Summary
In 2004, Gloria Mark and her colleagues at the University of California, Irvine, shadowed office workers with stopwatches and found that people stayed on one screen for about two and a half minutes before switching. By 2012 it was 75 seconds. Since around 2016 it has been about 47 seconds, and when other researchers repeated the measurement with computer logging they got a median of 40. That is not a story about a generation that cannot concentrate. The same people, given a quiet hour, focus perfectly well. It is a story about an environment that interrupts them every couple of minutes and a set of habits, half of them self-inflicted, that fill the gaps. This piece goes through what the research found, what a team's own activity data shows, and the six changes that reliably get focus back.
E-Monitor shows focus blocks, switch counts and interruption patterns by team, so you can see where attention is going before trying to fix it. Book a demo →
What the Research Actually Found
Mark's finding is often reported as if human attention spans had shrunk. Her own explanation is more precise and more useful. The measurement is of how long people stay on a single screen before moving to another, and the fall from two and a half minutes to 47 seconds tracks the arrival of email on every desk, then chat, then the phone that is never quite off. People did not lose the ability to concentrate. They gained a dozen reasons per hour to stop.
Two other findings from the same body of work matter for anyone managing a team. About half of the switches are self-interruptions: nobody pinged, the person just reached for the inbox. And after an interruption, getting back to the original task takes a long time, with the widely cited figure being around twenty-three minutes, though Mark herself has said that number is often quoted more precisely than the data supports. The honest version is that recovering from a switch costs minutes, not seconds, and the cost compounds across a day of switches.
Microsoft's Work Trend Index adds the volume. Its telemetry across Microsoft 365 finds an interruption, meaning a meeting, an email or a message, every two minutes on average during the working day, and the most interrupted hour, 11am, is also when most people are at their natural cognitive peak.
What It Looks Like in Activity Data
If you have monitoring data, you can see the attention problem directly rather than taking the research on trust. Three figures tell the story.
Application switches per hour. A focused hour has a handful. A fragmented one has thirty or forty, and most knowledge workers sit between fifteen and twenty-five. The number is visible in application tracking and it is the closest proxy for Mark's stopwatch.
Focus blocks. Stretches of twenty-five minutes or more on one application. Count them per day. Teams that report feeling scattered usually show two or three; teams that feel in control show six or more. The productive hours guide gives the ranges by role.
Where the blocks break. Overlay the switch timeline with chat and email activity. If most focus blocks end with a chat window, the interruptions are coming from colleagues. If they end with a browser tab, they are self-inflicted. The fix is different in each case.
The point of looking is to replace argument with evidence. A manager who believes the team is distracted and a team that believes it is overloaded are usually both right, and the data shows how.
It Is the Environment, Mostly
The temptation is to treat short attention as a personal failing and respond with training, or worse, with monitoring that counts switches and reports them. That gets the causation backwards. The same employee whose screen attention is 40 seconds during a day of stand-ups, pings and a shared inbox will hold a document for an hour on a Friday afternoon when the calendar is empty. The capacity is intact. The conditions are not.
The conditions are set by the organisation. Response-time expectations on chat decide whether a message can wait. Meeting placement decides whether the best hours are available. Tool sprawl decides how many windows a task needs. Notification defaults decide whether the phone buzzes. Every one of those is a decision somebody made, usually by not making it, and every one can be changed by a manager without any employee changing their character. The context switching guide puts a cost on each.
Six Changes That Protect Attention
These are the interventions that show up as more and longer focus blocks in the data within a few weeks. None of them requires willpower from the people whose attention is being protected.
- Agree a chat response time. "Within two hours" turns chat from a pager into a queue. The single biggest change in most teams
- Move meetings off the peak. Late morning is the worst time for a stand-up and the best time for the hard task; the peak hours guide shows how to find yours
- Protect one block a day, two hours, status set to away, calendar blocked, measured afterwards so it does not quietly fill with meetings
- Batch the inbox into two or three sittings; the email itself takes the same time, the fragmentation does not
- Cut the tool count. Every application a task touches is a switch; an integration that removes a copy-paste step removes dozens a day
- Turn off the defaults. Badge counts, desktop banners, sound. People rarely re-enable them
Teams that do all six typically double their daily focus blocks inside a month, with no change in hours worked. The meeting-free days guide covers the larger version of the third item, and the deep work guide the longer-term practice.
The Self-Interruption Half
Half the switches are nobody's fault but our own, and that half responds to a different approach. It is not discipline, which fails by Wednesday. It is friction and visibility. Closing the email client between sittings adds just enough friction that the reflexive check does not happen. A phone in a drawer rather than face-down on the desk does the same. And seeing your own switch count on a dashboard, which employees with self-view can do, turns an invisible habit into a number that is oddly motivating to bring down.
This is the one place where monitoring data serves the employee more than the manager. A manager should never be counting an individual's switches. An individual who can count their own usually wants to.
A Worked Example: The Team With No Afternoon
A nine-person analytics team came to its lead convinced it was underwater. Hours were normal. Output was slipping. The activity data showed switch counts averaging thirty-one an hour and focus blocks averaging 2.4 a day, nearly all of them before 10am. After that the day was dust: a stand-up at 10, a shared inbox that chimed on every ticket, a Slack channel where the convention was to answer within minutes, and three recurring syncs that between them touched every afternoon.
The lead made four changes in one week. Stand-up moved to 12:30. The inbox got a rota so that one person watched it per hour instead of nine. Slack got a stated two-hour response expectation, written in the channel description. Two of the three syncs merged. A month later, switches were down to fourteen an hour, focus blocks were up to 5.7 a day with most of them now in the afternoon, and the backlog that had been growing for a quarter was shrinking. Nobody had worked a minute longer. The attention had always been there; it had been spent on each other.
Frequently Asked Questions
1. What is the average attention span at work?
About 47 seconds on a single screen before switching, according to Gloria Mark's research at UC Irvine, down from two and a half minutes in 2004. Other researchers using computer logging found a median of 40 seconds. It measures switching in an interrupting environment, not the ability to concentrate.
2. Has attention span really decreased?
Screen attention has, measurably. Underlying concentration has not: the same people focus well when interruptions stop. The fall tracks the arrival of always-on email, chat and phones, and about half of all switches are self-initiated rather than caused by someone else.
3. How long does it take to refocus after an interruption?
Minutes rather than seconds, and it compounds across a day. The commonly quoted 23 minutes comes from Mark's early work and is often stated more precisely than the data supports; the practical point is that each switch has a recovery cost.
4. How can managers improve team attention span?
Change the environment rather than the people: agree a chat response time, move meetings off the late-morning peak, protect a daily focus block and measure it, batch email, reduce the number of tools a task touches, and switch off notification defaults. These show up as more and longer focus blocks within weeks.
5. Can monitoring data measure attention?
It can show application switches per hour, focus blocks per day and what ends them, which together are a close proxy. Use it at team level to diagnose the environment and give individuals their own view; never report an individual's switch count to a manager.
6. Is it the employee's fault if they cannot focus?
Mostly no. Switch rates are set by chat norms, meeting placement, tool sprawl and notification defaults, all organisational decisions. The self-interruption half responds to friction and self-visibility, not to discipline or surveillance.
See where the attention goes E-Monitor shows switch counts, focus blocks and what interrupts them by team, and gives each employee the same view of their own, so the environment gets fixed rather than the people. Sign up →