The day a prototype finally works is genuinely worth celebrating. It is also the day most inventors quietly stall, because the next step is not obvious in the way the previous ones were. Building the thing had a clear goal: make it function. What follows has five goals running at once, most of them commercial rather than technical, and the natural instinct — start showing it to people and see what happens — burns months without producing a decision. This article lays out what actually comes next, in what order, and roughly what each piece costs.
First, be honest about which prototype you have
"It works" covers a wide range. A benchtop rig with a laptop attached and a bundle of jumper wires proves the physics. A hand-assembled unit that looks like the product and survives a week in a stranger's house proves something entirely different. Before planning anything, place your build on the ladder — the distinction between a looks-like model and a works-like model is the one that determines how much work remains.
Ask three questions. Does it work when you are not in the room? Does it work for someone who has never seen it before? Does it work the tenth time, after being dropped in a bag? If any answer is no, you are still in development, and the honest next step is another iteration rather than a manufacturing quote. That is not failure; it is cheaper than finding out during a production run.
The five tracks that run in parallel
After a functional prototype, work stops being sequential. Five tracks advance at the same time and constrain each other.
1. Design for manufacture
A prototype is optimized for being built once, by someone who understands it. A product is optimized for being built ten thousand times by people who have never met you. That means part count reduction, tolerance analysis, material selection with real suppliers behind it, assembly sequence, and test points. This is the design for manufacturing pass, and it typically changes more of the design than founders expect — sometimes the whole architecture, if the prototype leaned on machined parts that will never be affordable in volume.
2. Cost
Take your bill of materials and price it at the volume you can realistically afford for the first run, not at the volume in your business plan. Add assembly, test, packaging, freight, duty, and scrap. That number is your cost of goods, and it has to survive being multiplied on the way to shelf price — the arithmetic in markups, margins, and distribution math is where a lot of otherwise good products die. Do this before tooling, not after.
3. Compliance and certification
Find out now which marks your product needs and what testing they require. A radio means FCC. Anything that plugs into a wall usually means a safety listing. Children's products, food contact, and anything health-adjacent each bring their own regime. Certification is rarely the biggest line item, but it is frequently the longest lead time, and it can force a board respin. Start from which certifications a new electronic product actually needs and get a written answer, not a guess.
4. Intellectual property
A working prototype is often the point where a patent application becomes worth filing, because you finally know what the invention really is rather than what you hoped it would be. It is also the point where you start showing the thing to people, which starts clocks. Read when to file a patent before the first trade show, not after.
5. Channel
How will this reach a buyer? Direct online, retail, distributor, licensing to an established manufacturer? The answer changes the packaging, the price, the margin structure, and sometimes the product itself. It also determines whether you need to build a company at all — licensing versus manufacturing your invention is a fork worth taking seriously while you can still change direction cheaply.
What each stage typically costs
Numbers vary enormously by product complexity, but for a modest consumer device with some electronics, planning ranges look roughly like this:
| Stage | Typical range | What you get |
|---|---|---|
| Design for manufacture pass | $8,000–$40,000 | Production-intent CAD, BOM, drawings, supplier-ready package |
| Engineering validation builds | $5,000–$30,000 | A handful of units built the way production will build them |
| Certification testing | $3,000–$25,000 | Test reports and the right to put marks on the label |
| Injection tooling (if molded) | $4,000–$60,000 per tool | The mold that makes per-part cost collapse |
| First production run | Varies | Sellable inventory and the first real quality data |
Two things surprise people in this table. The first is that engineering after the prototype often costs more than the prototype did. The second is that tooling is a one-time payment that buys you a lower per-unit price forever, which is why the run size question in how many units your first production run should be is a financial decision more than a manufacturing one.
The validation you should buy before the tooling
Between the prototype and the mass production order there is a step worth every dollar: a small batch built with production methods and put in front of real users. Urethane casting, 3D printed housings finished to look production-grade, or a short CNC run can produce 20 to 100 units without committing to steel. Those units go to beta testers, to a trade show, to the distributor you are courting, and to a drop-test bench.
What comes back is the list of things nobody predicted — the button people press by accident, the instructions nobody reads, the part that loosens after a week. Fixing those in CAD costs a day. Fixing them after tooling costs a tool modification and six weeks. Scaling from prototype to production is mostly the discipline of pushing discoveries earlier.
Choosing a manufacturer without getting burned
Only now — with a real design package, a cost target, and validated units — are you ready to talk to factories seriously. Approaching a manufacturer with a prototype and a hope produces vague quotes that mean nothing and often a polite decline. Approaching with drawings, a BOM, a tolerance spec, and a quantity produces comparable numbers you can actually evaluate. Finding a manufacturer for your product covers the search. Read every quote carefully: three quotes for the same drawing package routinely turn out to describe three different products, with different finishes, different test coverage, and different assumptions about who pays for the fixtures.
A realistic sequence
- Classify your prototype honestly and iterate if it is not ready.
- Lock a requirements document, so the design stops moving.
- Run the design for manufacture pass and cost the BOM.
- Confirm the regulatory path and file IP before public disclosure.
- Build a small production-intent batch and put it in real hands.
- Fix what that reveals, then quote, tool, and run.
Expect this to take longer than the prototype did. Six to eighteen months from working prototype to sellable inventory is normal for a physical product, and the projects that go faster are usually the ones that did steps three and four early instead of last.
Projects House takes products through exactly this stretch — from a prototype that works to a manufacturing package a factory can build from. If you have a unit on your bench and are not sure what the next check should pay for, describe it through the contact form.