Written by Dr. Leo Maren · Co-founder of Fwocus · Behavioural scientist · PhD, Imperial College London
You have probably read that it takes 23 minutes to refocus after an interruption. The number is real in one sense: a researcher said it in an interview. But it does not measure what most articles say it measures. This page explains what the research found, and how I use a 23-minute block because of it.
Where "23 minutes to refocus" comes from
The number comes from Gloria Mark, a professor of informatics at the University of California, Irvine. In a 2006 Gallup interview, she said that most interrupted work was resumed on the same day, and that it was resumed after an average of 23 minutes and 15 seconds. She added that people did about two other tasks before they returned.
I repeated the "23 minutes to refocus" line in a lab presentation during my PhD. A colleague asked me for the paper. I could not find the number in any paper, and I felt a little foolish. It was a useful lesson about my own field. Digital cognition has many numbers that travel faster than their sources.
What the research measured
The closest published source is Mark, Gonzalez and Harris (2005). The researchers observed 24 managers, financial analysts and software developers at an IT services company. A researcher followed each person with a watch and a notepad for about three working days. The paper reports more than 700 hours of formal observation.
The main results:
People spent about 11 minutes on one project before they switched to another. The paper calls these projects "working spheres."
57% of these work periods were interrupted.
77.2% of interrupted work was resumed on the same day.
When it was resumed, the average gap was 25 minutes 26 seconds, with a very large spread (standard deviation of 54 minutes 48 seconds).
Before they returned, people worked on an average of 2.26 other projects.
The interview figure of 23 minutes 15 seconds is close to the paper's 25 minutes 26 seconds, but it is not the same. I do not know which analysis produced the interview number.
What the research does not show
The 25-minute gap is time until return, not time to refocus. During those minutes, people were not staring at the screen and trying to concentrate. They were doing other work: a phone call, an email, a second project. The study did not measure how long it took to get back to full concentration after they returned.
Three more limits are important:
The average hides a wide spread. A standard deviation of about 55 minutes means some gaps were very short and some were hours. "Average" does not mean "typical" here.
The sample was small and specific. Twenty-four people at one company in the early 2000s, before smartphones and team chat.
Many interruptions came from the workers themselves. The study counted internal interruptions, where a person stops on their own. For analysts and developers, internal and external interruptions were about equal.
So the honest statement is this. In one observational study, interrupted work often waited more than 20 minutes and two other tasks before people came back to it. That is still a strong finding. It is just a different one.
What happens to your attention after a switch
Other research looks at the cost inside your head. Leroy (2009) described attention residue: part of your attention stays on the task you left. In her experiments, people who left a task unfinished performed worse on the next task. The residue was stronger when the first task was not complete.
A later lab study by Mark, Gudith and Klocke (2008) found a surprise. Forty-eight participants answered emails while the researchers interrupted them every few minutes by phone or instant message. The interrupted groups finished faster, with no loss in quality. But they reported more stress, frustration, time pressure and effort.
I recognize both results in my own work. After a Slack message, I rarely lose 23 minutes. I lose the thread of the argument, and I work faster and more tensely to get it back. The time cost is small. The cost in tension is what builds up over a day.
Why a 23-minute block works as a defense
A 23-minute block is short enough that most messages can wait until the end. That is the main reason I use it. If someone needs me, 23 minutes is rarely a problem. If I stop to answer, I start the chain of other tasks that the 2005 study observed.
The number 23 has no special power. I chose it as a reminder of the research, and it is a little shorter than the classic 25-minute Pomodoro. You can use the 25 minute timer or the 20 minute timer for the same method.
How I run a 23-minute block
Type the task in the task line. Make it a specific action, such as "Draft section 3, paragraph on sampling." The line is not saved, but it stays visible while the timer runs.
Silence the sources you control. Close email and chat, or set them to do-not-disturb for 23 minutes.
Keep a paper note for self-interruptions. When you think "I must check the invoice," write "invoice" and continue. This is my answer to the internal interruptions in the 2005 study.
If someone interrupts you, leave a marker first. Type a short note at your cursor, such as "NEXT: add Figure 2 reference." Mark and colleagues noted that a visible cue, like the cursor at the last word, helps you reorient when you come back.
At the chime, process the paper note. Then decide if you start another block.
The step that changed the most for me was step 4. Before, I answered interruptions immediately. After, I spent ten seconds on a marker. When I returned, the marker told me the next action, and I did not need to reread three pages to find my place.
Timer controls
Press Space to start, pause and resume. Press R to reset, and F for fullscreen. Fullscreen hides other tabs, which removes one cause of self-interruption.
The − and + buttons change the length by 1 minute, before or during a session. The Up and Down arrow keys do the same. The chime at zero is soft by default. If you work with headphones and need a quiet signal, the silent timer has no sound at all.
The browser tab title shows the time left. If you must check another tab for your task, you can still see how much of the block is left.
FAQ
Does it really take 23 minutes to refocus after an interruption?
No study shows that exact claim. Gloria Mark's 2005 study found that interrupted work was resumed after about 25 minutes on average, and that people did about two other tasks first. That measures the time until people came back to a task, not the time it takes to concentrate again.
Is the 23-minute figure made up?
Not exactly. Gloria Mark gave the figure of 23 minutes 15 seconds in a 2006 Gallup interview, and it is close to the 25 minutes 26 seconds in her 2005 paper. The problem is how people describe it, not the number itself.
Is a 23-minute block better than a 25-minute Pomodoro?
There is no evidence that 23 minutes is better than 25. Choose a length that you can protect from interruptions most days. If you often get pulled away before 25 minutes, a shorter block is a good start. The ADHD focus timer has more on starting with short blocks.
Can I use this timer for the GRE Verbal section?
Yes. The second Verbal Reasoning section of the GRE is 23 minutes for 15 questions, according to ETS. For pacing advice on timed GRE practice, see the 21 minute timer, which covers the first Quant section.