The Discovery Protocol: Agility Over the "Perfect" Plan

A plan is a hypothesis about what will happen. It becomes useful when reality is allowed to correct it.

Detailed planning has a place when the work and its operating conditions are reasonably understood. Early in a venture or uncertain project, precision can disguise assumptions that have not yet been tested. Good execution cannot rescue a plan built around the wrong problem.

Discovery and execution therefore need different modes of work.

When the model is known

When demand, delivery, and constraints are understood, process and efficiency help the organization perform consistently. Planning can coordinate people, resources, and dependencies because the underlying model has evidence behind it.

When the model is uncertain

When the problem, customer, or business model is still unclear, the immediate task is learning. The team needs to identify the assumptions that carry the most risk and find proportionate ways to test them.

This does not mean acting without a plan. It means treating the plan as revisable.

The discovery protocol

1. Isolate the critical assumption

Name the assumption that must be true for the project to work. Make it specific enough that evidence could strengthen or weaken it.

2. Choose the smallest credible test

Design a prototype, conversation, observation, or limited release that can produce relevant evidence. The smallest test is not always the fastest one. It must still represent the question being asked.

3. Decide what the result would change

Before running the test, state what evidence would cause you to continue, revise, or stop. Otherwise, it is easy to reinterpret every result as support for the original plan.

4. Involve the people affected

Bring customers, users, or operators into the process early enough to influence the work. Feedback is evidence, but it is not automatically representative. Pay attention to who responded, the conditions of the test, and what remains unknown.

5. Update the plan

Use what you learned to change the assumption, the proposed solution, or the next test. Do not optimize a process whose value has not yet been established.

Discovery does not guarantee success, and shipping quickly is not a virtue by itself. The purpose is to reduce consequential uncertainty before committing more time, money, and attention than the evidence supports.