Context Switching: What It Costs and the 23-Minute Myth

Context Switching: What It Costs and the 23-Minute Myth

Context switching is moving your attention from one task to another, and every switch has a cost: slower, less accurate work right after the transition, plus leftover thoughts about the task you left behind. The famous "23 minutes to refocus" stat is real data, but it measured how long people took to get back to an interrupted task, not how long refocusing takes. The term comes from computing, where a CPU saves one process's state and loads another's. Your brain does the same thing, minus the save file.

Key takeaways:

• The 23-minute figure comes from field observation of 24 workers. It measures time until they returned to an interrupted task, not time spent in a fog.
• The "40% productivity loss" is a researcher's estimate, not a measured outcome from any single study.
• In one study of 137 workers, people toggled between apps about 1,200 times a day, adding up to nearly four hours a week reorienting.
• The best-tested fix for attention residue is a short "ready-to-resume" note before you switch.

The context switching numbers, and what each one actually measured

What is context switching?

In computing, a context switch saves one process's registers and loads another's. Clean, reversible, predictable overhead. The human version loses the save file. When you leave a half-written email for a Slack thread, your brain stores a fuzzy, emotionally tagged impression, not a snapshot, and part of your attention stays with the unfinished email whether you want it to or not.

Mark's 2005 field study found that a large share of interruptions were self-interruptions, where the worker opened a new task unprompted: about half for analysts and developers, and 40.8% for managers. Self-interrupted work took longer to resume (29 min 1 s versus 22 min 37 s for externally interrupted work, p < .055, a trend). You are your own worst interrupter.

Does it really take 23 minutes to refocus after an interruption?

Not the way people quote it.

Gloria Mark and colleagues observed 24 information workers (managers, financial analysts, software developers) at one company, shadowing each person for about three days, over 700 hours total. People spent an average of 11 minutes and 4 seconds in a working sphere before switching. 57.1% of segments were interrupted. Of those, 77.2% were resumed the same day, taking an average of 25 minutes and 26 seconds (sd 54 min 48 s), after working in 2.26 other working spheres first.

The rounder "23 minutes 15 seconds" comes from a 2006 Gallup interview where Mark summarized a slightly different data cut (81.9% resumed same day). She noted people did about two intervening tasks before returning.

The number describes how long the interrupted task sat waiting, not a 23-minute fog after every ping. The standard deviations were enormous (54 min 48 s on a 25 min mean), meaning some people resumed in minutes and others took hours. An average this noisy is not a universal law.

What a single switch costs in the lab

Rubinstein, Meyer, and Evans (2001) ran four task switching experiments where participants alternated between arithmetic and geometric classification. Per-switch costs were small (fractions of a second to a few seconds), scaling up with rule complexity and down with task cueing.

The damage is in the count. A 2022 study by Murty and colleagues in Harvard Business Review tracked 137 workers for five weeks: nearly 1,200 app toggles per day. Including the reorientation time after each switch, that added up to just under four hours per week, roughly 9% of annual work time.

The "40% of productive time lost" figure is Meyer's estimate quoted by the APA, not a measured outcome. When people ask how much context switching hurts productivity, that number comes up, but treat it as a back-of-the-envelope characterization, not data.

What is attention residue?

Attention residue is the piece of your mind that stays with Task A after you move to Task B. Sophie Leroy named the concept in a 2009 study (Organizational Behavior and Human Decision Processes): when people leave a task unfinished, part of their attention remains stuck on it, and performance on the next task suffers.

Across four studies in Leroy and Glomb (2018), published in Organization Science, a short "ready-to-resume" plan (write down where you were and what you will do next) reduced residue and protected performance on the interrupting task. One sentence, written in seconds. Most articles about the "23-minute problem" never mention it.

Interrupted people work faster, and pay for it in stress

Mark, Gudith, and Klocke (2008) ran 48 participants through an email task with and without interruptions. Interrupted people finished in less time with no drop in quality.

The catch: after only 20 minutes of interrupted performance, participants reported higher stress, frustration, workload, effort, and time pressure. Switching does not just eat clock time. It eats bandwidth, and you compensate by speeding up until the stress catches you. The cost of context switching is not only minutes lost; it is the stress of running faster to stay in place.

How to reduce context switching: a batching plan by role

Engineers

• Block one or two 90-to-120-minute no-meeting windows. Batch code review into two daily sessions instead of reviewing each PR as it arrives.
• Before any switch, write a one-line "next step" in the ticket. This is the ready-to-resume plan.
• Turn off Slack during deep work blocks.

Managers

• Cluster meetings into half-days. Give your team a stated response window (e.g., "I check Slack at 10 and 2").
• Keep one protected block for writing. A time-blocking template built around your energy helps enforce it.

Sales and support

• You cannot batch inbound, but batch outbound and admin into fixed blocks.
• Use cued switches (a fixed queue order) per Rubinstein et al. (2001); they cost less than uncued jumps.
• Between back-to-back calls, a 60-second reset can help keep residue from bleeding over.

For everyone

• Close the tab you keep checking. Self-interruptions were roughly half of all interruptions for the analysts and developers in Mark's data. If you feel overstimulated and cannot focus, it might not be discipline; you may have toggled 600 times by lunch.

Context switching: what to keep and what to drop

The cost of context switching is real. The per-unit task switching cost is small, but 1,200 toggles a day turns small into substantial. The 23-minute figure is real data, but it describes wait time, not fog time. The 40% number is not data.

What is well supported: attention residue makes your next task worse when you leave the previous one unfinished under pressure. The ready-to-resume plan reduced that residue across four studies. It is free, it takes seconds, and it held up across four studies. Fewer switches, planned transitions, and a one-line note at the boundary. That is the evidence-based playbook.

Frequently Asked Questions

What is context switching in simple terms?

Context switching means stopping one task and starting another. The term comes from computing, where a CPU saves one program's state and loads another's. For humans, the cost is a brief delay on every switch plus "attention residue," leftover thoughts about the task you just left. More switches per day means more accumulated cost.

How long does it take to refocus after an interruption?

The popular "23 minutes" comes from a 2006 Gallup interview summarizing Gloria Mark's observation of 24 workers. It measures how long the interrupted task waited before the person returned, not how long refocusing took. People did about two other tasks in between. Per-switch refocusing in lab settings takes seconds, not minutes.

Is multitasking the same as context switching?

They overlap but are not identical. Multitasking usually means trying to do two things at once, which the brain cannot do for attention-demanding work. Context switching is the transition between tasks, one at a time. Both carry costs. See why multitasking fails in the brain for a deeper look.

What is attention residue?

Attention residue is the portion of your attention that stays with an unfinished task after you move on. Sophie Leroy identified and named it in 2009. It is worst when you leave a task unfinished under time pressure. A "ready-to-resume" plan (writing down where you stopped and what you will do next) reduces it.

Is the 40% productivity loss from context switching real?

It is a researcher's estimate, not a lab result. David Meyer has been quoted by the APA saying productivity can drop by as much as 40%. No single study measured that number. The ~1,200 daily toggles and ~4 hours/week reorienting from the HBR 2022 study are more concrete.

How do you reduce context switching at work?

Three evidence-backed moves: (1) Batch similar tasks into blocks instead of scattering them. (2) Use cued transitions (a fixed order or checklist), because cued switches cost less than uncued ones. (3) Write a one-line "ready-to-resume" note before every switch. Turning off non-urgent notifications during focus blocks helps with self-interruptions.

The Switch You Can Plan For

Each switch is cheap; the accumulation is not. Switch fewer times and handle each transition better: a ready-to-resume note, a protected block, a cue that says this window is for one task only.

That cue can be physical. A 2025 meta-analysis (Payne et al., Nutrition Reviews) found L-theanine plus caffeine improved attention-switching accuracy versus placebo (SMD 0.33), and Giesbrecht et al. (2010), n=44, found better task-switching and alertness at 97 mg L-theanine + 40 mg caffeine. Giesbrecht's authors were Unilever-affiliated; Payne's lead author's PhD is funded by Lipton Teas and Infusions, and two authors work for Lipton or Unilever. The evidence covers the combination, not Roon.

Roon is a zero-nicotine pouch built on that pairing: 80 mg caffeine, 60 mg L-theanine, 25 mg methylliberine (Dynamine), and 5 mg theacrine (TeaCrine). In an n=12 crossover PK study (Mondal et al., 2022), methylliberine cut caffeine's oral clearance from 41.9 to 17.1 L/hr and extended its half-life from 7.2 to 15 hours, with no higher peak. That study used 100 mg methylliberine + 150 mg caffeine, 4x Roon's methylliberine and about 1.9x its caffeine, so only the direction carries over: same peak, longer tail, a 6-to-8-hour felt window. Stop at least 8 hours before bed.

One pouch at the start of a protected block gives you a sustained window to use it. From $9.16 per tin on a 12-tin subscription. Try Roon for your next focus block.

Related from Roon

• The L-theanine and caffeine task-switching trials
• Deep work vs shallow work
• A time-blocking template built around your energy
• How to reset between back-to-back meetings

Written by Roon Team