Decision Frameworks

The reality

A founder fires a senior project manager on a Thursday afternoon after a heated call with a major client. By Saturday the founder has realised the project manager was right about the technical position and the client was wrong. By the following Tuesday the founder has paid a settlement, lost the project manager's institutional knowledge across 14 active projects, and absorbed two months of recovery cost trying to hire a replacement. The bad call traces to the conditions. 90 seconds of pressure, no defined process for a senior termination, no second voice in the room, no waiting period for an irreversible call. Most consequential founder decisions are made under conditions that guarantee the outcome.

Read this if

• A consequential decision in the last quarter was made in less than an hour and was later regretted
• The founder cannot name the difference between a decision that needs five minutes and one that needs five days
• Hires, firings, or major spend decisions happen in the heat of the moment rather than in scheduled windows
• The same kind of decision (a key client conversation, a partnership agreement) keeps producing the same kind of regret
• The founder has no defined "second voice" for irreversible decisions
• A decision that turned out to be wrong was made with information the founder could have had if they had waited 24 hours

What dysfunction costs

Reversal cost. A consequential decision made under pressure and reversed within a week costs the business the original decision plus the reversal. A senior termination reversed in three days costs the legal exposure, the team's confidence, and the relationship with the person who is now back in the business knowing the founder almost let them go.

What success looks like

When decision-making is a discipline:

• The founder can sort decisions into categories (reversible, partially reversible, irreversible) and apply a different process to each
• Irreversible decisions (senior termination, major spend, key partnership commitment) have a mandatory waiting period and a named second voice consulted before the call
• The founder has a "decision log" of consequential decisions, the inputs, and the outcome 30/60/90 days later
• Decisions made under pressure are rare because the windows for pressure decisions have been engineered down
• The team knows the categories and brings decisions to the founder in the right window
• A reversed decision is reviewed for both the call and the conditions that produced it

The framework

Decision-making runs as four layers. Each layer answers a different question about how to make a consequential call well.

Layer 1: Categorise the decision

Every decision sits in one of three categories, and each category needs a different process.

Reversible. A decision that can be unwound within 30 days at low cost. Hiring a junior team member. Trying a new vendor on one project. Running a small marketing experiment. These can be made fast. The cost of being wrong is the cost of trying.

Partially reversible. A decision that can be unwound but with real cost. A senior hire who has signed a contract. A six month vendor commitment. A pricing change communicated to clients. These need more deliberation, a clear input list, and a defined window (24 to 72 hours).

Irreversible. A decision that cannot be unwound without serious damage. A senior termination. A major partnership signed. A capital commitment above a defined threshold. A public statement about the business. These need the full process: waiting period, second voice, written rationale.

The behaviour to adopt this week: list five recent consequential decisions. Sort them into the three categories. Notice which category produced the most regret.

Layer 2: The waiting period

Irreversible decisions get a mandatory waiting period. 24 hours minimum for terminations. 72 hours for major spend or partnership commitments. The waiting period is the part of the decision where the founder's brain switches from threat-response to deliberation, not a delay. The same decision made in 90 seconds and made after 24 hours produces a different output more often than founders expect, and rarely a worse one.

When the waiting period is skipped: the founder almost always cites urgency. The urgency is almost always lower than it feels. A senior termination on Thursday afternoon does not need to happen on Thursday afternoon. The team will absorb the wait. The founder will not absorb the reversal.

The behaviour to adopt this week: write the waiting periods for the three or four categories of irreversible decision. Tape them to the desk. Hold them.

Layer 3: The second voice

Irreversible decisions need a second voice in the call. A cofounder, a senior team member, a trusted advisor, a peer. The second voice has one job: ask the question the founder is not asking, surface the input the founder is not weighing, or simply say "are you sure?" at the moment when the founder is most certain.

The second voice cannot be the personal partner unless the personal partner is formally inside the business. The second voice has to have full operational context, or the input is partial.

The behaviour to adopt this week: name the second voice for each category of irreversible decision. Hire, fire, major spend, partnership. The same person can serve multiple categories. The point is that the role is named.

Layer 4: The decision log

Every consequential decision goes into a written log. Date, category, decision, top three inputs, second voice, expected outcome at 30/60/90 days. Once a quarter the founder reviews the log against actual outcomes. The review is the founder learning where their judgment runs strong and where it runs weak. Done well, it surfaces the patterns the founder cannot see in any single decision.

When the log does not exist: every decision feels novel because the founder has no record of how the last one went. Patterns of regret repeat without being noticed.

The behaviour to adopt this week: open a single document. Title it "Decision Log". Add the last three consequential decisions. Add a 30 day check-in date for each.

A founder you might recognise

A founder runs a 26 person construction services business in Jebel Ali. AED 13M (USD 3.5M) last year. Through 2024 and 2025 he made three senior decisions under pressure that all needed reversal: a senior hire fired in week six and rehired in week ten, a vendor partnership signed and unwound within four months, and a pricing change communicated to clients then walked back. The pattern was visible to the team. He could not see it himself.

In Q1 2026 his accountability group pushed him to write a decision log. Reading the log, the founder noticed that 11 of his 14 consequential decisions in the previous year had been made within 30 minutes of a triggering event. None of them had a second voice. None had a waiting period.

He defined three categories with waiting periods (24 hours for termination, 72 hours for major spend, 7 days for partnership commitments). He named his cofounder as the second voice for hires and fires, and his accountability advisor as the second voice for partnerships. He kept the decision log running.

In the next two quarters he made eight consequential decisions. Six went through the waiting period and the second voice as written. Two were genuine emergencies that did not. Of the eight, one needed a small adjustment. None needed a reversal. The founder had not become smarter. The conditions had become better.

Working through it

1. Write the three categories: reversible, partially reversible, irreversible. List five examples of each from the business in the last quarter. The categories become real when the founder can see their own decisions in them.

2. Define the waiting period for each irreversible category. 24 hours, 72 hours, 7 days. Write the rule. Communicate it to the cofounder, the senior team, and any second voice.

3. Name the second voice for each category. Hire, fire, major spend, partnership, public statement. The second voice is a named person.

4. Open the decision log. Date, category, decision, top three inputs, second voice, expected outcome at 30/60/90 days. Backfill the last three consequential decisions.

5. Review the log quarterly. What did each decision actually produce? Where did the framework hold? Where did it not? Where was the framework skipped, and what did skipping cost?

Common mistakes

• Treating urgency as a category override. Almost every "I have to decide right now" feeling is wrong. The team will absorb a 24 hour wait. The founder will not absorb a reversal.
• Skipping the second voice because "they will not see what I see." The second voice is valuable precisely because they are not in the founder's head. Their job is to surface what the founder is missing.
• Writing the decision log and not reviewing it. A log without a quarterly review is journaling. The review is what produces the learning.
• Categorising irreversible decisions as partially reversible. A senior termination is irreversible in practice even if the legal mechanism allows reversal. The relationship does not come back. Categorise on the relational reality.
• Building the framework and not communicating it to the team. A team that does not know the categories cannot bring decisions in the right window. They will surface a hire-and-fire on Thursday afternoon because they do not know that category needs 48 hours.

Self-assessment

Y or N for each.

1. Can the founder name the three categories of decision and apply the right process to each?
2. Are waiting periods written down for each category of irreversible decision?
3. Has the second voice been named for each irreversible category?
4. Does a decision log exist with the last three consequential decisions, inputs, and expected outcomes?
5. Is the decision log reviewed at least quarterly against actual outcomes?
6. Has the founder waited the full waiting period on at least one irreversible decision in the last quarter?
7. Has the team been briefed on the categories so they bring decisions in the right window?

Five or more "yes" answers means decision-making is doing the work it is supposed to do. Three or four is the band where the framework exists in part but the discipline has not taken hold. Two or fewer means the next consequential decision is being made under conditions that guarantee a worse outcome than the founder is capable of producing.