Executive Task Switching and Decision Latency: What the Evidence Shows

Executive task switching and decision latency

Executives rarely experience a day as a clean sequence of decisions. A board question may arrive during a hiring review. A customer escalation may interrupt financial planning. A calendar change may force a return to a half-finished approval. Each event is ordinary. The difficult question is whether the switching pattern is slowing decisions, increasing rework, or pushing judgment into less suitable hours.

This review examines cognitive and workplace evidence on task switching, interruptions, and decision timing. The research supports a practical conclusion: interruptions can change the time, effort, and quality conditions around work. It does not support a single benchmark for how quickly a CEO, founder, or other executive should decide. Executive roles differ too much in authority, risk, information quality, and crisis exposure for that inference.

TL;DR

Task-switching experiments consistently find a switch cost, even when people know a switch is coming. Interruption research adds a second cost: people may need time to reconstruct the state of the original task, and they may change their strategy or make more errors. Those effects are not identical in every study. Some work shows faster completion under interruption, with higher stress and effort rather than lower measured quality.

For executive operations, the right measurement is local. Track when a decision request became ready, when a decision was made, how often it returned for missing context, and whether an interruption displaced planned work. Compare like with like. A short urgent decision is not comparable to a high-stakes hiring, legal, or investment decision.

Contents

• Methodology
• Limitations
• Key statistics
• Key takeaways
• Evidence and comparison table
• Analysis
• Sources
• Frequently asked questions
• Related research

Methodology

I reviewed ten sources retrieved on August 1, 2026. The set includes peer-reviewed experiments, a systematic review, a recent interruption simulation, institutional research summaries, and public workplace datasets. Preference went to primary papers or the institution that produced the data. Each source was recorded with its publication date, population, geography, measure, and denominator where the source reported one.

The review distinguishes three terms. A task switch is a change from one task to another. An interruption is an external event that stops the primary task and requires attention to a secondary task. Decision latency is the elapsed time between a decision becoming ready for review and a decision being recorded. A message arrival, a meeting count, or time spent online is not automatically decision latency.

The executive interpretation is an inference. Most studies do not sample CEOs or senior executives. Findings from students, clinicians, general knowledge workers, or simulated teams can describe mechanisms that may matter in executive work. They cannot establish executive-specific effect sizes without executive samples and a comparable definition of the outcome.

Limitations

The evidence uses different tasks, populations, settings, and outcome measures. Laboratory switch costs are not the same as a senior leader moving between a pricing decision and a customer call. Clinical interruption studies involve safety-critical workflows, where interruptions can sometimes be necessary. Workplace surveys can show association, not causation. Institutional reports may use proprietary samples or definitions that are not directly comparable.

The often-repeated claim that every interruption takes 23 minutes to recover from should not be used as a benchmark here. The primary 2008 paper linked below reports a different result: participants completed interrupted email tasks faster, with no measured quality difference, but reported greater stress, frustration, time pressure, and effort. A separate field observation reported time to return to a task, which is not the same measure as restored concentration. The distinction matters.

Key statistics

• The APA review describes two components of switch cost: reconfiguring control settings and interference from the previous task. More preparation reduced the first component but did not remove the full cost.
• The interruption systematic review included 79 papers, of which 63 experimental studies were analyzed. It found overall negative effects across completion time, resumption lag, work strategy, decision-making process, and error in the clinical literature it reviewed.
• Mark, Gudith, and Klocke studied 48 participants in a controlled email-writing experiment. Interrupted participants finished faster, with no reported quality difference, but reported more stress, frustration, time pressure, and effort.
• A 2025 interruption-model study used 312 participants working as nurses in a simulated clinical team. Participants were more likely to accept an interruption as the interrupter's time pressure increased relative to their own. This is evidence about a simulated clinical task, not executive decision speed.
• A 2017 study on blocking distractions examined workplace focus and productivity. Its population and operational definitions should be read from the paper before applying the findings to an executive team.

Key takeaways

• Protect complex work from avoidable switching, but keep a path for genuine urgency.
• Prepare decisions before they reach the executive. The request should state the decision, owner, deadline, relevant context, options, and recommendation.
• Measure latency from readiness to decision, not from message receipt. Otherwise the metric mixes waiting for information with waiting for attention.
• Record rework. A fast decision that returns twice because the brief was incomplete is not necessarily fast in operational terms.
• Treat an executive-specific benchmark as an internal baseline that can be revised when role, team, or risk changes.

Evidence and comparison table

Source and publication date	Population, geography, and sample	What was measured	Denominator and use limit
APA, Multitasking: Switching costs, retrieved 2026-08-01	Research summary, not a new sample; underlying task-switching experiments	Switch cost, preparation, control settings	No single pooled denominator. Use for mechanism, not an executive benchmark.
Rubinstein, Meyer, and Evans, 2001	Laboratory participants; geography and final sample vary by experiment	Executive control in task switching, response time, and accuracy	Experimental trials, not executive work. Supports a switch-cost mechanism.
Gonzalez and Mark, 2004	14 information workers observed in workplace settings	How workers organized tasks and switched between work spheres	14 observed workers. Descriptive, not a causal productivity estimate.
Mark, Gudith, and Klocke, 2008	48 participants in an email-writing experiment; United States and Germany authorship	Completion time, quality, stress, frustration, time pressure, effort	48 participants. The task was controlled email work, not executive judgment.
Westbrook et al., 2010 systematic review	Clinical interruption literature across included studies	Completion time, resumption lag, work strategy, decision process, error	79 papers, 63 experimental studies. Clinical setting limits generalization.
Monteiro et al., 2022	Medical residents in three diagnostic-reasoning experiments	Diagnostic reasoning under interruption	Study-specific samples. The paper says causal evidence in this domain remains limited.
Mark, Iqbal, and Czerwinski, 2017	Workplace focus study; authors at UC Irvine and Microsoft Research	Effects of blocking distractions on focus and productivity	Study sample and task definitions are paper-specific. Do not convert into an executive percentage.
Crocker and Keller, 2025	312 nurses in a simulated clinical team	Interruption and response decisions under time pressure	312 simulated participants. Clinical simulation, not executive decision latency.
Microsoft Work Trend Index, 2023 edition	Global workplace survey and telemetry-based reporting; report-specific samples	Work patterns, meetings, and digital interruptions	Denominators vary by edition and measure. Read the methodology before citing a percentage.
U.S. Bureau of Labor Statistics American Time Use Survey, 2024 data release	U.S. civilian population age 15 and over; nationally representative time-use survey	Time spent in activities, by population categories	Population time-use denominator. It does not measure executive decision latency.

Analysis

Switching changes the conditions around a decision

The cognitive literature gives a simple explanation for why a crowded executive day can feel slow even when every item receives attention. A switch requires a person to activate the rules for the new task and suppress or carry the rules from the old one. When the original task resumes, the person also has to reconstruct what was known, what remained open, and what the next step was.

Preparation helps. A decision brief that states the question and options reduces the work needed to rebuild context. It does not make switching free. If the executive must interrupt a complex analysis to answer a low-context message, the cost may show up later as re-reading, a missed dependency, or a decision that needs to be revisited.

Interruption is not always harmful in the same way

The evidence does not justify a blanket rule that every interruption makes output worse. In the 2008 experiment, interrupted participants completed the task faster and did not show a measured quality difference. The tradeoff appeared in stress, frustration, time pressure, and effort. Other studies, especially in clinical settings, report slower completion, longer resumption, changed strategies, or more errors.

The practical lesson is to measure both output and cost. A team that finishes more items by compressing work may be paying through fatigue, rework, or lower willingness to take on difficult decisions. A team that blocks every notification may miss a real escalation. Routing rules need an urgency definition, an owner, and a fallback channel.

Decision latency needs a clean start point

Executives often receive a request before it is ready for a decision. The request may be missing a recommendation, a budget impact, a legal review, or the person responsible for implementation. Counting from the first message makes the executive appear slow when the real delay is preparation.

Use two timestamps instead:

1. Ready time: the decision owner has the minimum agreed context and the request is explicitly ready for review.
2. Decision time: the decision or next step is recorded and the owner is clear.

Then add a third measure, rework rate, for decisions returned because the brief was incomplete or circumstances changed. Keep separate queues for routine approvals, people matters, external commitments, and high-risk decisions. A median can describe a queue, but it should not replace the distribution. A few urgent incidents can distort an average.

A practical operating test

Run a two-week baseline without changing notification rules. Record planned focus blocks, interruption category, decision ready time, decision time, number of missing-context returns, and after-hours spillover. Do not copy confidential message content into the dataset.

After the baseline, change one rule. For example, an assistant can batch routine approvals into two windows, protect a scheduled analysis block, and route urgent items through a named channel. Review the same measures after two weeks. Keep the change only if decision turnaround or rework improves without shifting unreasonable load to the support team or the executive's evening.

The executive remains responsible for priorities and sensitive judgment. An assistant or chief of staff can improve the intake, prepare options, group related work, and close follow-up loops. This is an operating design choice, not proof that delegation eliminates cognitive costs.

Sources

1. American Psychological Association, "Multitasking: Switching costs," retrieved 2026-08-01
2. Rubinstein, Meyer, and Evans, "Executive control of cognitive processes in task switching," 2001
3. Gonzalez and Mark, "Constant, constant, multi-tasking craziness," 2004
4. Mark, Gudith, and Klocke, "The cost of interrupted work: More speed and stress," 2008
5. Westbrook et al., "A systematic review of the psychological literature on interruption and its patient safety implications," 2010
6. Monteiro et al., "Failure to demonstrate effects of interruptions on diagnostic reasoning," 2022
7. Mark, Iqbal, and Czerwinski, "How blocking distractions affects workplace focus and productivity," 2017
8. Crocker and Keller, "Testing a computational model of interruptions," 2025
9. Microsoft, Work Trend Index
10. U.S. Bureau of Labor Statistics, American Time Use Survey

Frequently asked questions

Does task switching slow executive decisions?

It can add time and effort around complex work, but available studies do not establish one executive-specific delay. Measure from decision-ready time to recorded decision time, then compare similar decisions.

Is the 23-minute interruption claim a reliable benchmark?

No. The number is often presented as recovery of concentration, while the primary sources distinguish returning to a task from restoring full focus. The 2008 experiment linked here found faster completion under interruption with higher reported stress and effort.

How should an executive team measure decision latency?

Record when a request became ready for review and when the decision was recorded. Add missing-context returns, rework, decision type, and urgency. Keep message receipt as a separate operational measure.

What can an executive assistant do about switching costs?

An assistant can improve intake quality, batch routine work, protect agreed focus blocks, route urgent matters, and maintain follow-up records. The executive should retain authority for sensitive or high-risk decisions.

Related research

• Executive Notification Overload
• How to conduct a CEO time audit
• What tasks waste the most CEO time?

CTA

If interruptions are obscuring priorities, book a free consultation call to discuss a measured executive support workflow.