difference in what they can finish is obvious to everyone in them.