Learning to Make Decisions With Incomplete Information

Every significant decision founders make involves incomplete information. Should we pursue this market? Should we hire this person? Should we take this deal? You never have all the data you'd like before deciding.

Waiting for certainty isn't an option. By the time you have perfect information, the opportunity has passed. Markets have shifted. Competitors have moved. Candidates have accepted other offers.

The skill of deciding well with incomplete information separates great founders from everyone else. It's not about being right every time. It's about making good bets consistently and adjusting as you learn.

Good process, not just good outcome:

• Luck affects outcomes
• Good decisions can have bad outcomes
• Bad decisions can have good outcomes
• Judge yourself on process, not just results

Appropriate information gathering:

• Enough to inform, not enough to delay
• Diminishing returns on more research
• Know when to stop gathering and decide
• Action generates information too

Clear about what you're betting on:

• What assumptions is this decision based on?
• What would need to be true for this to work?
• What signals would tell you to reverse?
• Explicit beliefs, not vague hopes

Reversibility considered:

• How reversible is this decision?
• Can you unwind it if wrong?
• One-way doors deserve more care
• Two-way doors can be faster

Identify what would change your mind:

• Before researching, ask: what would cause me to decide differently?
• Focus research on those specific questions
• Don't gather information that won't affect the decision
• Targeted research beats comprehensive research

Talk to people who know:

• Expert opinions are information-dense
• One good conversation may beat hours of desk research
• People with direct experience > people with theories
• But calibrate for biases and conflicts

Look for disconfirming evidence:

• Actively seek reasons you might be wrong
• The thing you don't want to find is often the thing that matters
• Challenge your preferred conclusion
• Steel-man the alternative

Time-box the research:

• Set a deadline for deciding
• Gather what you can before deadline
• Accept you won't have everything
• Decision with 70% information often beats delay for 90%

Expected value thinking:

• Probability × outcome for each option
• Choose option with best expected value
• Helpful but limited (assumes you can estimate probabilities)
• Good for structuring thinking even if numbers are rough

Regret minimization:

• What would you regret more: trying and failing, or not trying?
• Useful for big, irreversible choices
• Jeff Bezos famously used this for Amazon decision
• Different time horizons change the calculation

Reversibility-based speed:

• One-way doors: decide slowly and carefully
• Two-way doors: decide quickly, correct as needed
• Most decisions are more reversible than they seem
• Default to speed for reversible decisions

Pre-mortem analysis:

• Imagine the decision failed
• What would have caused the failure?
• Use that to identify risks to monitor
• Not to avoid deciding, but to decide more wisely

Set deadlines:

• Decision will be made by X date
• Stick to deadline regardless of comfort level
• Discomfort with uncertainty isn't reason to delay
• Deadline creates forcing function

Recognize diminishing returns:

• First 20% of research gives 80% of insight
• Additional research rarely changes decisions
• Be honest when research is procrastination
• Enough to be informed, not enough to be certain

Accept that all options have flaws:

• Waiting won't reveal a perfect option
• You're choosing between imperfect alternatives
• Looking for certainty that doesn't exist
• Best available beats hypothetically perfect

Default to action:

• When in doubt, act
• Action generates information
• Mistakes can often be corrected
• Inaction is itself a decision with consequences

Expect to be wrong sometimes:

• Good decision-makers are wrong regularly
• The goal is to be right more often, not always
• Build error correction into your approach
• Being wrong isn't shameful; not learning from it is

Set up to learn quickly:

• Define what success looks like before deciding
• Create checkpoints to evaluate
• Build feedback mechanisms
• Know when you'll revisit the decision

Separate decision quality from outcome:

• Good decision + bad outcome: learn, don't self-flagellate
• Bad decision + good outcome: don't congratulate yourself
• Track decision quality over time
• Build calibration of your own judgment

Update and move on:

• When evidence shows you're wrong, change course
• Don't throw good money after bad
• Ego shouldn't prevent reversals
• Being wrong once doesn't mean being wrong always

Track your decisions:

• Log significant decisions and reasoning
• Record what you believed and why
• Review outcomes periodically
• Identify patterns in your thinking

Get feedback:

• Ask advisors to evaluate your reasoning
• Debrief decisions with co-founders
• Learn from others' decision processes
• Invite challenge to your thinking

Study decision-making:

• Read about cognitive biases
• Learn probabilistic thinking
• Understand your personal tendencies
• Knowledge improves judgment

Practice deliberately:

• Make decisions consciously
• Apply frameworks explicitly
• Reflect on process, not just outcomes
• Treat decision-making as skill to develop

How do I know when I have enough information? When additional research isn't changing your confidence significantly. When you keep finding the same things. When the opportunity cost of delay exceeds the benefit of more research. Usually sooner than feels comfortable.

What if my decision affects many people? Higher stakes warrant more careful process, but still not infinite delay. Get input from those affected. Be transparent about uncertainty. Create mechanisms to adjust if wrong.

How do I decide when my gut says one thing and data says another? Investigate the conflict. Your gut often has information your conscious mind hasn't processed. But gut can also be bias. Look for what might explain the divergence. Sometimes gut wins, sometimes data wins, but examine the conflict before deciding.

What about big, irreversible decisions? These deserve more time and more careful analysis. But even irreversible decisions can't wait forever. Set a higher bar for confidence, get more input, but still decide within a reasonable timeframe.

How do I get comfortable with uncertainty? Practice. Start with smaller decisions. Accept that discomfort is part of the job. Reframe uncertainty as opportunity. And remember that the alternative (not deciding) is also uncertain.