Process-Over-Outcomes Discipline
Precommit to an advantaged process before judging the result
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 96%
Process-over-outcomes discipline separates decision quality from the luck embedded in any single result. Before acting, the decision-maker defines how the problem will be approached, what evidence will count, and why that approach should provide an advantage. Assumptions and criteria are recorded so they cannot be rewritten after the result is known. Afterward, the review considers both what happened and whether the predefined process was followed. A favorable result reached through weak reasoning should not be celebrated as proof, while an unfavorable result reached through a sound probabilistic process should not automatically trigger abandonment. Across repeated decisions, this creates cleaner organizational learning, more consistent execution, and less hindsight bias. The process itself evolves when accumulated evidence reveals a genuine weakness.
Origin
Matthew Granade attributed his commitment to process over outcomes to his Bridgewater training during his Masters in Business interview.
Core principles
- 01A good outcome does not prove that a decision was good.
- 02A bad outcome does not prove that a sound process was wrong.
- 03Design the approach before uncertainty resolves.
- 04Explain why the chosen process should create an advantage.
- 05Improve repeatable decisions by auditing how outcomes were reached.
How to run it
- 1
Define the decision
State the choice, objective, constraints, and uncertainty before the outcome is visible.
Pro tip Use a short written decision record.
Watch out A vague objective makes later evaluation subjective.
- 2
Predefine the approach
Specify the evidence, analysis, criteria, and sequence that will govern the decision.
Pro tip Set thresholds for acting, waiting, or rejecting.
Watch out Do not move criteria simply because an attractive option appears.
- 3
State the expected advantage
Explain why this process should outperform a plausible alternative.
Pro tip Identify the mechanism rather than relying on reputation or habit.
Watch out Elaborate process without an advantage is bureaucracy.
- 4
Execute and document
Follow the agreed process while recording deviations, new evidence, and the final rationale.
Pro tip Separate contemporaneous notes from retrospective commentary.
Watch out Undocumented exceptions make honest review difficult.
- 5
Review process and outcome separately
Assess whether the process was sound and followed before considering how the decision turned out.
Pro tip Ask what would have been concluded if the outcome were hidden.
Watch out Do not let one lucky success validate a weak method.
- 6
Update from repeated evidence
Change the process when a pattern of decisions reveals a structural weakness or a better mechanism.
Pro tip Aggregate comparable decisions before making major revisions.
Watch out Overreacting to one result reintroduces outcome bias.
In the wild
An investment team defines its market, product, customer, management, and return tests before diligence begins. After investing, it evaluates whether those tests were completed and correctly interpreted separately from whether the startup happened to succeed.
→ The team learns which parts of its method predict quality without mistaking luck for skill.
A research group predefines its validation periods, success thresholds, and deployment rules. A profitable live month is not enough to excuse leakage or undocumented parameter changes.
→ Only results produced through a credible development process influence future capital allocation.
Common mistakes
Celebrating lucky wins
A positive outcome can conceal weak reasoning and encourage the organization to repeat an unreliable decision process.
Punishing sound losses
Probabilistically correct decisions sometimes fail, and abandoning them after one loss can damage long-term performance.
Designing the process afterward
Post hoc criteria allow hindsight to rewrite what the decision-maker actually knew and intended.
Is it for you?
Best for
It is best for repeated decisions made under uncertainty, including investing, hiring, experimentation, and product selection.
Not ideal for
It is not ideal for deterministic tasks where outcomes directly and reliably reveal execution quality.
From the transcript
“I'm a very big believer and I think this comes from my my Bridgewater training UM in sort of process over outcomes.”
“you have to think ahead of time about how you're going to approach a problem and why that's going to give you an advantage in…”
“you don't you don't look at outcomes, you look at how you got to those outcomes.”
From the episode
Matthew Granade Discusses Quantitative Investment (Podcast)
Matthew Granade