Prototyping

Overview

A prototype is the cheapest way to test whether an idea is worth building. We build prototypes that prove the concept, or kill it, before serious money is committed to a full implementation.

The point of a prototype is not to look impressive. The point is to answer a specific question quickly, with the smallest possible investment.

When to Prototype

Prototyping pays off in a handful of specific situations:

When stakeholders cannot agree on the requirements: a working artefact moves the conversation from abstract opinion to concrete reaction.
When a new technology is being adopted: a small build surfaces the awkward edges that the vendor's slide deck did not mention.
When an unfamiliar problem domain is being entered: the act of building something forces the team to confront the parts they do not yet understand.
When an investment decision is pending: a prototype de-risks a large commitment with a small one.

Types of Prototype

Different questions call for different prototypes. We typically build one of:

Clickable mockups (Figma): for testing user-experience flows and getting stakeholder feedback on layout and journey, without writing any code.
Interactive HTML prototypes: one step beyond clickable, useful when the interaction model itself needs to be validated.
Functional spike implementations: small builds against the real technology stack, to confirm that a chosen approach is feasible.
Throwaway proof-of-concepts on real data: the most expensive prototype, but the most convincing, for confirming that an idea works against the actual messy data the production system will face.

Output

Every prototype engagement ends with two things:

A working artefact you can show to stakeholders, demo to a board, or hand to a usability tester.
A clear written recommendation: proceed, redesign, or stop, and if proceed, what to build next and what to discard.

Our Approach

Prototypes are deliberately disposable. We build the cheapest thing that proves or disproves the point, then we throw it away and build the real thing properly. Trying to "evolve" a prototype into the production system is one of the more reliable ways to ship a fragile product.

We are honest about which mode we are in. A prototype is not a beta. We do not pretend otherwise to inflate a deliverable.