The DECIDE BETTER® method

Your gut is a data set. Audit it.

Intuition is compressed experience: reliable in familiar situations, unreliable in novel ones. Intuition and information are partners, and you need a method for when they disagree. When they do, that is the cheapest research you will ever get.

Five rules

The gut sets the bar. The data tests it. You stay accountable.

  1. 01

    Set the bar before you see the data

    Make your gut testify. Before you open the report or run the model, write down what you expect and the result that would change your mind. Say it to your team.

    Set the bar first guards against 26: Anchoring bias, Dunning-Kruger effect, HALT: Hungry, Angry, Lonely, Tired

  2. 02

    Check confirming data as hard as contradicting data

    If the numbers agree with you, audit them as aggressively as the numbers that do not. We are just not wired to be contrarian to our own beliefs.

    Audit the good news too guards against 33: Confirmation bias, Devil's Advocate, Groupthink

  3. 03

    Make sure the data answers your question

    Every data set answers somebody's question. Make sure it's yours, or you end up precisely wrong and only generally right.

    Data answers your question guards against 40: Availability heuristic, Framing effect, Survivorship bias

  4. 04

    Decide on the ingredients, not just the answer

    Name the criteria and weigh them before you argue about the winner. Not everything matters equally; saying so is just not wanting to do the work.

    Decide on the ingredients guards against 40: Decoy effect, Endowment effect, Hyperbolic Discounting

  5. 05

    Never make a decision you can't explain

    If you cannot say how you got there, the tool decided, not you. The gut sets the bar. The data tests it. You stay accountable for the call.

    Explain the decision guards against 19: Hindsight Bias, Post-mortem analysis, After-Action Review

Match rigor to stakes

Mundane, Important, Epic

Spend more time on decisions you cannot undo. Spend almost none on the ones you can.

Mundane

Set it, forget it.

Low stakes and easy to undo. Let system one handle it, use a template, and save your attention for what matters.

For example: Vendor renewals, routine scheduling, a team lunch.

Important

Measure twice.

Real consequences, partly reversible. Bring information to the table, weigh the criteria, and invite a second view.

For example: A key hire, a new program, a budget reallocation.

Epic

Measure twice, cut once, seek wise counsel.

Hard or impossible to undo: acquire, shut down, let a key person go. Spend more time on decisions you cannot undo. Run the premortem and the red team.

For example: Mergers, closures, decade-defining investments.

Premortem

Why wait for the postmortem? Run the premortem.

Before you act, imagine it is a year from now and the decision has failed. Each person writes down why. The reasons that surface are the risks you can still do something about. Some teams call it a before-action review.

Red team

Red team your decisions before they are made

Give someone the job of arguing the other side, with real time and real authority. Those who are often quiet have the most to say: at the edges of the organization is where you will find the greatest signal to noise. Then disagree and commit.

The library

Biases, fallacies, razors and practices

Each entry has a meaning, an example, an evidence grade and the DECIDE BETTER® rule that guards against it. Search it here, and get the countermeasures, favorites and completions in the app.

biases, fallacies, razors and practices in the library

Category
Evidence grade

Featured entries

Anchoring biasCategory: Cognitive BiasesType: PerceptionGrade
Also called
Anchoring effect, Anchoring and adjustment
Meaning
Relying too heavily on the first number or idea you hear, then adjusting too little from it.
Example
A board sets next year's budget by nudging last year's figure up 3 percent instead of asking what the work actually needs.
Mitigation
Set the bar before you see the data: write your own estimate first, then compare it with the anchor on the table.
Open the full entry: Anchoring bias
Availability heuristicCategory: Cognitive BiasesType: MemoryGrade
Also called
Availability bias
Meaning
Judging how likely something is by how easily examples come to mind.
Example
After one vivid donor complaint, a team concludes donor satisfaction is falling, though the survey says otherwise.
Mitigation
Ask for the base rate. Count the cases instead of recalling them.
Open the full entry: Availability heuristic
Confirmation biasCategory: Cognitive BiasesType: BeliefGrade
Also called
Myside bias
Meaning
Seeking out and favoring information that confirms what you already believe.
Example
A program director backs a new intake process. She shares the pilot results that show time saved with the board, and sets aside the staff survey that found more errors. The board approves it after hearing only one side.
Mitigation
Check confirming data as aggressively as contradicting data. Assign someone to argue the other side.
Open the full entry: Confirmation bias
Decoy effectCategory: Behavioral EconomicsType: PerceptionGrade
Also called
Asymmetric dominance effect
Meaning
Adding a clearly worse third option changes which of the original two people pick.
Example
A middle pricing tier looks like a bargain only because a lopsided premium tier sits next to it.
A caution
Strong with numeric options, weaker with real products and pictures.
Mitigation
Remove dominated options and compare the real alternatives head to head.
Open the full entry: Decoy effect
Devil's AdvocateCategory: Best PracticesType: LearningGrade
Also called
Dialectical inquiry (related)
Meaning
One person deliberately argues against the prevailing view to test it.
Example
In a hiring decision, one panelist argues the case for the runner-up candidate.
A caution
Assigned dissent helps less than genuine dissent; rotate it and give it real authority.
Mitigation
Rotate the role so dissent is expected, not personal.
Open the full entry: Devil's Advocate
Dunning-Kruger effectCategory: Cognitive BiasesType: PerceptionGrade
Meaning
The idea that people with little skill in an area rate their own ability too high. Later work finds a smaller effect, and part of the pattern comes from statistics.
Example
A new manager is certain the vendor selection is simple, while the procurement lead sees six open risks.
A caution
The popular chart, with beginners as the most confident people in the room, is not what the study found. Low performers do overrate themselves, but much of the pattern can come from statistics (regression to the mean), and top performers still rate themselves highest.
Mitigation
Ask the people closest to the work to rate your plan. Compare self-ratings with outcome data.
Open the full entry: Dunning-Kruger effect
Endowment effectCategory: Behavioral EconomicsType: MoneyGrade
Meaning
Valuing something more simply because you already own it.
Example
A team rates its homegrown tool far above an equivalent product it would never have built.
A caution
Size depends on how the exchange is framed and run.
Mitigation
Ask what you would pay to acquire it today if you did not already have it.
Open the full entry: Endowment effect
Framing effectCategory: Cognitive BiasesType: PerceptionGrade
Also called
Attribute framing, Risky-choice framing
Meaning
Reaching different decisions from the same facts depending on how they are presented.
Example
"90 percent of grants succeed" wins approval where "1 in 10 grants fails" would not.
Mitigation
Restate the option in the opposite frame before you decide.
Open the full entry: Framing effect
GroupthinkCategory: Cognitive BiasesType: SocialGrade
Meaning
A group's desire for harmony overrides a realistic look at the alternatives.
Example
Nobody objects in the meeting; three people raise concerns in the hallway afterward.
A caution
Janis's full model has weak support; the core danger (pressure to agree) is well supported.
Mitigation
Collect inputs privately before discussion. Invite the quiet voices at the edges first.
Open the full entry: Groupthink
HALT: Hungry, Angry, Lonely, TiredCategory: Emotional BiasesGrade
Also called
HALT check
Meaning
A checklist from recovery programs: hungry, angry, lonely or tired. The idea is that these states make poor choices more likely.
Example
A late-night email approving a vendor change goes out after a 14-hour day.
A caution
A rule of thumb from recovery programs. Sleep loss and anger do change judgment. Two famous findings often cited for it are disputed: the "hungry judges" study, and "ego depletion" (willpower as a fuel that runs out), which failed large replications. Treat HALT as a sensible pause.
Mitigation
Before an important decision, check HALT. If any apply, set a time to decide tomorrow.
Open the full entry: HALT: Hungry, Angry, Lonely, Tired
Hindsight BiasCategory: Cognitive BiasesType: MemoryGrade
Also called
"Knew-it-all-along" effect
Meaning
After an outcome is known, believing you saw it coming all along.
Example
After a launch fails, everyone remembers having doubts that no one wrote down.
Mitigation
Journal your expectation and your reasons before the result arrives, then compare.
Open the full entry: Hindsight Bias
Hyperbolic DiscountingCategory: Behavioral EconomicsType: MoneyGrade
Also called
Present bias (related)
Meaning
Preferring smaller rewards now over larger rewards later, more than reason supports.
Example
A program skips the data cleanup that would pay off next year for a quick dashboard this month.
A caution
Some patterns are better explained by uncertainty than by hyperbolic curves.
Mitigation
Put the later payoff in today's terms and commit in advance.
Open the full entry: Hyperbolic Discounting
Imposter SyndromeCategory: Emotional BiasesGrade
Also called
Impostor phenomenon
Meaning
Doubting your own competence despite evidence, and fearing you will be found out.
Example
A capable new director defers every call to the board rather than trust her own analysis.
A caution
A self-perception pattern, not a clinical diagnosis or a judgment bias.
Mitigation
Write down the evidence for your competence and decide on the ingredients, not on your nerves.
Open the full entry: Imposter Syndrome
Loss AversionCategory: Behavioral EconomicsType: MoneyGrade
Meaning
Losses feel about twice as painful as equal gains feel good.
Example
A foundation turns down a promising program to avoid a small chance of a public misstep.
A caution
Well supported on average: a 2024 meta-analysis finds losses weigh about twice as much as equal gains. The effect is smaller, or missing, for small routine stakes, and a minority of researchers argue it is overstated.
Mitigation
Frame each option as both a gain and a loss, and compare expected values.
Open the full entry: Loss Aversion
Optimism biasCategory: Emotional BiasesGrade
Also called
Unrealistic optimism
Meaning
Believing you are less likely than others to experience a bad outcome.
Example
A team assumes its data migration will be the one that has no surprises.
A caution
Optimism in plans, budgets and forecasts is well documented. Some lab measures of "unrealistic optimism" have been criticized as statistical artifacts (Harris & Hahn 2011).
Mitigation
Run a premortem and assign a probability to each failure path.
Open the full entry: Optimism bias
OverconfidenceCategory: Cognitive BiasesType: PerceptionGrade
Also called
Overconfidence effect, Overprecision, Overestimation, Overplacement
Meaning
Being more sure of your judgments than the evidence justifies.
Example
A team gives a 90 percent chance of hitting a date it has missed three times before.
Mitigation
Think probabilistically and speak probabilistically: give ranges, then check your track record.
Open the full entry: Overconfidence
Post-mortem analysisCategory: Best PracticesType: LearningGrade
Also called
Postmortem, Lessons learned
Meaning
A structured review after an outcome to learn what happened and why.
Example
After a grant cycle, the team compares what it expected with what happened, criterion by criterion.
Mitigation
Review the process, not just the outcome; good decisions can still meet bad luck.
Open the full entry: Post-mortem analysis
Pre-mortem analysisCategory: Best PracticesType: LearningGrade
Also called
Premortem, Prospective hindsight
Meaning
Imagine the decision has already failed, then explain why it failed, before you act.
Example
Before a migration, the team lists the five most likely reasons it would be a disaster in six months.
A caution
Promising, but the evidence is thin: a few studies show it surfaces more risks and trims overconfidence. There are no large trials on decision outcomes.
Mitigation
Why wait for the postmortem? Run the premortem. Assign an owner to each failure path.
Open the full entry: Pre-mortem analysis
Red TeamingCategory: Best PracticesType: LearningGrade
Also called
Red team
Meaning
Assigning people to attack a plan the way a skeptic or adversary would.
Example
A red team argues that a new portal will overwhelm the two staff who must approve submissions.
Mitigation
Red team your decisions before they are made. Give the red team real time and real authority.
Open the full entry: Red Teaming
Scenario PlanningCategory: Best PracticesType: LearningGrade
Also called
Scenario analysis
Meaning
Building several plausible futures and testing a decision against each one.
Example
A ministry tests its five-year plan against a currency shock, a leadership change and a giving surge.
Mitigation
Pick scenarios that would change your choice, not just ones that are easy to imagine.
Open the full entry: Scenario Planning
Status Quo biasCategory: Cognitive BiasesType: BeliefGrade
Also called
Default bias (related)
Meaning
Preferring things to stay as they are, even when a change would serve you better.
Example
A ministry keeps a spreadsheet process that costs a full day a week because switching feels risky.
Mitigation
Score the current state as one of the alternatives, with the same criteria as the others.
Open the full entry: Status Quo bias
Straw Man fallacyCategory: Cognitive BiasesType: BeliefGrade
Also called
Straw man
Meaning
Misrepresenting someone's argument so it is easier to knock down.
Example
"So you want to stop funding the field entirely?" in reply to a request for one more site visit.
Mitigation
Restate the other person's argument until they agree it is fair, then respond.
Open the full entry: Straw Man fallacy
Sunk Cost fallacyCategory: Behavioral EconomicsType: MoneyGrade
Also called
Sunk cost effect, Concorde fallacy
Meaning
Continuing an effort because of what you already spent, not because of what it will return.
Example
A failing system stays in place because "we already paid for three years of licenses."
A caution
Reliable on average. Size varies by the kind of decision, and it is strongest when people feel personally responsible for the first choice.
Mitigation
Ask: if we were starting today, would we choose this? Judge only future costs and benefits.
Open the full entry: Sunk Cost fallacy
Survivorship biasCategory: Cognitive BiasesType: MemoryGrade
Meaning
Drawing conclusions only from the cases that made it through a filter, and ignoring the ones that did not.
Example
A grant committee studies its five most successful projects and copies their traits, never looking at the projects that failed.
Mitigation
Ask what the missing cases would show. Pull the full list, including the failures.
Open the full entry: Survivorship bias
The IKEA effectCategory: Behavioral EconomicsType: PerceptionGrade
Also called
Effort justification (related)
Meaning
Placing extra value on things you built yourself.
Example
A custom intake form survives every review because the team assembled it by hand.
Mitigation
Have someone outside the build team score it against the alternatives.
Open the full entry: The IKEA effect
The Mom TestCategory: Best PracticesType: LearningGrade
Meaning
Ask people about their past behavior, not their opinion of your idea, so even your mom cannot lie to you.
Example
Instead of "Would you use a donor portal?" ask "How did you last check on a gift you made?"
Mitigation
Ask about specifics in the past. Listen for what people already spend time and money on.
Open the full entry: The Mom Test
The Ulysses ContractCategory: Behavioral EconomicsType: MoneyGrade
Also called
Ulysses pact, Commitment device, Precommitment
Meaning
A decision made in advance that binds your future self against a known temptation.
Example
A board agrees before the cycle opens that no grant skips due diligence, however urgent it feels.
Mitigation
Write the rule down while calm, with a named person to hold you to it.
Open the full entry: The Ulysses Contract
Zeigarnik effectCategory: Cognitive BiasesType: MemoryGrade
Also called
Ovsiankina effect (related)
Meaning
The idea that unfinished tasks are remembered better than finished ones. Research does not reliably show a memory edge, but people do feel pulled to go back and finish what they started.
Example
A director leaves a staffing decision open on Friday. All weekend she keeps returning to it in her head. On Monday it still pulls at her attention during unrelated budget talks.
A caution
A 2025 meta-analysis found no reliable memory advantage for unfinished tasks. The urge to go back and finish them (the Ovsiankina effect) is better supported.
Mitigation
Close the loop: decide, or set a dated check-in so the decision can rest.
Open the full entry: Zeigarnik effect

All entries