Action Produces Information
Take a bounded step into uncertainty to reveal the next better move
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 5
- Confidence
- 99%
When the path ahead is unclear, Armstrong recommends replacing endless hypothetical debate with a small action that can generate evidence. The first move does not have to be correct; its purpose is to expose reality that planning cannot reveal. Before Coinbase existed, Armstrong and a friend built an Android Bitcoin wallet that ran a full node on the phone. Launching it immediately revealed an unusable multi-minute synchronization delay. That failure clarified that key blockchain services would need to run in the cloud and helped make the next product direction visible. The method works by choosing a bounded experiment, observing what breaks or surprises users, and allowing that information to guide the next move.
Origin
Armstrong describes this as a favorite prism from the Y Combinator and Airbnb ecosystem and illustrates it with his pre-Coinbase Android wallet.
Core principles
- 01Thinking alone cannot clear every unknown
- 02A small action can expose hidden constraints
- 03A wrong first move can still reveal the right second move
- 04Bounded experiments turn fog into evidence
How to run it
- 1
Name the fog
Write down the decision you cannot resolve and identify which fact is missing. Distinguish a real unknown from a fear stated as a hypothetical.
Pro tip Frame the unknown as something observable.
Watch out Do not keep debating a question only the market or product can answer.
- 2
Bound an experiment
Select the smallest reversible action that puts the assumption in contact with reality. Decide what result would be informative before starting.
Pro tip Optimize the first action for learning, not polish.
Watch out Avoid risking an amount you cannot recover from.
- 3
Act before certainty
Run the experiment even though it may be the wrong implementation. Pay attention to friction that was invisible in planning.
Pro tip Launch enough of the real experience to expose constraints.
Watch out Do not defend the prototype when its purpose is to teach.
- 4
Extract the information
Identify what the action made newly clear, especially failures or user behavior. Translate that evidence into a more precise next decision.
Pro tip Ask what you now know that you could not know before acting.
Watch out A failed implementation is not a failed experiment if it resolves the unknown.
- 5
Take the clearer step
Use the new information to run the next bounded action. Continue until the direction is clear enough for a larger commitment.
Pro tip Let each experiment narrow the search space.
Watch out Do not turn the method into random activity without a learning question.
In the wild
Armstrong and a friend built and launched a simple Android Bitcoin wallet while he still worked at Airbnb. The phone had to synchronize a full Bitcoin node whenever the app opened, producing a delay of several minutes. Seeing that failure in use made him realize that a serious company would need to run synchronization, backups, and security in the cloud.
→ A flawed prototype revealed the architecture and company opportunity behind Coinbase.
Common mistakes
Waiting for a complete map
Some facts emerge only after a product, conversation, or experiment encounters reality.
Taking action without a question
Motion produces little useful information when the builder has not named what the action should test.
Is it for you?
Best for
Builders facing a new problem where discussion cannot resolve the central unknown.
Not ideal for
Irreversible decisions whose downside cannot be bounded before acting.
From the transcript
“action produces information right”
“if you just take some steps into the unknown you might not be doing the right thing but oftentimes by doing that action you'll realize…”
“just taking action produced information which made the next step a little bit more clear”
From the episode
#627: Brian Armstrong, CEO of Coinbase — The Art of Relentless Focus, Preparing for Full-Contact Entrepreneurship, Critical Forks in the Path, Handling Haters, The Wisdom of Paul Graham, Epigenetic Reprogramming, and Much More
Brian Armstrong