Start → Field guides

Bearing 270

Field guides

Short, practical guides for people facing software decisions.

Guide 1

Build, buy or adapt?

Buy when your process is common and a mature product already fits most of it. Accept its way of working for the parts that do not matter to you.

Build when the process is what makes your organisation distinctive, or when no product fits without heavy compromise.

Adapt when you already have something that works partly; improving it is often faster and less disruptive than replacing it.

Write down the two or three criteria that matter most before looking at any options, so the options cannot quietly redefine the criteria.

Guide 2

Reading a discovery report

A good report names the users and their real tasks, separates evidence from assumption, compares at least two options with costs and risks, and makes a recommendation. It includes something tested with people, even if only a paper prototype.

Warning signs: a feature list with no users in it, no options other than the one being sold, and estimates more precise than the uncertainty allows.

Guide 3

Signs of a lost project

  1. Nobody can say what the next demo will show.
  2. Status reports describe activity, not results.
  3. The list of "nearly done" items keeps growing.
  4. Users have not seen anything for over a month.
  5. Decisions are revisited without new information.
  6. The original problem is no longer mentioned.
  7. Only one person understands how to deploy.

To reorient: stop, list what works today, pick the single most valuable next result and plan a two-week leg to reach it.

Glossary

Map terms

Discovery: a short phase of research and prototyping before building. Spike: a time-boxed technical experiment. Increment: a small, complete piece of working product. Staged rollout: releasing to a share of users first.

Try the compass tools.