- Define the real goal.
- Ignore the current implementation.
- State the desired outcome in one sentence.
- Identify the current mechanism.
- What specific approach are you using to achieve the goal?
- Ask: "If I couldn't use this mechanism, how else could I accomplish the same goal?"
- Brainstorm 2–5 fundamentally different pivots.
- Prefer pivots that remove complexity or test one assumption.
- They do not have to be production-ready.
- Choose the cheapest pivot.
- The goal is to learn something quickly, not build the final solution.
- Execute the pivot.
- Observe what changes.
- Record what assumptions were confirmed or eliminated.
- Either:
- Keep the pivot if it's better.
- Or use what you learned to improve the original implementation.
- Repeat until the goal is achieved.
Principles
- Preserve the goal, not the implementation.
- Every pivot should teach you something.
- Prefer experiments over theories.
- Reduce assumptions whenever possible.
- Temporary solutions are valuable if they increase understanding.