Task switching cost is the measurable loss in speed and accuracy that occurs each time you move between two tasks. It has been measured in laboratories since the 1990s, in setups as simple as alternating between naming colours and naming shapes. The switch itself takes time, independent of either task.
In a lab the cost is a few hundred milliseconds. In a working day, with tasks that carry context, it is minutes.
Why it is not the same as attention residue
The two terms get used as synonyms and describe different halves of the problem.
Task switching cost is the price of the transition. Your brain reconfigures: different rules, different goal, different working memory contents. That reconfiguration takes time, and it happens even for trivial task pairs.
Attention residue is what leaks across the transition. Part of the old task remains active in the new one. It is heaviest when the old task was left unfinished. The full definition is in attention residue.
You pay both, every switch. One is the cost of arriving; the other is the cost of not fully having left.
Why phones are the worst case
Three properties stack up.
The switches are frequent. Not scheduled, and often not noticed.
The tasks are maximally different. Moving from a spreadsheet to a video feed is a larger reconfiguration than moving between two spreadsheets.
The old task never completes. A feed has no ending, so it leaves the heavy kind of residue every time.
That combination is why a day with thirty small checks can feel much harder than a day with two long meetings, despite the meetings taking more clock time.
What it does not mean
It does not mean multitasking is impossible or that any interruption is damaging. Switching is normal and often necessary; a job that never permits it is rare. It also does not mean the cost is fixed. Practised, predictable switches are cheaper than novel unpredictable ones.
And it does not mean the answer is to work in four hour blocks with the phone in another room. That advice fails for most people, for the reasons in strict mode made it worse.
If the answer you want is a timer that protects a block of work, the category and its limits are covered in focus timer apps.
What actually helps
Batch the switches. Ten checks at three moments cost less than ten checks scattered across an afternoon, because you pay the reconfiguration fewer times.
Make the switch visible. The expensive switches are the ones nobody decided to make, described in autopilot scrolling and I picked up my phone without deciding to. A switch you noticed is a switch you can price. One you did not notice arrives with no receipt.
Give the old task an ending. Even a token one. A moment of contact before moving on lowers what leaks across.
reCalm makes the invisible switches visible. The apps you shield open onto one breath, which is a receipt for the switch you were about to make without noticing. Batching, endings and noticing all get easier once the reach itself has an edge.
Frequently Asked Questions
What is task switching cost?
The time and accuracy lost each time you move between two tasks, over and above the time either task requires. It is measurable in laboratory conditions at a few hundred milliseconds, and much larger for real work.
Is task switching the same as multitasking?
Mostly. What is usually called multitasking is rapid switching, since attention does not genuinely split across two demanding tasks. The costs are switching costs.
How much does context switching cost at work?
Estimates vary too much to quote a single figure honestly. The direction is well supported: frequency of switching predicts loss better than the total time spent switched away.
How do I reduce task switching cost?
Reduce the number of switches rather than their length, batch them into fewer moments, and make unnoticed switches noticeable so they become decisions rather than reflexes.