Remote Work6 min read
The Hidden Cost of Context Switching for Remote Teams
Refocusing after an interruption takes about 23 minutes. Here is the math for a remote workday and three protocols that recover real output.
By FlowQuest Editorial · 2026-04-28 · Updated 2026-04-28
The 23-minute number you may have heard before
In a widely cited line of research conducted at the University of California, Irvine, Gloria Mark and her collaborators measured how long it takes a knowledge worker to fully return to a task after an interruption. The figure that emerged from her observational studies of office workers is approximately 23 minutes and 15 seconds — that is the average time for a worker to return to the original task they were working on before being interrupted, accounting for the intermediate tasks they typically handle in between. The original empirical work appeared in studies including 'No Task Left Behind? Examining the Nature of Fragmented Work' (Mark, González, Harris, 2005) and was popularized in her later book 'Attention Span' (2023). The number is not precise to the second across every workplace and every kind of interruption, and the researchers themselves are careful about that. But the general result has held up across more than a decade of follow-up work: the cost of a single interruption is not the seconds it takes to read a message, it is the multi-minute tail of broken concentration and partial recovery that follows. For remote teams this number matters more than it did for the in-office workers Mark originally studied, because the remote workday has many more interruption surfaces and far weaker norms for protecting attention.
Doing the math for a remote workday
Take a generous remote workday. Eight hours on the clock. Subtract a lunch break and a couple of short context switches that are unavoidable. You start with roughly six hours that could in principle be productive deep work. Now apply realistic interruption rates. A typical remote worker on a chat-heavy team is interrupted, by a notification or an unscheduled message that demands a response, somewhere between eight and fifteen times per day. Take the conservative end: ten interruptions. If each interruption costs 23 minutes of refocus time, ten interruptions cost roughly 230 minutes — that is, just under four hours. Subtract those four hours from your six hours of nominal deep-work capacity and you are left with about two hours of actual focused output. The eight-hour day became a two-hour day. This is not a rhetorical exaggeration; it is roughly what happens to teams that ship slowly, run on heroics, and cannot explain where the time went. The interruption rate does not need to climb to twenty per day for the math to break. Ten is enough. The interventions that work are the ones that cut the interruption rate first and the recovery cost second.
Protocol 1: async-first communication
Async-first does not mean nobody talks. It means the default mode of internal communication is the message that does not demand an immediate reply, and the synchronous interruption is the explicit exception. In practice this looks like longer written messages, decisions made in threads instead of meetings, and an explicit team norm that says nobody is required to respond to a non-urgent ping inside a defined response window — typically four hours during work hours. Two cultural rules carry most of the weight. First, urgency is named, not assumed. If something is genuinely urgent, the sender says so out loud and uses a separate channel reserved for it. Everything else lives in the default channel and is read on the recipient's schedule. Second, replies in the default channel do not need to be instant, and senders do not interpret a delayed reply as disrespect. These rules sound obvious. They are also exactly the rules most remote teams quietly violate within a month of writing them down. The protocol works only when leadership models it visibly — and especially when leadership stops responding to non-urgent messages within thirty seconds, because that single behavior establishes the expected response speed for the entire team.
Protocol 2: focus blocks announced in calendar
A focus block is a calendar event placed on your own calendar that says, in effect, 'I am not available between 10 and 12.' The block is visible to teammates, marked busy, and treated with the same respect as a meeting with a customer. The block does not need to be defended in conversation; the calendar defends it for you. The protocol scales from individual to team in two steps. Step one: every individual on the team puts at least one focus block per day on their own calendar, ideally at a time that matches their personal energy curve. Step two: the team agrees on at least one shared focus block per week — typically a long block on the same morning for everyone — during which no internal meetings are scheduled, no internal pings are expected, and the channel goes quiet by convention. The shared block is the highest-leverage piece. It signals organizationally that deep work is real work, not a personal indulgence. It also creates a predictable window where the most cognitively demanding tasks of the week can land without negotiation.
Protocol 3: paired silence sessions
Paired silence is the practice of two teammates joining a video call together, briefly stating what each of them intends to work on, then muting and working in parallel for a fixed block — typically 50 minutes. Cameras can be on or off. There is no talking during the block. At the end, each person briefly reports what they got done. The mechanism is social: knowing another human is also working, on camera, in silence, raises the cost of getting up and switching tasks. It also externalizes the start ritual, which is one of the hardest parts of focus for ADHD users specifically. Paired silence is not a meeting. It does not require agendas, and it works best when the pairing is informal and rotates. For distributed teams across time zones it can be scheduled once or twice a week as an open optional session — anyone who shows up gets the benefit, anyone who does not is not required to. Three sessions a week is usually enough to materially raise the team's deep-work hours without adding a single new meeting.
Measuring whether it actually worked
Pick one number and watch it for a month. The cleanest is the count of days per week, per person, on which at least one continuous focus block of 90 minutes happened with no interruptions. Self-reported is fine. The number is small at first — often zero or one — and the goal is to move it to three or four. The protocols above will move it. Once the number is moving, the team's qualitative output usually changes faster than the metric does: shipping rate goes up, meeting count drops, and the question 'where did the day go' stops being the most common Friday afternoon thought.