| name | apply-single-tasking |
| description | Use when switching between tasks frequently, when you find yourself doing multiple things simultaneously, or when work quality feels inconsistent — to eliminate context-switching costs and recover the cognitive capacity lost to multitasking. |
| source | Mark, Gloria et al. "No Task Left Behind? Examining the Nature of Fragmented Work" (CHI 2005); Meyer, David E. et al. "Executive Control of Cognitive Processes in Task Switching" (Journal of Experimental Psychology, 2001); Newport, Cal. "Deep Work" (2016) |
| tags | ["focus","multitasking","context-switching","attention","cognitive-performance"] |
| verified | true |
Apply Single-Tasking
Work on one task at a time with complete attention, eliminating task-switching to recover cognitive performance lost to context fragmentation.
Why This Is Best Practice
Adopted by: Attention research at UC Irvine (Gloria Mark), MIT (Earl Miller), and University of Michigan (David Meyer) consistently demonstrates that multitasking is a myth — the brain does not process multiple cognitively demanding tasks in parallel. Cal Newport's Deep Work (2016) built the practitioner framework on this research. Single-tasking is explicitly recommended by Google's productivity programs and McKinsey's leadership development curriculum.
Impact:
- After an interruption, it takes an average of 23 minutes 15 seconds to return to the original task (Mark et al., 2005)
- Multitasking reduces IQ by up to 15 points — equivalent to missing a night of sleep (Hewlett-Packard study, 2005)
- Context-switching reduces productive output by 40% (Meyer et al., 2001)
- Heavy media multitaskers are worse at filtering irrelevant information than light multitaskers (Ophir, Nass & Wagner, Stanford 2009)
Why best: The brain has one executive attention system that serializes tasks. What feels like multitasking is rapid task-switching, which incurs a "switch cost" — a cognitive loading period every time focus changes. Single-tasking eliminates switch costs by holding attention on one context until a natural completion point.
Steps
Step 1: Understand attention residue
When you switch from Task A to Task B, part of your attention remains on Task A — this is attention residue (Sophie Leroy, University of Washington, 2009). Even if Task A is "done," unresolved questions about it leak into Task B's cognitive space.
Implication: interrupting yourself between tasks (to check Slack, glance at email) creates residue even when the check takes 30 seconds. The cost is measured in the quality of the resumed task, not the duration of the interruption.
Step 2: Define and isolate the current task
Before starting any work:
- Name the task in one sentence ("Write the API design doc for the auth service")
- Open only the applications and tabs required for this task
- Close everything else — email client, Slack, news, social media
- Set a physical or digital timer for 25–90 minutes
The isolation step is not optional — the presence of unread notifications creates anticipatory distraction even when you don't act on them (Stothart et al., 2015: receiving a notification, even without checking, reduces task performance).
Step 3: Use completion or natural stopping points to switch
The optimal time to switch tasks is at a natural stopping point — a completed section, a logical pause, a finished sub-task. Switching mid-sentence, mid-function, or mid-analysis maximizes residue.
If an interruption arrives (someone stops by, a message appears):
- If urgent: note exactly where you are ("line 47, debugging the null check") and then handle the interruption
- If not urgent: write it down for later and continue to the stopping point
The note removes the open loop from working memory, eliminating anticipatory residue.
Step 4: Batch similar low-focus tasks
Not all tasks require deep single-tasking. Routine tasks (email, Slack, admin) can be batched:
- Check email twice daily at fixed times (10am, 4pm)
- Process Slack messages in two 15-minute blocks rather than continuously
- Handle administrative tasks in one batch
Batching clusters similar cognitive modes together, reducing the switch cost between heterogeneous tasks throughout the day.
Step 5: Design the environment to support single-tasking
Willpower alone cannot sustain single-tasking in a notification-rich environment. Make distractions structurally harder:
| Intervention | Mechanism |
|---|
| Phone in another room or bag | "Out of sight" reduces urge frequency by 2× (Ward et al., 2017) |
| Website blockers (Freedom, Cold Turkey) | Remove volitional choice from distraction |
| Status set to "Do Not Disturb" | Signals unavailability without conflict |
| Single-app fullscreen mode | Removes visual cues for other apps |
| Headphones (even without audio) | Social signal for focus; reduces interruptions |
Rules
- One task open at a time — close everything else before starting
- No notifications during task sessions: phone silent, Slack/email closed or DND
- Switch only at natural stopping points — never mid-task
- Write down interruptions rather than acting on them immediately; process after the session
- Batched reactive work is not multitasking — it is scheduled serial processing
Common Mistakes
Keeping Slack/email open "just in case": The presence of the unread badge impairs task performance even without checking. Close the app.
Calling rapid task-switching "multitasking": The brain does not multitask; it switches rapidly and pays the residue cost on every switch. Recognition that it's happening is the first step to stopping it.
Treating single-tasking as only for deep work: Shallow tasks also benefit from single-tasking. Processing email with a Zoom call in the background produces worse email and worse Zoom presence.
Believing you are the exception: High multitaskers consistently overestimate their multitasking ability. The Stanford study found that heavy multitaskers performed worse on every metric — they were not skilled at it; they were simply addicted to it.
When NOT to Use
- Physical tasks that don't require cognitive load can run concurrently (walking + podcast, cooking + audiobook)
- Monitoring tasks (waiting for a build, watching a deployment) allow light parallel reading — not cognitively demanding tasks
- Pair programming is single-tasking on the same task with two people — not an exception to this skill
