The word custom hides a spread of four different projects with four different price tags, timelines and failure modes. A buyer who says "we want a custom product" may mean a logo on a stock housing, or a new motherboard, or a device that has never existed — and the factory hearing the word quotes against the rung they assume, which is how both sides end up disappointed by a process that was never defined. Naming the rung before the first quotation is the cheapest project management that exists.
Rung one: branding on a stock platform
The lowest rung changes nothing inside the product. Logo on the housing, custom carton, translated manual, a chosen colour from the existing palette. The platform, the PCB, the firmware and the app stay as shipped to every customer. This rung has real value — it is what makes a product sellable as a brand — but it does not differentiate the hardware, and a competitor can reach the same rung by paying the same setup fees. Timeline: weeks. Tooling: print plates and packaging. Risk: almost none.
Rung two: configured variants
The second rung changes the configuration of an existing platform: a different lens, a different radio module, a housing colour from new tooling, a stripped or added feature on the firmware build, a different plug standard. The platform carries the change; the bill of materials is edited, not designed. This is where most successful brand programmes actually live, because it buys real market fit — a product shaped to the destination market — at a fraction of full development. Timeline: two to four months where a new housing colour is involved. Tooling: one or two moulds. Risk: low, but the change list must be tested as a whole, because a swapped module changes power consumption, which changes the battery claim, which changes the packaging copy.
Rung three: platform extension
The third rung adds capability the platform never had: a new sensor type on an existing hub, a new product category that joins an existing app ecosystem, a private protocol bridge. The board may be redesigned around existing firmware architecture. This rung is where the app becomes part of the specification — a new device type needs screens, pairing flows and server-side support, and app work is invisible in factory quotations until someone asks. Timeline: four to nine months. Risk: medium, and concentrated in integration rather than hardware.
Rung four: ground-up development
The fourth rung is a new product: new industrial design, new board, new firmware, certification from scratch. It is also the rung where the buyer's assumptions cost the most, because a ground-up project starts with a development phase whose output is still a prototype — and the quotation buyers expect is only available after that phase defines what is being quoted. Factories that take this rung seriously quote it in stages: a development fee, a prototype milestone, then tooling and mass production pricing that depends on the prototype's bill of materials. A single all-in number for rung four is a warning sign in either direction.
Standing on the right rung
The useful exercise for a buyer is to write down which rung the project actually needs — not which one sounds ambitious. Most market problems are solved at rung two: the destination market needed a different plug, a local cloud, a different chime standard, a colour that matches the brand. Paying rung-four money for a rung-two problem is common and unnecessary. Underestimating a rung-four requirement and forcing it into rung two is rarer and more expensive, because it ships a product that solves the wrong problem well.
XingLiao runs custom smart home development across all four rungs — branding, configured variants, platform extension and ground-up programmes — with the rung named and priced before the first sample. The same platforms behind the manufacturer range support the higher rungs.
The short version
Branding, configuration, extension, ground-up: four rungs, four price structures, four timelines. Name the rung first and the quotation, the timeline and the risk all become comparable. The most expensive mistake in custom hardware is not overpaying — it is buying the wrong rung and discovering it at the certification stage.
Top comments (0)