Skip to main content
Steady at the Top
← All articles

How to reduce decision fatigue in the CEO’s workday

Reduce executive decision overload by removing repeated choices from the CEO’s queue and improving the information attached to consequential requests. Audit the work before assuming that all tired decisions share a single psychological cause. Treat finite-willpower explanations as contested rather than as a proven daily decision limit.

Cognitive load is the demand a task makes on mental resources. The American Psychological Association’s 2018 defining reference, APA Dictionary of Psychology: cognitive load, relates the demand to the task rather than to a universal daily quota. Decision volume counts choices; it does not establish how much mental work each choice requires. An incomplete request can demand more analysis than several clear requests combined.

A private moment that makes the issue visible

Composite scenario, not a client case: You can approve an acquisition discussion in the morning and still spend dinner choosing the wording of an internal reminder. Your phone holds a queue of minor exceptions. At the end of the day, you postpone a staffing decision because the queue never produced a clear stopping point.

Use the research within its limits

Barsade (2002): The Ripple Effect: Emotional Contagion and Its Influence on Group Behavior studied mood transmission in a laboratory managerial decision task. It gives a reason to examine how a crowded queue changes expressed conduct in meetings, without diagnosing the CEO or explaining every delay through emotion. A decision audit must inspect the requests themselves as well as the interactions around them.

State the decision before discussing personalities

The practical question is which choices require your authority or judgment. Counting every decision equally hides differences in consequence, complexity, and the information needed to decide.

Use the CEO decision-queue audit

  1. Record a representative workday. List incoming decisions and their actual consequences. Note missing information, repeated topics, interruptions, and requests that arrive without a recommendation. Capture requests that arrive by informal message as well as scheduled meetings. The invisible queue may explain why the calendar understates the workload.

  2. Separate recurring rules from judgment. Identify choices that can follow an agreed threshold or standard. Preserve a route for exceptions instead of requiring personal approval for every ordinary case. Check whether the recurring choice genuinely follows a stable rule. A superficially routine request can carry unusual delivery or contractual consequences.

  3. Assign decision ownership. Give a named owner authority within the boundary. Clarify what they must report and what change in consequence requires escalation to you. Tell requesters where to send ordinary cases. Delegation will not shorten the queue if everyone continues copying the CEO for implicit approval.

  4. Improve the consequential request. Require the decision, alternatives, relevant evidence, and deadline. Return incomplete briefs for clarification rather than doing every missing analysis yourself. Ask the requester to distinguish the decision deadline from the date they would prefer an answer. Preserve genuine urgency without accepting every request as immediate.

  5. Check the queue after redesign. Review whether routine requests moved and whether material risks still reached you. Restore a boundary that fails without returning every choice to the CEO. Inspect a few delegated outcomes for missed escalation conditions. Improve the boundary where needed, rather than equating less CEO involvement with success.

Count the work hidden inside an approval

Start with the request as it arrived, not the polished decision eventually made. Record what you had to discover before answering. Did you locate an owner, reconstruct customer terms, compare conflicting numbers, or decide which problem the requester meant? Those activities belong in the workload description even if the final approval took only a moment.

Distinguish requests for judgment from requests for permission. A manager may know the appropriate action but lack authority to take it. Another may have authority but lack the information needed to exercise it. These require different repairs. Granting a broader approval limit will not help someone who cannot see the delivery consequence; improving a report will not resolve an unclear authority boundary.

Also mark duplicate decisions. The same exception may appear in a message, an executive meeting, and a follow-up call. Count the underlying choice once while recording the repeated contacts it creates. This prevents the audit from mistaking communication volume for distinct business complexity and points toward a shared place to record the answer.

Sort the queue by the redesign it needs

Some requests need a rule. Ordinary purchases within an approved budget might follow a defined owner and threshold. Other requests need a better brief because the tradeoff changes each time. A consequential pricing exception may require customer context, delivery exposure, alternatives, and a recommendation. Do not force that judgment into a rule merely to reduce the count.

A third category needs a scheduled forum. Several related staffing requests may depend on one resource allocation choice. Reviewing them separately can produce contradictory answers. Bring the shared constraint into one discussion with the relevant owners. Keep urgent exceptions available, but distinguish them from the ordinary allocation cycle.

A fourth category should stop. A report may exist because a past CEO once requested personal review. Verify whether any current obligation or operating risk still requires that review. If it has no continuing use, end the requirement explicitly and tell the people producing it. Silence can leave employees performing the work while assuming you still inspect it.

Design a brief the CEO can decide from

A decision brief should make the requested choice visible near the beginning. State the consequence of waiting, the viable alternatives, and the recommendation with its evidence. Include what remains uncertain. A short document can still be incomplete, while a longer attachment may be useful if the core choice is clear and the supporting detail is accessible.

In a hypothetical request, a manager asks whether to accept a customer deadline. The CEO should not have to discover that accepting it displaces another committed project. The brief names that tradeoff, identifies the affected owner, and presents the options already checked. If the displacement requires a decision at the CEO’s level, the request now contains the information needed to make it.

Return an incomplete brief with a precise question. Ask for the missing capacity comparison or the owner’s recommendation, rather than asking for more detail generally. Explain when a rapid provisional answer is necessary and what assumption it depends on. This keeps preparation standards compatible with real operating urgency.

Test whether the redesign transfers work responsibly

Choose a recurring category for an initial change. Give its owner the necessary access, authority, and escalation conditions. Tell the rest of the team how the route changes. If the CEO continues responding to every copied message, employees may reasonably conclude that the delegation was ceremonial.

Review both the CEO’s queue and the delegated work. Fewer messages can mean improved ownership, but they can also mean that a risky exception went unreported. Inspect examples from inside the boundary and near its edge. Ask whether the owner understood when consultation was required and whether another function supplied its part of the decision.

Do not set a target number of daily decisions as a psychological prescription. The useful comparison concerns the composition of the work: fewer repeated permissions, clearer consequential requests, and functioning exception routes. A day with fewer choices can still be demanding if one choice has serious consequences.

Keep the audit brief enough to finish. Once a category has a workable route, stop recording every instance unless a new problem warrants attention. The purpose is to redesign the queue, not create a permanent diary of executive busyness. Retain the decision boundaries and the next review date as the operating result.

Put the boundary into words

This request needs a decision owner before it needs my opinion. Show the consequence, the options, and your recommendation. If it is inside the agreed boundary, decide and record the exception only when the threshold is crossed.

Work through a bounded example

Hypothetical application: A CEO receives every discount request because nobody knows which terms delivery can absorb. Finance and operations define a reviewable boundary. Sellers can approve ordinary trades within it, while unusual scope and margin consequences escalate. The redesigned queue removes repetition without assuming that every delegated choice is low risk.

Check the failure mode before scaling

Ego depletion, the proposed draining of a limited self-control resource by prior exertion, is contested. Hagger and colleagues’ 2016 preregistered multilab replication found little evidence for the tested effect. That result does not establish that heavy workload is harmless. It does rule out presenting a fixed daily willpower allowance as settled science.

Repeated difficulty concentrating can have causes a decision log cannot assess. Use this audit to reorganize requests and ownership. If symptoms persist or affect daily life, consult a licensed clinician. Do not use a lower approval count as evidence that a health concern has been resolved.

Review the work, then revise the arrangement

Compare the new queue with the original sample. Check decision quality through specific cases, including a delegated exception, rather than claiming success solely because fewer requests reached you.

The finite-resource account also needs qualification. Hagger and colleagues (2016): A Multilab Preregistered Replication of the Ego-Depletion Effect (primary abstract) reported a multilaboratory replication that did not support a dependable depletion effect under the tested protocol. That finding does not make workload harmless. It means the queue audit should address observable demands without claiming that executives exhaust a scientifically established daily supply of willpower.

Connect this decision to the next one

Address the wider conduct issue with How CEOs can handle pressure without passing it down. If open requests follow you home, see How to stop work worries from taking over your evenings. Fear about survival warrants How to lead when you are afraid the business will fail. Assign authority through Who should decide what, and when should a decision escalate.

Sources

Steady at the Top