TThe Diary of a CEO
← All frameworks
Innovation

Bias-to-Action Experiment

Replace prolonged certainty-seeking with a small test that teaches

Difficulty
Easy
Time to result
~weeks to results
Steps
6
Confidence
96%

The Bias-to-Action Experiment treats action as a way to learn rather than merely the final step after complete analysis. Define the uncertainty, form a hypothesis, and choose a small, bounded test that can distinguish useful alternatives. Set the evidence and time frame before starting, run the experiment, and judge the quality of the method separately from whether the outcome happened to be favorable. Bartlett describes rewarding teams for conducting the experiment rather than for getting a positive result, while Pink frames planning and certainty as overvalued relative to action and experimentation. The mechanism is faster feedback: a contained test can answer a practical question that prolonged speculation cannot. The framework applies only where experimentation is responsible and reversible; it does not justify exposing people to hidden risks or bypassing safeguards.

Origin

Pink and Bartlett develop the method while discussing inaction regrets, business procrastination, and the scientific habit of testing hypotheses.

Core principles

  • 01Complete certainty is often unavailable before action
  • 02Acting can be a form of figuring out
  • 03Small tests reduce the cost of learning
  • 04Teams should distinguish experiment quality from outcome luck
  • 05New evidence should trigger the next test

How to run it

  1. 1

    Name the uncertainty

    State the decision you cannot resolve through current information alone. Separate the unknown from facts already established.

    Pro tip Write the decision as a choice between observable alternatives.

    Watch out Do not experiment around legal, ethical, or safety constraints.

  2. 2

    Form a hypothesis

    Predict what will happen under one bounded change and why. Define what result would count against the prediction.

    Pro tip Prefer a hypothesis that can be disproved by the test.

  3. 3

    Shrink the test

    Choose the smallest action that can generate decision-relevant evidence without creating disproportionate cost or harm.

    Pro tip Test one important variable where practical.

    Watch out A test too small or noisy may create false confidence.

  4. 4

    Set the frame

    Specify the duration, evidence, stop conditions, and decision threshold before starting. Make the experiment finite.

    Pro tip Record the baseline so change can be compared honestly.

    Watch out Do not move the success criterion after seeing the result.

  5. 5

    Run and observe

    Execute the test and collect the agreed evidence. Evaluate whether the process was sound independently of a favorable or unfavorable outcome.

    Pro tip Reward disciplined learning rather than lucky results.

  6. 6

    Decide the next move

    Use the result to proceed, stop, revise the hypothesis, or run a better test. Convert learning into another concrete decision.

    Pro tip Write what the experiment changed in your view.

    Watch out Repeated testing without a decision can become another form of avoidance.

In the wild

Testing instead of waiting two years

Bartlett contrasts spending years considering a business idea with running a roughly three-month experiment that can reveal whether to continue. He says his teams used diligent time frames and were rewarded for performing the experiment rather than for producing a favorable outcome.

The organization receives evidence sooner while reducing pressure to manufacture a winning result.

Common mistakes

Waiting for certainty

Some practical unknowns remain unresolved until a real action produces feedback.

Rewarding only positive outcomes

Outcome-based incentives can discourage honest tests or encourage teams to manipulate the apparent result.

Running an unbounded test

An experiment without limits, safeguards, or a decision rule can create more risk than learning.

Is it for you?

Best for

Reversible business and creative decisions where a bounded experiment can produce meaningful evidence.

Not ideal for

Irreversible, unsafe, unethical, or high-stakes actions that require stronger evidence and approval before testing.

From the transcript

acting is a form of figuring it out

Daniel Pink · (1:29:00)

planning as an asset, overvalued; acting as an asset, undervalued

Daniel Pink · (1:30:30)

they have a hypothesis, they test the hypothesis

Daniel Pink · (1:32:00)

From the episode

The Real Trick To Long Term Motivation: Daniel Pink