Looks-like or works-like: which prototype do you build first?

A functional prototype, also called a works-like, working or engineering prototype, has the real mechanism, electronics and battery inside the real volume. It is built to prove that the product works in its final form, even if it does not look finished. It is tested until it fails, because a failure found at this stage costs a revision, not a mold.

Two other things go by a similar name. In software and UX design, a "functional prototype" is a clickable mockup of an app, which is not what this article is about. And a proof of concept is not a functional prototype either: it tests one principle on its own, with a development board and parts bought off the shelf, before the product has a shape. The works-like comes later and puts that principle inside the product.

We build works-like prototypes from CNC-machined parts and cast parts in materials close to the production ones, with custom PCBs and firmware when the product has electronics.

A visual prototype, also called a looks-like prototype, an appearance model or an aesthetic prototype, has the final shape, colors and finish, and nothing inside. It is made for hands, photos and investors. It tells you whether the size, the grip and the finish are right, whether the color reads as premium and whether the buttons fall under the thumb. It cannot tell you whether the engineering fits inside it.

A looks-like is machined or printed, then finished and painted to the final color and texture. It often hides how it was put together: glue where screws will go, parts that will never come apart. That is fine for a photo, and it is also why a looks-like proves nothing about assembly.

Neither one tells you the unit cost. That question belongs to design for manufacturing and to a small batch, after both.

Because the works-like needs a volume to fit into. A works-like puts the real parts inside the final shape. If that shape changes after the works-like is built, the engineering is done twice. Settling the shape first, with a looks-like model people have held, means the works-like is built once, inside a shape that will not move.

Many guides give the opposite rule: build the works-like first when there is technical risk. We separate two kinds of doubt. A doubt about the principle, whether the idea works at all, is the job of a proof of concept, and it comes before both prototypes. A doubt about integration, whether the real parts fit, stay cool and survive inside the real shape, is the job of the works-like, and it needs the shape.

Two cases skip the separate looks-like. If the shape is already settled, by an existing product line or a brief that fixes it, go straight to the works-like. And if your product uses known technology in a new shape, there is no proof of concept to make: you go straight to development.

Not sure which case is yours? Six questions tell you which prototype to build next.

Yes, and most projects get there: a later round often combines both. It takes more work, because the shell has to be finished like a looks-like while everything inside works like a works-like. It is worth it when the same prototype has to be filmed for a launch and used for real.

Photonix, a solar lawn mower, is one of those. Its full-size prototype was built to be filmed for the launch, and it was fully functional. The shell was shaped from CNC-cut frames, foam and fiberglass, then painted. Underneath, a chassis plate cut on a CNC carried the batteries, the drive and a servo that set the cutting height from the handlebar. It mowed real lawns, its solar panel charged the batteries, and every feature in the brief worked.

A prototype that looks and works like the product answers the first two questions at once. It still leaves the third one open: whether the product can be made in numbers at the right cost. That is the job of design for manufacturing, which comes next.

It depends on who is looking. Many founders raise their first money at the concept stage, with renders and a brief. A campaign page runs on how the product looks and what it does. On Kickstarter, the rules for hardware projects ask creators to show working prototypes and do not allow photorealistic renderings, so the works-like has to exist before launch (Kickstarter's rules for hardware and product design projects).

KORU, an air purifier and smart planter, funded on Indiegogo at more than ten times its goal with the renders we made for it. Its team wrote that the renders are often confused with product photos.

Whatever the platform, backers expect delivery. The works-like has to be on its way before the campaign, not after it.

For a molded consumer product, the prototype step of our guide takes about 16 weeks, with an indicative budget of $7,500 to $30,000 for the looks-like and works-like rounds together. These are orders of magnitude, not a quote: the effort can vary fifty times from one product to another, so we give a fixed price once the brief is written.

Single builds come faster. A simple FDM print takes 24 to 48 hours, CNC-machined parts typically 5 to 7 days, and a fully assembled functional prototype 2 to 4 weeks, depending on its complexity and on component lead times.

Three things move the price most: the number of parts, electronics and a battery inside, and certification. The full breakdown is in time and cost, in the guide.

Design for manufacturing, then a small batch. Once the works-like holds, every part is reviewed for the way it will be molded and assembled. Then, before any steel mold, 10 to 200 units are cast from one master part in silicone molds, by vacuum casting or gravity casting in resin: units for testers, certification labs and first backers. Only then does a factory cut the production mold.

A factory quotes what it can see. Without a works-like, the quote is a guess, and so is your margin. The small batch step of the guide explains what that batch is for.