Between the idea and the prototype there is a stretch that tends to be underestimated. It is not the technical stretch — it is the one where you decide what, exactly, the prototype has to prove. Skip that step and you build a handsome object that answers nothing. What follows is a practical route.
1. Write the hypothesis before you draw
State in one sentence what you believe and have not yet proved. “I believe a vibration sensor fitted to the sieve can flag misalignment before it affects the product.” That sentence defines what the prototype measures and what can be left out.
Without it, every detail looks equally important — and the project grows out of control.
2. Decide which risk you are attacking
Prototypes serve different risks, and the right type depends on which risk dominates:
- Desirability risk — does anyone want this? A look-alike prototype, with no function, already answers.
- Technical feasibility risk — does it work? Here appearance is irrelevant; a bench with loose wires is enough.
- Usability risk — can the person operate it? This needs shape, weight and interface close to the real thing.
Trying to settle all three in the same prototype is the mistake that delays projects most.
3. Use the cheapest material that still answers
Cardboard, foam and a polymer printed part answer far more questions than people expect. Moving up in material too early costs money and, worse, creates attachment: whoever machined it in aluminium resists changing the design.
4. Define the success criterion before the test
Write down, before testing, what will count as a pass. Without that, the result gets read in favour of the expectation — and every prototype “sort of works”.
A good criterion is verifiable: “detects the misalignment within 30 seconds, in 8 out of 10 runs”. A bad one is “it turned out fine”.
5. Test with whoever will use it, not with whoever built it
Whoever designed it knows where to press. Hand the prototype to someone who has never seen it, ask them to carry out the task and simply watch — without explaining. The discomfort of watching someone get it wrong is the most useful information in the process.
6. Record what did not work
The discarded path has value. With no record, six months later someone on the team proposes again the alternative that was already tested and failed — and the cycle repeats.
If you have reached the hypothesis and want to come out of it with a working prototype, talk to LIPS. We work with companies, institutions and researchers, and the first conversation exists precisely to define what the prototype has to prove.