Single-Tasking: The Case Against Multitasking (And What to Do Instead)

Multitasking is listed as a skill on résumés.

It should not be. The research on multitasking is one of the most consistent bodies of evidence in cognitive psychology: humans cannot genuinely do two cognitively demanding tasks simultaneously. What we call multitasking is rapid serial switching — rapidly shifting attention between tasks, paying a cost each time, and doing both tasks worse than if we had done them one at a time.

The belief in multitasking as a productivity strategy is one of the most expensive cognitive errors in modern work life.

The Neuroscience: Why Multitasking Is Impossible

The brain has a bottleneck in its attention system. Only one stream of information can receive full processing resources at a time. When two tasks both require conscious attention, the brain must switch between them — it cannot fully process both simultaneously.

This is not a limitation that can be overcome with practice. It is a structural feature of how the prefrontal cortex allocates attention. The small number of people who appear to multitask effectively have been shown in controlled studies to not be doing genuine simultaneous processing — they are rapid switchers with shorter context-switch costs, not parallel processors.

The Switching Cost

Every time your attention moves from one task to another, there is a transition cost: the previous task’s context needs to be stored, the new task’s context needs to be loaded, and the cognitive system needs to reorient. This takes time — typically 0.5–5 seconds for simple tasks — and it produces what researcher Sophie Leroy calls “attention residue”: a portion of your cognitive resources that remain focused on the previous task even after you have nominally switched to the new one.

Attention residue means you are never fully present with the new task immediately after switching. The first 1–15 minutes after a context switch, you are working with diminished cognitive capacity.

In a workday full of constant switching — email check, task work, message notification, task work, phone call, task work — the attention residue accumulates. By mid-morning, your effective cognitive capacity is a fraction of what it would be without the switching.

The Error Rate Increase

Studies consistently show that task-switching increases error rates by 20–40% compared to single-tasking. This is not because people are less skilled — it is because errors occur disproportionately at the points of transition, when attention residue from the previous task interferes with the current one.

For knowledge work where quality matters, this error rate increase means rework, missed details, and lower output quality — all of which cost more time than the supposed efficiency gain of doing two things “at once.”

Forms of Multitasking to Eliminate

Background Notifications

Every notification is a context switch. Even when you do not respond to a notification — even when you merely glance at it — your attention has moved, attention residue has been created, and you will need time to return to full engagement with what you were doing.

Research by Gloria Mark at UC Irvine found that after a task interruption, it takes an average of 23 minutes to fully return to the original task. Even brief, chosen interruptions (not forced interruptions) take several minutes to recover from.

The fix: notifications off during any focused work period. Not on silent — off. The awareness that notifications might be arriving is enough to maintain some background monitoring that reduces available attention.

Monitoring Multiple Screens

Having email, Slack, or social media open in a background window while working — even without actively checking — imposes a monitoring cost. Part of your peripheral attention is watching for changes in those windows.

The fix: close everything that is not needed for the current task. Not minimized — closed.

Media Consumption During Cognitive Work

Background music without lyrics is mostly neutral for cognitive work. Podcasts, TV, videos, or music with lyrics during cognitively demanding tasks actively impairs performance. Language-processing systems compete for the same cognitive resources as reading and writing. You cannot fully process the podcast and fully process the document — one or both suffers.

The fix: silence or instrumental music during demanding cognitive work. Reserve podcasts and spoken media for genuinely low-demand tasks (exercise, household chores, commuting).

Parallel Task Management

Managing many projects simultaneously — keeping all of them in progress at the same time — produces the cognitive version of multitasking over a longer time scale. The context-switching between projects as you move between them throughout the day or week creates the same residue and overhead as within-task switching.

The fix: batching (work on one project for a full day or half-day before switching) and project queuing (finish or reach a defined resting point before picking up a new one).

Single-Tasking in Practice

The One-Tab Rule

When doing focused work, have only the tabs and applications open that are directly necessary for the current task. If you need Google Docs, have Google Docs open. Not Google Docs plus email plus Slack plus two research tabs you might need.

Opening additional resources as needed is slightly slower than having them pre-loaded. The cognitive cost of the pre-loaded distraction is much higher than the friction of opening them on demand.

The Task Commitment Card

Before starting a work session, write down the one thing you are working on for this session. Put it somewhere visible. Everything that is not that thing is off the table until the session ends.

This sounds trivially simple. It is not — because the common failure mode of work sessions is the drift: you are working on the report, you think of something to check, you open a new tab, you are now working on something else without having made a conscious decision to switch.

The task commitment card creates a decision point. To switch tasks, you have to consciously choose to switch — and see that you are choosing to switch — rather than drifting without awareness.

Protecting Transition Time

Rather than switching from one task to another immediately, build a short transition buffer between tasks: 2–5 minutes to close out the previous context (notes, status, what comes next), let your mind rest briefly, and then orient to the new context before beginning.

This feels like it wastes time. In practice, it significantly reduces error rates and warm-up time for the new task. The transition buffer means you arrive at the new task with less residue and reach full engagement faster.

The Input Queue

Interruptions — questions from colleagues, unexpected requests, new ideas mid-task — are a major source of involuntary context-switching. The standard response is to handle them immediately, which breaks focus.

The alternative: maintain an input queue. When something new arrives while you are in a focused task, write it down and return to the task. Handle the queue between tasks or in a designated review window.

This requires communicating to people around you that they will get a response, just not immediately — which requires setting expectations and being consistent about honoring the queue. Over time, the people you work with learn to expect a prompt but not immediate response, which is a reasonable norm.

Multitasking That Is Actually Fine

Not all forms of “doing two things at once” are cognitively costly.

One cognitively demanding task + one automatic task: Walking while thinking, listening to music while doing a repetitive manual task, exercising while listening to a podcast. These combine a demanding task with one that is fully automatic and requires no cognitive resources. The automatic task has no bottleneck claim on your attention.

Two automatic tasks: Most household tasks can be combined. Walking + podcasts. Cooking + audiobooks. Commuting + planning your day mentally. These are fine.

The general principle: genuine multitasking harm occurs when two tasks both require active cognitive resources. One demanding + one automatic is parallel but not competing. Two automatic is fine. Two demanding is the problem.

The Output Calculation

Here is a useful way to see why single-tasking beats multitasking on pure output:

Multitasking scenario:

• 2 hours spent alternating between Task A and Task B
• Each context switch costs 5 minutes of reduced effectiveness
• 20 context switches in 2 hours = 100 minutes of degraded attention
• Effective focused time: ~60 minutes, distributed incoherently
• Error rate: elevated throughout

Single-tasking scenario:

• 1 hour on Task A, 1 hour on Task B
• 1 context switch between them: 5 minutes transition cost
• Effective focused time: ~115 minutes at full capacity
• Error rate: normal

Same clock time. Significantly different output quality and quantity.

Building the Single-Tasking Habit

The multitasking habit is deeply entrenched for most people — the constant availability of distraction has trained rapid switching into something that feels normal and even virtuous.

Rebuilding the single-tasking habit takes deliberate practice:

Week 1: One 25-minute single-tasking session per day. One task, everything else closed, for 25 minutes.

Week 2: Two 25-minute sessions. Extend to 35 minutes if the first week was comfortable.

Week 3: Morning block of 90 minutes single-tasking on your most important work.

Ongoing: Default to single-tasking for all focused work. Multitasking only for explicitly automatic tasks.

The discomfort of single-tasking in the early weeks is not the difficulty of the work — it is the withdrawal from the constant stimulation that multitasking provides. That discomfort passes. The resulting focus capacity does not.

Single-tasking is the execution layer of the Deep Work system. Read the Deep Work Guide for the full practice — single-tasking is what happens inside each deep work session.