A company that already manufactures something has advantages a startup would trade almost anything for: a factory, a trained workforce, tooling that is paid off, a distribution channel, customers who answer the phone, and a finance department that understands unit economics. It also has problems a startup never faces, and they are why so many established manufacturers launch new products more slowly than a three-person team in a rented shop.
The difference is not talent. A startup optimizes for one thing — getting a product to exist — while a manufacturer has to do that without disrupting the products already paying the bills. Everything in the internal process follows from that constraint.
Where the idea comes from
Startup ideas usually begin as a founder's conviction. Manufacturer ideas come from four places, each with a different risk profile:
- Customers asking. The strongest signal available. A distributor or key account saying "we would buy this" is worth more than any market study — but customers ask for what they know, so this channel produces line extensions more often than new categories.
- Sales losing deals. A pattern of lost bids on a specification you cannot meet is a documented, quantified market gap. This is often the best-evidenced idea in the building and the one that gets least attention because it arrives as a complaint.
- Production and service knowledge. The people who repair the products know which failure modes annoy customers most, and the people running the line know what your equipment could make that it currently does not.
- Competitive and category pressure. Reacting to a competitor's launch is the weakest starting point, because you begin at least a year behind and you inherit their positioning.
The organizational problem is that these ideas arrive continuously, from people whose main job is something else, and there is nowhere for them to land. The first fix in most manufacturers is not a development process at all — it is a simple intake: a place ideas get written down, someone who screens them, and a monthly meeting where they are killed or advanced. Without it, the projects that happen are whichever ones had an executive sponsor.
Stage gates, and why they exist here
Startups iterate because a wrong decision costs a prototype. A manufacturer's wrong decision costs a $70,000 mold, a production slot, and a sales channel's credibility. So money is released in tranches, each requiring a decision meeting with the authority to stop the project.
| Gate | Question answered | Typical spend to reach it |
|---|---|---|
| Concept screen | Does this fit our strategy, channel, and capabilities? | A few days of internal time |
| Business case | What volume, price, margin, and payback — and who says so? | $5,000–$25,000 |
| Feasibility and concept design | Can we actually build it, and roughly at what cost? | $20,000–$80,000 |
| Detailed design and prototype | Does a real unit meet the spec in customers' hands? | $50,000–$250,000 |
| Tooling and pilot run | Can the line build it repeatably at the target cost? | Tooling plus a pilot build |
| Launch | Are sales, service, and supply ready? | Launch and inventory costs |
Two things make gates work rather than turn into theater. First, a gate must be able to say no; a process where every project passes every gate is a reporting exercise. Second, each gate needs named deliverables agreed in advance, so the meeting is a review rather than a debate about what should have been prepared. The mechanics are covered in stage-gate development with real decision points, and the underlying sequence of technical work is the same one described in the product development process step by step.
Who owns the specification
This is the question that decides more manufacturer projects than any technical choice. In practice there are three models, and the failure mode of each is predictable.
- Sales owns it. The spec reflects what customers ask for, which is good, and grows without limit, which is not — every request is individually justified by a deal, and the product ends up costing too much to hit its price.
- Engineering owns it. The spec is coherent, buildable, and often solves a slightly different problem than the market has.
- A product owner owns it. One named person accountable for the business result, gathering input from sales, engineering, production, and service and then deciding. This is the model that works, and most manufacturers have to create the role deliberately because it does not exist in a company organized around making an existing product.
Whoever owns it, the output is a written requirements document with numbers, not a slide deck — target cost, target price, performance limits, environment, regulatory scope, and what is explicitly out of scope. Writing it well is a skill in itself, laid out in how to write a product requirements document. A manufacturer's version has one extra section: the constraints imposed by the plant.
Designing around what you already own
This is the manufacturer's real superpower, and it is routinely underused because designers default to a blank sheet.
Before the first concept sketch, list what the company already has: existing molds and whether any could be modified rather than replaced, standard extrusion profiles, sheet metal thicknesses already stocked, motors and power supplies already qualified and bought in volume, fasteners in the standard kit, packaging sizes the warehouse already handles, and the process capability of every machine on the floor. Then design toward that list.
The payoff is substantial. A part that fits an existing family mold with an insert change might cost $8,000 instead of $70,000. A product using the same power supply as three other models buys better and needs no new safety file. A design that runs on the existing line needs no new fixtures, work instructions, or retraining. This is value engineering applied at the concept stage rather than after launch, where it is worth several times more.
The trap on the other side is designing a compromised product to avoid buying one new tool. If the constraint list is producing something worse than customers can buy elsewhere, the constraint is wrong, not the market.
Cannibalization: the conversation nobody wants to have
A startup's new product competes with strangers. A manufacturer's new product competes partly with its own catalog, and this genuinely stops good projects.
The honest framing: the question is not "will this take sales from our existing line" but "will someone take those sales anyway." If a cheaper or better version of your product is technically feasible, a competitor will eventually build it. Cannibalizing yourself keeps the margin in-house; declining hands the transition to someone else. Manufacturers manage it by positioning the new product at a different price tier or use case, selling it through a different channel, timing launch to the old model's natural refresh, and stating in the business case how much forecast volume is genuinely new versus transferred. A forecast that quietly assumes zero cannibalization will be missed.
What actually differs from a startup, summarized
- Cost of being wrong. A startup pivots. A manufacturer has channel commitments, service obligations, and inventory, so it validates earlier and commits later.
- Speed limiter. Startups are limited by money. Manufacturers are limited by attention — the same engineers who would develop the new product are firefighting the current one. Ring-fencing development time, or buying capacity outside, is usually the binding constraint.
- Cost target. A startup discovers its cost. A manufacturer knows the price the channel will bear on day one and designs to a margin from the start.
- Definition of done. A startup ships when the product works. A manufacturer ships when service, spares, documentation, training, and the line are ready too.
- Reuse. A startup starts from nothing. A manufacturer starts from a parts library, a supply chain, and a factory — and the projects that exploit that fully are the ones that hit their numbers.
Most established manufacturers end up with a hybrid: internal engineering keeps running the current products, and the new product's development is run alongside with outside help for the disciplines the company does not employ full time — industrial design, electronics, firmware, regulatory. The tradeoffs of that split are worked through in outsourcing development versus building an in-house team, and a company weighing whether to develop its own product at all will recognize the questions in a small business that wants its own product. Companies that instead brand someone else's design are running a different model entirely, compared in OEM versus ODM.
Projects House works with manufacturers who have the factory and the customers but not the development bandwidth — running the design, the prototypes, and the production readiness against your existing tooling, suppliers, and cost targets. If you have a product idea and a plant already running, start a conversation through our contact form.