Concept and draft
Research, a prototype you can operate, tests with the people meant to work with it, and several loops out of that.
Between the commission and the first line of code lies the phase that decides whether the right thing gets built. It is the shortest and the most consequential.
How it runs
Research comes first. I gather data about the people who will use it, or work with what is already there. Out of that comes a proposed concept, and out of that, as early as possible, a prototype you can operate.
I take that to the people meant to work with it later, collect their responses and rebuild. That takes several loops, not one. Stopping after the first means you paid for the prototype and left the insight behind.
In parallel I check what can be built and what can be run afterwards. Feasibility and economics decide along the way, not at the end.
What stands at the end
A draft you can put in front of someone, and the findings from the tests: what worked, what did not, and what was changed as a result.
The most common outcome is something considerably simpler than what we started with. Talking to users early means cutting features rather than building them.
What does not happen in this phase
No specification for the drawer. A document nobody reads costs more than a prototype somebody tries.
And no finished design. What is made here looks deliberately unfinished: a draft that looks like a product gets feedback on colour instead of on substance.
Frequent questions
How long does this phase take?
That depends on how much is already settled. The number is in the scope from the first conversation, and it is the one that shrinks most if you bring groundwork with you.
Can we skip this phase?
If it really is settled what should be built, yes. It is settled less often than it feels.