Experimentation Toolkit

The operational detail behind every chapter, and why you write it down

The chapters of this book carry the judgment: what you are looking for, what makes a good one, how to know you have arrived. This part carries the procedure: one page per diamond, each running the same three phases, plus a gallery for the two artifacts people most often get wrong in ways that look right.

Page What it holds
Community Guides Diamond 1. Casting a wide net, committing to a community, testing access.
Community Choice Gallery Badly and well scoped communities side by side. Scope is a resolution setting, not a description.
Pain Guides Diamond 2. Conversation, observation, role play, secondary research, then clustering through to pain testing.
Pain Statement Gallery Weak pain statements beside strong ones, with what changed named.
Solution Guides Diamond 3. Generating many solutions, narrowing, testing before you build.

Each page runs in the same three phases as its diamond: diverge, converge, test. That repetition is deliberate. The shape is the method, and meeting it three times is how it becomes a habit rather than a diagram.

There is a fourth rendering of the same procedure. The Method Layer is written for an AI rather than for you, so a model can run the mechanical half alongside you. It is the same method these pages describe, addressed to a machine. Neither is the real one and neither is the lesser path.

Why You Document Every Experiment

Innovation begins in uncertainty. You do not pretend to know more than you do, and you do not guess blindly. You run experiments to reduce uncertainty, and you record them so that what you learned survives contact with next month.

Documentation is not bureaucracy. It is the act that turns an experience into evidence, and a hunch into something you can be wrong about in public.

Whether you are watching a person in a setting, testing a part, or sitting with a customer, the work counts only if it is recorded in a way you and others can still understand later. Your memory will smooth what happened into a story that supports whatever you now believe. Notes written at the time will not.

Without documentation there is no evidence. Without evidence there is no progress under uncertainty.

What Counts as Documentation

It does not have to look like a lab notebook. A field log, an interview summary, a data table, a short write-up, a photographed wall of sticky notes — any of these works, as long as it captures four things:

  • The logic. What you believed before, and what you were trying to find out.
  • The method. What you actually did, including what went wrong with it.
  • The findings. What happened, including the parts that were inconvenient.
  • The update. What you now believe, how strongly, and how you could still be wrong.

The fourth is what makes the record worth keeping. Findings without an update are a diary.

Written out as the sections of an actual record, those four become six:

Unknown → Design → Execution → Evidence → Knowledge update → Next steps.

The first is the logic, the middle three are the method and the findings, and the last two are the update. Every entry in the Halo Alert demonstration follows that shape, and it is worth copying because the order stops you writing the conclusion first.

Your Data Room

Everything you gather lives in one place, and that place has a name in this book: your data room. It holds more than results. It holds the design of each experiment, what you expected, what came back, and what you concluded — because six weeks from now the question will not be what did we find but why did we believe that, and only the design answers it.

Keep it somewhere shared and durable rather than in a notes app on one phone. Name files so that the sequence is visible at a glance:

exp-[##]-[short-title]-[date]
exp-03-cafeteria-observations-2026-09-16

Maintain a running index of experiments so you can see the whole expedition on one screen. When a test surprises you, that index is the first thing you will read.

Practices That Make the Difference

  • Document immediately. Not tonight. The details that matter are gone within the hour.
  • Be honest about surprises, gaps, and mistakes. A record that only contains successes cannot teach you anything, because you already knew the successes.
  • Write to your future self and your teammates. Both of them will have forgotten the context you are currently assuming.
  • Link the raw material. Recordings, images, transcripts, and source documents attached to the entry, not scattered.

Most of your experiments will fail in the ordinary sense. That is expected, and it is not the measure. What matters is that each one teaches you something you could not have known otherwise, and that you captured it, questioned it, and acted on it.

This is not extra work alongside building a venture. Under uncertainty, this is the work.