Production methods, chosen.
Also called: Manufacturing strategy · Process selection · Make-vs-buy · Supply-chain orchestration
Deciding how the product is actually made at volume: which processes, who does each part, what you make yourself, and what you buy in.
A production method is a system, not a step. You pick the processes, then decide who runs each one. You don’t need to own a factory; you need to orchestrate trusted partners so parts and labour flow into finished, tested units. Asset-light beats asset-heavy for a first run.
What production methods are
A production method is the answer to one question asked at volume: how does a pile of raw material and a stack of bought-in parts turn into a finished, tested unit a customer can buy? Not the prototype on your bench, where you can fudge anything with patience. The thing that has to repeat 500 times without you in the room.
It has two halves, and people usually only think about the first. The first half is process choice: for each part of the product, which manufacturing process makes it. A ceramic shell is cast and fired. A control board is populated and reflow-soldered. A wooden collar is cut and finished. Each part carries its own process, its own tooling, its own quality risk.
The second half is the one that decides whether you sink or swim: make versus buy. For every process, you choose whether to do it yourself or hand it to someone who already does it well. This is where the system thinking starts. A product is not made in one place by one set of hands. It is made by a chain of partners, each doing the part they are best at, with the parts flowing between them in sequence. Your job is to design that chain and keep it flowing, not to perform every operation in it.
Here is the opinionated bit, and I will happily be wrong on a future project: for a first run of a physical product, owning the means of production is almost always the wrong move. Tooling, floor space, machines and trained operators are enormous fixed costs you take on before you have proven a single sale. The asset-light route is to find the makers who already own that capability, route your work through them, and pay only for the units you actually need. You stay light, you stay flexible, and you can change a partner far more cheaply than you can sell a kiln.
First run of 500 to 1,000 units, a £149 retail price on a bill of materials of £38 to 55, and not one square metre of owned factory. The product is assembled from a flow of parts, not stamped out of a machine the founder paid for.
- Buy a factory you don’t need.
- Take on tooling, kilns and floor space before a single sale is proven.
- Lock yourself into fixed costs you can’t unwind cheaply.
- Become the bottleneck for every operation in the chain.
- Orchestrate trusted partners; stay asset-light.
- Buy capability as a service from makers who already own it.
- Pay only for the units you actually need, when you need them.
- Own the design, the spec and the relationships; let partners own the kit.
How it fits the bigger picture
Production Methods is activity 09.10.03 in Stage 09 Manufacture. The processes you choose here decide what has to be checked, so the next activity is quality control docs (09.10.04), which writes down how each partner proves their part of the chain is right before it flows to the next.
What it can do
It turns a working prototype into a repeatable system of partners and processes that delivers finished, tested units at volume without you owning a factory. It keeps the fixed costs off your balance sheet and the flexibility in your hands, so you can change a supplier far faster than you could change a building.
What it can’t do
It can’t run itself. An asset-light chain only works if the parts flow in sequence and each handover is clean, and that depends on the test specs and acceptance criteria the next activity writes down. Choose the partners well and the method still fails if nobody defines what “good” looks like at each step.
See the full 10-stage process →
Try it yourself
List every part of your product and, beside each, name the process that makes it. Now ask one hard question per part: do I make this, or buy it from someone who already does it brilliantly? Default to buy. For each “buy”, name a candidate partner and the single thing you’d hold them to. What’s left in the “make” column is your real factory footprint. If that column is empty, good. You’ve designed an asset-light chain.
Want a structured first pass at how your product gets made? Start the Free Sprint → and the GPT will help you map the chain.
Your production-method checklist
Project notes: a factory made of partners
▸ From the notebook · optional reading
How the proofing box’s first run of 500 got made across three towns, with nobody owning a single machine.
3 min read · click to open
Dan came to the table assuming we’d need a unit, some machines and someone to run them. It is the natural picture in your head when you say the word “manufacture”. I asked him a different question: “For a first run of five hundred, what do we actually need to own?” The honest answer turned out to be: nothing with a plug on it.
The chain we drew
We worked the product back into its parts and gave each one to whoever already did it best. The ceramic shell went to an established maker in Stoke-on-Trent who has been firing this kind of body for decades; their kilns, their moulds, their problem to solve, not ours. The low-voltage control board went to a small contract assembler in Manchester who builds to a drawing and tests every board. The wooden collar came from a joinery shop. And the final marriage of all of it, plus the BS EN 61010 end-of-line test, sat with a small finishing partner who packed and shipped.
So the “factory” was four partners across three towns, with parts flowing between them in sequence: ceramic to the finisher, boards to the finisher, finisher tests and ships. Nobody in that chain owned the whole thing. The Hartleys owned the drawings, the bill of materials, the UKCA file and the relationships. That was the entire owned footprint.
What that bought us
The numbers told the story plainly. A first run of 500 to 1,000 units at a £149 retail on a £38 to 55 bill of materials was viable precisely because there was no kiln to amortise and no floor space sitting idle between runs. When the second batch needed a tweak to the heater spec, we changed the instruction to the Manchester partner in an afternoon. Had we bought our own assembly line, that same change would have meant retraining and downtime we’d already paid for.
The one thing this model demands in return is discipline at the handovers. Because the parts move between independent businesses, every interface has to be specified, which is exactly why the next activity, the quality control docs, exists. The chain only flows as cleanly as the specs that govern each junction in it.
— Manufacture stage, project notes, 2026
— Next in Manufacture → Quality control docs