1.4 The development process: from requirement to production

A moulded part does not spring from a single designer. It travels through a long, structured sequence of steps — and usually across several different companies — before it reaches mass production. Two things are worth understanding separately: the development process every automotive part follows (this section — not to be confused with the physical moulding process of §1.2), and the stakeholders who carry out each part of it (the next section).

1.4.1 The industry standard: APQP

Automotive product development follows a formal, published standard called APQP — Advanced Product Quality Planning, defined by the industry body AIAG (the Automotive Industry Action Group). Its purpose is to make sure a new part is properly planned, designed, validated and released, with agreed deliverables at each step. The European automotive industry uses a closely aligned equivalent, VDA Maturity-Level Assurance.

APQP organises the work into five phases, ending in a formal part sign-off called PPAP (the Production Part Approval Process):

For an injection-moulded part, the five phases play out roughly like this:

1. Plan & Define. Turn the customer's needs into concrete part requirements — function, size, material, cost and volume targets — and settle the industrial design (the styled shape).
2. Product Design & Development. Engineer the part itself: the detailed CAD geometry, the choice of material, engineering review, and the design-for-manufacturability checks (the design rules of §1.3). This is where the part's mouldability is largely decided.
3. Process Design & Development. Engineer how it will be made: design the mould (the tool), plan the moulding process, and run mould-flow analysis to confirm the tool will fill and cool well.
4. Product & Process Validation. Cut the tool, mould prototype parts, test them against the requirements, correct any problems, and obtain PPAP sign-off that the part and the process are approved for production.
5. Launch, Feedback & Corrective Action. Ramp up to mass production, then monitor quality and feed problems back into corrective action.

1.4.2 Why no single actor can do it all

Notice how much ground these five phases cover: market analysis and styling, detailed geometry and material choice, tool design and mould-flow simulation, press set-up, testing and formal sign-off. These are genuinely different competencies, and no single person — and rarely a single company — holds all of them. The workflow is therefore split across specialists, and usually across several companies, each owning a slice of it. (APQP describes what the phases are and what each must deliver; it deliberately says nothing about who performs them — that "who" is the subject of the next section.)

That division is also what makes moulded-parts engineering hard, for two reasons:

• Every hand-off is a data boundary. The designer's CAD model, the toolmaker's mould design and the moulder's process records live in different tools — and often in different companies. Knowledge that would be most valuable joined together (which geometries caused which defects, which mould changes fixed them) stays fragmented across those boundaries and rarely flows back upstream to the person who chose the geometry.
• Feedback is slow and travels the wrong way. What actually happened on the press — the scrap, the tuning, the warpage measured on the finished part — sits at the far downstream end of the chain, while the decisions that caused it were made far upstream. Closing that loop, so that production reality informs the next design, is one of the enduring structural challenges of the field.

A useful mental model: the workflow is a supply chain of hand-offs, and each hand-off is both a coordination point and a place where information is lost. Much of good moulded-parts engineering is the discipline of not losing that knowledge — carrying it upstream into the geometry, which is what the design rules of §1.3 encode.