spike

A spike is a throwaway experiment used to validate ideas and test feasibility fast. Use this rapid-learning approach for early-stage tasks where quick, iterative experiments determine the viability of an approach before committing to a full build.

To decompose an idea into testable spikes, break the concept down into 2 to 5 independent feasibility questions. Each question is tackled as a separate spike, allowing you to align on ordering, scope, and acceptance criteria before any implementation begins.

The best way to structure throwaway experiments is to apply a workflow that requires a clear decomposition step, lightweight execution artifacts, and a concise verdict write-up. This demonstrates outcomes with observable results and learnings to guide real development.

Yes, you can use rapid-learning spikes to evaluate risks early. Spikes provide a lightweight, disposable approach to surface feasibility, required research, and potential risks before committing resources, saving time by validating approaches upfront.

You should avoid using disposable prototype experiments when a task is not early-stage or lacks clear decomposition steps. Spikes require structuring ideas into discrete, testable questions, so they are less suited for tasks with predetermined approaches or no independent feasibility variables.