Before you use it
- Use when
- A batch is efficient locally but one waiting item may block a person, decision, shipment, repair or downstream task.
Keep this limit in mind
- The comparison is qualitative unless you have useful data. Do not invent precision or let every requester define their own item as costly to delay.
Current editorial reading
A delayed answer can be the expensive part
Batching saves repeated setup, but waiting also has a cost. Operations research makes this trade-off unusually visible. Karmarkar showed in batch manufacturing models that lot size, queueing and lead time are linked, and that larger batches can create costs through longer lead times and work in process. Knowledge work is not a factory, so this is an analogy rather than direct evidence for inbox or project decisions. The useful transfer is the system view: local efficiency is incomplete if it leaves other work waiting.
People can trade switch cost against delay
Mittelstädt and colleagues tested voluntary task switching in a controlled environment where repeating the same task gradually required waiting longer for the next stimulus. Participants tended to switch when that environmental delay roughly matched their own switch cost, and they improved local task performance by adapting to the trade-off. The task was a laboratory paradigm, not a software team or support queue. Still, it directly challenges a rigid “never switch during a batch” rule: the better action can change when the cost of staying put rises.
The switch cost is broader than elapsed time
A switch can be worthwhile and still be disruptive. In an experiment by Mark, Gudith and Klocke, interrupted participants completed the assigned work faster without a detected quality loss, apparently by working more quickly. They also reported more stress, frustration, time pressure and effort. That result is useful counter-evidence to a simple productivity story. Do not compare only two clock times. A justified exception should consider the downstream wait it removes and the recovery load it creates for the person who switches.
If you break the batch, make the return cheap
Trafton and colleagues found that participants who received an eight-second warning before an interruption used that interval to prepare and resumed the interrupted task faster than participants interrupted immediately. Practice reduced some of the disadvantage in the immediate condition, so the effect is not a fixed tax. For deliberate exceptions, the practical implication is modest: before switching, leave enough state to restart — the current point, the unresolved question or the next action — then answer only the blocking part and return. A restart cue does not make an interruption free; it reduces how much context must be reconstructed.
What is a strong enough reason to interrupt now?
Look for an observable consequence of waiting until the planned batch boundary. Strong candidates include several people unable to continue, a deadline that will genuinely be missed, a long external process that cannot start until you act, or a decision whose delay creates a larger queue. A sender saying “urgent” is not evidence by itself. The exception is strongest when you can name what is waiting, for how long, and what your small action releases.
Is “I am blocked” enough reason?
No. Check the blocker rather than the label. The other person may have another task, a workaround, or a response window that makes the delay harmless. Conversely, one short clarification may really hold up several dependent tasks. Ask what stops, when it matters, and whether a partial answer is enough. This keeps the fast path based on delay cost instead of whoever asks most forcefully.
What if I cannot estimate either cost?
Do not invent a score. Use a reversible default: keep ordinary work in the batch, but allow a short exception when the downstream consequence is concrete and near-term. Keep the response narrow, leave a restart cue, and notice what happened afterward. If exceptions repeatedly save little or regularly dissolve the whole focus block, tighten the rule. If planned batches repeatedly leave people or processes idle, shorten the response boundary.
Take it with you
Judge a batch by total flow, not by how perfectly you avoid switching. Break it when a concrete waiting cost is clearly larger than a bounded switch-and-return cost, answer only enough to release the dependency, and leave a restart cue before moving. The comparison is usually qualitative; the point is to make the trade-off visible, not to manufacture precision.
Sources and evidence (4)
What we checked
In a self-organised task-switching experiment, participants tended to switch when the growing environmental delay for repeating a task approximately matched their individual switch cost.
This directly supports trading switching costs against another time cost, but the experiment used a controlled cognitive task. It does not provide a numerical threshold for professional interruptions.
Operations research on batch manufacturing shows systematic relationships among lot sizes, queueing-related lead times and work in process, with long lead times carrying additional costs.
This is a manufacturing model, not evidence about human task switching. It is used only to make the local-efficiency versus system-waiting trade-off explicit.
In an interruption experiment, participants completed interrupted work faster with no detected quality difference, while reporting more stress, frustration, time pressure and effort.
This challenges the assumption that interruption cost can be read directly from completion time or output alone. The tested tasks and interruption conditions do not establish that workplace interruptions are generally beneficial.
The Cost of Interrupted Work: More Speed and Stress · Abstract
Participants given an eight-second warning before an interruption prepared more and resumed the interrupted task faster than participants who were interrupted immediately; immediate-interruption participants improved with practice.
This supports brief preparation before a foreseeable switch. It does not validate any particular restart-note format, and practiced interruptions may carry different resumption costs.
Editorial sources (4)
- Balancing cognitive and environmental constraints when deciding to switch tasks: Exploring self-reported task-selection strategies in self-organised multitasking — Victor Mittelstädt, Insa Schaffernak, Jeff Miller, Andrea Kiesel
- Lot Sizes, Lead Times and In-Process Inventories — Uday S. Karmarkar
- The Cost of Interrupted Work: More Speed and Stress — Gloria Mark, Daniela Gudith, Ulrich Klocke
- Preparing to resume an interrupted task: effects of prospective goal encoding and retrospective rehearsal — J. Gregory Trafton, Erik M. Altmann, Derek P. Brock, Farilee E. Mintz
Save and practise this idea
My learning
Read. Recall. Repeat.
Save this idea and return to it on any device.
Keep exploring
Connected ideas
You are readingProtect the flow, not the batching streak
Original edition evidence and sources
These records belong to the original released edition. The current editorial reading has its own evidence above; neither set silently replaces the other.
An integrative review reports performance costs in many dual-task and task-switching conditions compared with appropriate single-task controls.
Switching can be necessary and adaptive; the review does not imply that uninterrupted work or large batches are always optimal.
A systematic review and meta-analysis of laboratory studies found interruption-management interventions improved primary-task accuracy and reduced resumption lag on average, with effects varying by intervention and task.
Laboratory findings do not identify one universal batching schedule for real professional work.
All sources (2)
- Cognitive structure, flexibility, and plasticity in human multitasking—An integrative review of dual-task and task-switching research
- Effects of interventions to reduce the negative consequences of interruptions on task performance: A systematic review, meta-analysis, and narrative synthesis of laboratory studies
Original edition, downloads and provenance
The reading above is editorial version 1.0.0, bound to this exact source edition. The original data and saved downloads retain that edition and the separate editorial layer. They may contain earlier wording; use the current reading above when interpreting it.
Version 0.1.0
No review details are supplied for this edition.
Read the original edition and overlay as JSON · Release details
