"Too basic," said the reviews of the $25,000 Slate truck. Here is an uncomfortable observation from two decades of defect Paretos: basic is not the insult. Basic is the prize nobody budgets for. Every reviewer who climbed into that cabin and lamented the missing touchscreens and soft-touch plastics was performing a consumer ritual, not a quality assessment. Meanwhile, somewhere in an engineering office, someone made a deletion list — and if that list came from field data rather than a purchasing spreadsheet, it is the cheapest quality program the EV industry will run this decade.
Complexity is the failure mode no one funds removing
Do the arithmetic of subtraction honestly and it stops being philosophy. Every feature you add to a vehicle carries its own PFMEA lines, its own test station, its own torque spec, its own PPAP scope, its own training module, its own traceability burden. Every variant multiplies all of it again. And complexity doesn't fail loudly. It fails as a mis-picked latch at station 40, a wrong-configuration lock shipped to the wrong line, a gauge nobody calibrated because that part number hasn't run in six weeks.
I lived this at WITTE Automotive. Latch-and-lock portfolios breed part numbers the way damp garages breed mould — thousands of them, differing by a microswitch or a cable orientation a tired operator cannot distinguish at takt speed. Our QRQC data told the same story quarter after quarter: variant confusion and rarely-run configurations sat permanently in the top three defect families. Not design flaws. Not supplier scrap. Sheer variant mass confusing otherwise capable processes. When we later ran the A3 work that cut failure costs substantially, I'd estimate half the improvement came from deletions — killing low-runner variants, consolidating part numbers, retiring tests that existed only because a variant existed. We didn't add a single control.
A feature you never launch is a defect you never fund.
The industry funds the opposite behaviour. Launch budgets pay for additions. Nobody gets promoted for deleting a test station. But every OEM planning de-contented EV trims for affordability is going to run this experiment whether they intend to or not — the only question is whether they do it with discipline or by spreadsheet.
The two kinds of basic
This is where the Slate debate gets interesting, and where a quality engineer refuses to accept "basic" as a single category. There are two kinds, and they look identical on a window sticker.
Corner-cutting deletes the visible cues customers use to judge build quality — the consistent panel gaps, the door thunk, the seals that don't whistle at 130 km/h on the autobahn. It keeps whatever is cheap underneath and calls the result a trim strategy. Designed subtraction deletes variation generators nobody actually wants — the motorised vent, the fourth trim level, the software-dependent mirror adjust — and keeps the fundamentals over-engineered.
The evidence that separates them is not the bill of materials. It's the warranty data and the field Pareto. If a manufacturer deleted features whose failure modes were fat entries on the warranty Pareto, that's designed subtraction. If they deleted features the field data never flagged and quietly kept the ones it did, that's corner-cutting with better press handling. You will be able to read the answer in the claims data within eighteen months of launch.
What a stripped truck does to the control plan
Fewer torque specs. Fewer inspection stations. Shorter reaction paths when QRQC fires, because the possible-cause tree has fewer branches. A stripped vehicle doesn't just cost less to build — it fails in more legible ways, and legible failure is the entire game on a shop floor.
When I stood up the QA/QC department for a 900-plus-employee greenfield plant at SNOP, the most consequential decisions were not what to inspect. They were what not to inspect. Every station you leave out saves equipment, headcount, maintenance and — the part nobody prices — false-alarm noise. A marginal test that rejects good parts teaches operators to distrust the system, and an operator who distrusts the system overrides it. That's how detection theatre becomes a defect generator. Part of the 70% defect-cost reduction we achieved there came precisely from removing tests that added theatre rather than detection.
The same logic applies to the AI inspection wave the industry is currently funding — cameras that never blink, vision systems racing into every plant. Fine technology. I build and deploy AI systems myself, so this is not a purist's objection. But bolting intelligent detection onto a product you made unintelligently is paying interest on a complexity debt you chose to take on. The cheapest detection system in any plant is the test you never have to staff because the failure mode never existed.
Key takeaways
- Every added feature multiplies PFMEA lines, gauges, PPAP scope and training — treat each deletion as a funded quality project, not a cost-down line.
- Corner-cutting versus designed subtraction is provable: check the deletion list against the field Pareto and warranty claims, not the BOM.
- Variant confusion in rarely-run configurations is a permanent top-three defect family — consolidating part numbers is often cheaper PPM reduction than any new control.
- Before adding a test station, ask whether it detects or performs. False-alarm noise erodes operator trust faster than any defect erodes margin.
So the first page of your next quality program should not be a control plan. It should be a deletion list — a ranked, evidence-backed inventory of the parts, variants and tests you are going to stop making and stop staffing. The cheapest PPM reduction in any plant is the part you never build. Knowing which part that is — that's the whole job. The truck is basic. The question every OEM copying it should be asking is whether their discipline is.