How to build a decision process that works under pressure
Build your decision process around the consequence of a choice, the authority to make it, and the evidence required before commitment. Give routine decisions a light route and consequential decisions an explicit owner, challenge process, and review point. Do not apply the same paperwork to an ordinary purchase and a choice that changes people’s responsibilities or the company’s future exposure.
Bounded rationality is decision-making constrained by available information and cognitive capacity, following Herbert Simon’s work, including his 1955 paper as discussed in the Stanford Encyclopedia of Philosophy. The practical response is to design a usable process around those limits. It is not to promise that a better checklist will eliminate uncertainty or make every judgement objectively correct.
A decision system differs from a decision-making personality. It specifies who decides, what they need to consider, who can challenge the proposal, and when the choice may be reopened. You can inspect those arrangements without diagnosing whether an executive is naturally decisive, cautious, or intuitive.
Notice when the system becomes your personal availability
Consider this composite. You are meant to be at a family gathering, but your phone keeps filling with requests to approve changes. One concerns an ordinary expense, another a customer concession, and another a proposed leadership exit. Each message uses the same label: urgent. You answer them in arrival order because distinguishing their consequences would require a process no one has agreed.
The issue is not only how quickly you respond. Ask why each decision reached you, what authority existed below you, and whether the requester supplied the information the choice required. If every uncertainty needs your personal reassurance, the company has a dependency on your availability rather than a clear allocation of decision responsibility.
This composite is a pattern to examine, not evidence about a particular company. You do not need to withdraw from decisions to repair it. Separate the choices that require your authority from those that require a clearer instruction, an exception threshold, or access to a person who already has authority to act.
Use the consequence-and-ownership design
This proposed design method creates a decision route that fits the work. It is not a validated score or a replacement for legal, financial, technical, or other specialist judgement. When a choice concerns employment, governance duties, regulated activity, or a formal obligation, involve the relevant professionals before treating the route as sufficient.
1. Name the decision precisely
Write the choice in a sentence that identifies what will change. Separate approval of an investigation from approval of an outcome, and a recommendation from an authorized commitment. A request to “support the project” may hide several decisions with different owners. Divide them before deciding which forum should handle them.
Ask what happens if no decision is made now. Sometimes the current arrangement continues; sometimes a contract, delivery obligation, or external event makes delay consequential. Establish the actual condition. Do not infer a deadline from the urgency of the message or assume that waiting preserves the status quo without cost.
2. Classify the consequence
Examine the exposure created by the choice: what becomes committed, which people are affected, and what a reversal would require. Include dependencies that continue after the initial action. A small payment can initiate a larger obligation, while a large research exercise may leave the final commitment open. Price alone is not the classification.
Choose categories that your team can use consistently. You might distinguish routine delegated work, exceptions within an existing plan, and commitments requiring senior or formal approval. These are proposed categories, not universal thresholds. Define them from your organization’s actual authority and the obligations the decision creates.
3. Assign one accountable decision owner
Name the person authorized to decide and the people whose approval or advice is required. Distinguish consultation from consent. A specialist may provide an essential assessment without owning the commercial choice; a formal approval may be required regardless of whether the owner agrees with it. Make those relationships explicit.
Lerner and Tetlock’s review of accountability emphasizes that accountability arrangements differ. Apply that caution when designing the route. Asking many people to comment does not establish who must explain the choice. Require a named owner to state the basis of the decision and the limits of their authority.
4. Specify the minimum evidence and the unresolved assumptions
Identify what must be established before commitment and what may remain uncertain. Name the source, age, and limitations of the evidence. Separate an observed fact from an estimate and a desired outcome. Do not demand certainty that the decision cannot provide, but do not let a confident narrative replace a missing essential check.
For a consequential forecast, include a relevant comparison with earlier cases. Kahneman and Lovallo’s 1993 analysis warns against evaluating current plans only from the inside view of the plan. The operating application is to request a suitable comparison and document its limits. It is not a guarantee that historical averages predict your case exactly.
5. Give challenge a defined job
Ask a reviewer to test the assumptions or identify an exposure the proposal may have missed. Specify what they should inspect and when they must respond. Avoid appointing a generic devil’s advocate whose disagreement is treated as performance. A useful challenge identifies evidence or a decision condition, not merely an opposing attitude.
Keep the review proportionate to consequence. A routine delegated choice may need no extra reviewer. A consequential commitment may need an independent specialist or formal committee. Identify how an unresolved objection reaches the decision owner. Do not give every participant an unspoken veto simply because you have invited their opinion.
6. Record the choice and the review condition
Capture the decision, owner, principal reasons, unresolved assumptions, and implementation obligation. Use the shortest record that preserves the meaning. A routine choice may need a line in the normal workflow. A consequential decision may require a fuller memo or formal record. The objective is later understanding, not uniform document production.
Specify what would reopen the decision. It may be a changed assumption, a new obligation, a failed test, or an agreed review point. Separate new evidence from continued disagreement with a choice already authorized. Make it possible to stop when a real condition changes without requiring the owner to defend the original judgement indefinitely.
Design the route for a tired organization
Test the process with the people who will use it under ordinary workload. Can they identify the owner without calling you? Can they distinguish a request for advice from a request for approval? Can they find the relevant evidence and explain what remains unresolved? A policy that only works when its author is present needs further design.
Use a hypothetical urgent case to test absence coverage. Identify who can act when the usual decision owner is unavailable and which commitments must wait. Confirm the route through the responsible governance or legal process where necessary. Do not quietly transfer authority because the right person did not respond to a message quickly enough.
Also test whether an executive can bypass the route by presenting a preference as a settled instruction. If your own message overrides an agreed decision, make the override explicit. Record what changed and who now owns the consequence. Do not let the team carry conflicting directions and later judge it for lacking judgement.
Keep judgement separate from the appearance of diligence
A long paper can leave the core assumption untested. A short conversation can settle a routine choice appropriately. Review whether the process established what mattered, rather than how much material it produced. Ask the owner to state the strongest reason their preferred option might be wrong and the evidence that would change the recommendation.
When a result disappoints, examine the information available at the time. Do not assume the original process was weak solely because the outcome was poor. Equally, do not treat a fortunate result as proof that an unauthorized shortcut was sound. Preserve the record so the review can distinguish judgement, execution, and changed circumstances.
Your own pressure is relevant to the design but does not explain every error. Build a route that reduces unnecessary interruption and identifies decisions you should not make in an improvised exchange. If stress-related symptoms persist or affect daily life, consult a licensed clinician. A decision system is an operating intervention, not treatment.
Introduce the system with an actual example
A proposed introduction is: “We will use the consequence of the choice to determine its route. This decision belongs to this owner, requires these checks, and comes back to me only if this condition changes. Advice is welcome, but approval authority remains where we have explicitly placed it. We will record the reason and the next review condition.”
Walk through a current, appropriately shareable decision. Ask participants to identify which part they own and what they still need. Correct ambiguity before asking them to use the process on a higher-consequence case. Do not begin with a company-wide form that has not been tested against real work.
Connect the design with your confirmation-bias review when the evidence appears too convenient. Use a premortem workshop for a proposal whose failure modes need examination. For a group decision, use the executive meeting design to collect independent contributions before the discussion settles around your preference.
Review a sample, then simplify
Choose a small set of completed decisions and inspect whether the route fit their consequences. Look for avoidable escalations, missing authority, unresolved assumptions, and records that did not help anyone act. Ask the users which step added useful information and which step merely repeated an earlier approval.
Remove a step only after understanding the function it served. Preserve required professional and formal approvals. Refine the route so a person can make an authorized decision, explain it, and recognize when it needs review. Your objective is a company that can exercise judgement through clear responsibilities, including when you are away from your phone.