Experiment-to-Best-Practice Loop
Plan failures for learning, extract the data, and carry only what works forward
- Difficulty
- Moderate
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 98%
The loop reframes each project or product as an experiment rather than a final verdict on the entrepreneur. Before acting, define what is being tested and plan the experiment carefully enough that the result will produce usable data. Run it, observe both success and failure, and separate transferable best practices from assumptions, shortcuts, and mistakes that should be left behind. Carry the successful practices into the next build and repeat. Blair compares this to building houses: every completed attempt gives the builder knowledge that makes the next structure stronger and more valuable. The mechanism is not failure by itself. Progress comes from disciplined evidence extraction, deliberate reuse, and repeated construction, so the founder's accumulated judgement improves with every cycle.
Origin
Blair compares his entrepreneurial companies to successively larger houses, each built with best practices learned from prior successes and failures.
Core principles
- 01Projects and products are experiments under uncertainty
- 02Failure becomes useful only when it produces learning
- 03Meticulous planning makes evidence extractable
- 04Best practices should travel into the next build
- 05Repeated building improves future judgement
How to run it
- 1
Frame the experiment
Define the product, project, or operating belief being tested and the expected result. Make uncertainty explicit.
Pro tip Write the prediction before seeing the outcome.
- 2
Design for evidence
Choose the observations and data that will distinguish success, failure, and ambiguity. Plan the test meticulously enough to learn from it.
Watch out An unmeasured attempt can fail without teaching you why.
- 3
Run the build
Execute the smallest credible version that can answer the question. Treat setbacks as test results rather than personal condemnation.
Watch out Do not use the experiment label to excuse avoidable harm.
- 4
Extract the practices
Identify which actions produced useful results and which assumptions or decisions created failure. Convert the reliable successes into reusable practices.
Pro tip Describe the mechanism, not merely the outcome.
- 5
Build again
Apply the retained practices to the next project while discarding the failed assumptions. Design the next experiment around the remaining uncertainty.
Watch out Repeating the same design without carrying forward evidence is not iteration.
In the wild
Blair describes his first company as a house that became worth $25 million and a later one as a house worth roughly $700 million. He expects the next build to benefit from the old houses' best practices while leaving prior failures and shortsighted decisions behind.
→ Each venture becomes training data for a stronger subsequent company.
Common mistakes
Romanticizing failure
Failure has no automatic value; the experiment must generate data and change the next attempt.
Carrying every old practice forward
The loop improves performance only when failed assumptions and shortsighted choices are deliberately left behind.
Is it for you?
Best for
It is best for founders developing products, projects, or operating methods under uncertainty.
Not ideal for
It is not ideal for reckless bets where failure creates irreversible harm or where no evidence can be collected.
From the transcript
“if you view every project and every product as an experiment and experiments fail all the time that's what they're designed to do but from…”
“i plan my experiments in a very meticulous way so that i can extract the data from it and then i can apply the successes…”
From the episode
Ryan Blair: Conscious Business
Ryan Blair