A quote given before the prototype is a bet

The normal sequence is: you describe the project, a quote arrives, you sign, work starts. It works well for things that have been done a thousand times. For custom software it works badly, and the reason is simple: at the moment of the quote neither side yet knows what is being built.

What actually happens with an early quote

Whoever writes it has two options. The first is to build in a large margin to cover the uncertainty: the client pays more than was needed, and often never finds out. The second is to give a low number to win the work and then recover it through variations: the client pays more than they thought, and finds out halfway through.

In both cases the opening number was not an estimate. It was a bet, and the house always wins.

The prototype as a measuring instrument

We build the main screens — the ones where 80% of the time is spent — and we try them with you against the real process, with real data. Not a drawing: something you click.

Three pieces of information come out of that which did not exist before:

• How much there really is. Almost always less than was asked for, because trying it reveals that three screens imagined as different are the same screen with a filter.
• Where the difficulties are. Almost never where anyone expected. The piece that looked trivial is the one touching the system that will not expose its data.
• What the people using it actually want. A question nobody can answer in the abstract, and everybody can answer in front of a screen.

“But then I am paying twice”

No: the prototype is the first part of the work, not a preliminary to throw away. The screens you tested become the production screens, and the decisions made there are not remade.

What you avoid is the thing that genuinely costs: discovering at 70% of the build that the main flow needed thinking about differently. Changing halfway costs little; finding out at the end costs everything.

The rule that follows

If someone gives you a precise number for custom software without having seen how you work, they are not better than everyone else. They have simply decided in advance which of you will pay for the uncertainty.