Proof of concept, cheap and early.
Also called: Proof of concept · PoC · Functional rig · Principle test
Building the roughest possible rig that proves the core principle works, before spending money on anything that looks like the product.
A proof of concept tests one thing: does the central idea actually work. It should be ugly, cheap and fast, deliberately not the product. Prove the risky principle before you spend on tooling or finish. The pitfalls and checklist below keep it honest.
What a proof of concept is
Every product rests on one or two assumptions that, if wrong, sink it. The proof of concept is the cheapest possible test of those assumptions. It is not a prototype of the product; it is a rig that isolates the risky principle and proves it works, using whatever is to hand. Cardboard, breadboards, off-the-shelf parts, anything that answers the question without committing to the product.
Find the one thing to prove
Start by naming the single assumption that, if it fails, means there is no product. Not the things you already trust, the one you genuinely do not know. Then build the smallest, ugliest thing that can answer it, and decide in advance what result counts as a pass. The build should be embarrassing to look at and decisive to read.
How you prove it depends on what you are unsure of:
Breadboard the electronics. Test the sensing, control or output before you commit to a real board.
Rig the mechanism. Off-the-shelf parts and a rough fixture prove a motion, a load or a seal.
Mock the interaction. Watch a real person attempt the real action on a stand-in, and note where they hesitate.
Stress the unknown. Push the one material, module or condition you doubt until it tells you whether it holds.
For the proofing box, the one make-or-break unknown was thermal: could a low-power heater hold 26°C to within half a degree overnight, in a cold kitchen, with real dough soaking up heat. The rig ignored the product entirely. Here is the rig sheet from that proof of concept, so you can see the shape of a good answer rather than a generic template.
Pitfalls to avoid
Overbuilding. A proof of concept exists to prove one thing. Spending time on detail or finish is spending it on what was never the risk.
Testing the wrong risk. Prove what you do not know, not what is already safe. A rig that confirms the easy part teaches you nothing.
Hiding the jank. If it works, show it, duct tape and zip ties and all. A polished rig invites you to believe the product is further along than it is.
Skipping documentation. A rig that works once but never again has proved nothing. Record what you did and what happened, even a failure.
How it fits the bigger picture
Proof of concept prototype is activity 08.10.02 in the Develop stage. Behind it sit the technology review and the first physical model. Ahead of it sits the functional prototype, which builds the proven principle into something closer to the real product.
What it can do
Kill or confirm the project’s riskiest assumption for very little money, so you commit to expensive work only after the central idea has actually been shown to work.
What it can’t do
A rig that proves the principle does not prove the product; integration, finish and manufacture are all still ahead. Confusing “the idea works” with “the product works” is a classic and expensive mistake.
See the full 10-stage process →
Try it yourself
Write the one assumption that, if false, means you have no product. Decide in advance what result would count as proof. Then build the ugliest, cheapest rig that can test it, using whatever is to hand. Spend nothing on looks. Either it passes the bar you set, or you have just saved yourself a fortune.
Want a structured first pass? Start the Free Sprint → and the GPT will help you find the assumption worth testing first.
A good proof of concept, checked
Project notes: the bread-dummy rig
▸ From the notebook · optional reading
Project notes: half a degree, proved in a cardboard box
The proofing box’s whole viability rested on holding half a degree overnight. The Hartleys proved it in a crude rig before spending a penny on ceramic.
3 min read · click to open
The proofing box’s one make-or-break assumption was thermal: could a low-power heater hold 26°C to within half a degree, overnight, in a cold Stockport terrace, with two 1kg dough balls absorbing heat. If that did not work, nothing else mattered.
So the proof of concept ignored the product entirely. No ceramic, no wood, no nice OLED. Dan built a crude insulated box with a film heater, a sensor, a basic controller, and bags of flour-and-water standing in for dough. Total spend was maybe £60 of parts against a product that would retail at £149. It looked like a school project, which is exactly right.
We set the pass condition before building it: hold 26°C within half a degree across an eight-hour overnight run, drawing under 30W average. The rig did it, with a bit of control tuning. That single cheap result is what justified committing to the premium ceramic enclosure and the real board. I push every founder toward this order: prove the thing that could kill you cheaply, then spend on the things that make it sellable. Proving the half-degree in a cardboard box is the best-value afternoon the project ever had.
— Next in Develop → Presentation model