The compliance is licensed. The complexity is homegrown.

Here is an uncomfortable observation: when a mid-tier manufacturer buys a compliance platform, the vendor's sales cycle runs faster than the plant's PFMEA update cycle. Six weeks from demo to signature. The control-plan revision that platform was bought to serve has been open since spring, waiting on an engineer who spends half his week firefighting. That gap is not a software problem. It is an accounting problem. The complexity was not inherited and it is not weather – it was signed for, one accepted customer requirement, one standard, one variant at a time, each carrying a margin case and never a control-plan case. A running argument in the trade press says mid-tier manufacturers need a regulatory-technology strategy. Two decades of running quality in automotive and aerospace plants tell me the diagnosis is different. This industry does not need better vendor selection. It needs to price complexity at the point of acceptance.

Complexity arrives on a purchase order

No quality manager wakes up and orders four hundred extra active documents. They arrive as consequences. A new customer brings customer-specific requirements: a supplier portal, a labelling standard, a VDA 6.3 audit expectation, two critical characteristics nobody asked for. A new market brings a regulatory layer. A new variant brings, every single time, a control-plan entry, a reaction plan, a gauge strategy, a training record, a requalification loop, a revision to every document that touches the part. Each acceptance is individually rational. The sum is a quality system running three shifts of paperwork nobody costed.

The mechanism nobody prices is simple. The commercial team evaluates margin per part, and the part looks profitable. The quality system absorbs the rest as overhead, spread invisibly across engineering hours, audit preparation and firefighting. In consulting work – ArcelorMittal, Meopta and AVX Kyocera among the clients – I watched the same film: a plant drowning in document revisions reaches for a tool licence while the acceptance gate leaks like a sieve. One low-runner variant worth €40,000 a year quietly consumed €60,000 in control-plan maintenance, calibration and requalification. It appeared in no report. It appeared in overtime and escapes.

Regtech is an amplifier, not a filter

Strip away the demos and these platforms do a specific set of things: document control, training matrices, CAPA and change workflows, traceability, audit trails. They accelerate the documentation of quality. If the QMS is disciplined – clean FMEA structures, living control plans, 8Ds that close in weeks – that speed compounds and the subscription pays for itself. If it is not, the platform accelerates everything it touches, including the failures. You do not get fewer nonconformities. You get the same nonconformities, documented faster, at scale, on a per-seat licence.

Traceability is not control. A beautifully documented fire is still a fire.

The pricing model punishes the undisciplined. Per-seat costs mean the plant that accepted the most complexity now pays the most to administer it, monthly, indefinitely. And the current wave repeats the pattern one layer up: governance platforms arriving to govern the AI that was supposed to fix the last platform's problems. We now license software to govern the software we licensed to govern the plant. Nobody bills the complexity itself.

Failure cost is where complexity finally gets priced honestly, because the customer invoices it. At WITTE Automotive we attacked failure cost with QRQC, A3 and a Q-Wall and brought it down substantially – not because the tools were clever, but because the wall made the seams visible. Every escape pointed somewhere: a variant the control plan had never truly covered, a characteristic inherited from a requirement accepted years earlier by someone who never saw its cost.

Price the complexity before you accept it

The rethink happens at the gate, not at the licence renewal. Four moves, none of them expensive:

• A complexity gate in the quality plan. Every new customer requirement or variant gets a costed quality-impact assessment before acceptance – requalification hours, added control-plan lines, gauges, training, document revisions – signed by the same person who signs for the margin.
• One-in, one-out for low-runners. Plot revenue per variant against quality-system load. When the tail is 40% of your variants and 4% of your revenue, retire a variant each time you accept one.
• Change control as a costed decision. A change request carrying no price tag carries no information. If changes are free, you will get plenty of them.
• QRQC when escapes expose the seams. An escape means the control plan lied somewhere; the fix belongs in the acceptance logic, not only in the part.

I learned this the expensive way, then the cheap way. Building the QA/QC department for SNOP's greenfield plant – 900+ people, initially a fence and a drawing – I designed the document architecture before the first variant arrived, and deliberately simple: one FMEA structure per part family, modular control plans, a document count I could defend from memory in an audit. Complexity had to pass a gate before it earned a document number. The system absorbed variants instead of drowning in them, and the plant ran to a 70% defect-cost reduction and 98% customer satisfaction on modest software. Discipline first. Licence last.

So the strategic question for a plant manager is not which platform to license. It is who owns the cost-of-complexity number, and what authority comes with it. Give it to quality without authority and you get veto theatre; leave it with sales and purchase orders keep outvoting control plans. The plants that win the mid-tier are not the ones with the better licences. They are the ones that refuse the complexity they cannot control – then buy software only for the traceability they chose to need.