Two-thirds of the leadership is gone. The failure mode has tenure.

I have stood in a yard of contained parts that could not ship – never tens of thousands of vehicles, but enough to know the smell of the decision: cooling metal, dust, and a finance controller recalculating working capital by the hour. So when I see Ford parking Super Duty trucks by the tens of thousands while two-thirds of its manufacturing and engineering leadership is replaced, I recognise the reflex. It is the reflex of an organisation that still diagnoses its escapes as a people problem.

582 recalls in ten years. More than any other automaker has run up. That is roughly 58 a year – a recall a week, every week, for a decade. At that rate a recall is not an event. It is a routine output of a stable process, and a stable process always tells the truth about the system that built it. The uncomfortable observation is not that Ford has a quality problem. It is that the visible response – trucks in a yard, executives out the door – treats the staff and the symptom while the structure producing a recall a week survives every reorg untouched.

Replacing leadership is containment, not correction

Run the response through 8D discipline and watch it fail the taxonomy. Parking trucks is containment – D3, correctly executed, expensive and necessary. Replacing two-thirds of manufacturing and engineering leadership is also containment, of a psychological kind; it makes the board feel like correction. Neither changes a process, a control plan, a detection point or a single line of a PFMEA. The failure modes that generated 582 escapes will meet the new incumbents on Monday morning, undocumented and waiting in the same fixtures.

“A bad system will beat a good person every time.” — W. Edwards Deming

Deming said it long before anyone built a recall dashboard. Swap the whole cohort and the system still wins, because in any system tenure belongs to the failure mode, not to the people. The new leadership inherits the same process capability, the same detection rankings, the same supplier PPM numbers. Six months in, the first recalls under the new regime will be read as their personal failure. Watch that story get written.

Escapes are detection failures first

Every escape that becomes a recall shares one property: the failure mode existed, and nothing in the chain caught it. In PFMEA language the problem is occurrence and detection, not severity – severity was known when the part was designed. So the correction lives in control plans, reaction plans, layered audits and the authority of the person who can stop the line. None of those artefacts report to the executive suite. They report to the system that leadership turnover leaves exactly as it found it.

Blaming people is cheaper than auditing a system. A name closes the case in an afternoon; a process capability study takes weeks and indicts decisions signed off years ago by people who have since been promoted. So the org chart absorbs the punishment and the process quietly keeps its tenure. That is how a company runs up 582 recalls without once recalling its own control plans.

I have run this experiment. At SNOP, building a greenfield QA/QC department for a 900-plus-employee plant, we cut defect costs by 70% and closed a quarter with zero critical customer escalations – no purge required. At Witte Automotive before that, the toolkit was QRQC, A3 and a Q-Wall that made every escape visible within the shift it happened. QRQC worked because it fixed the loop instead of the blame: defect found, contained, root-caused and countermeasured fast, with the person who found it leading the discussion. I have led quality initiatives across a 2,000-plus workforce on multiple sites. I cannot name a single escape I ever prevented by changing a manager.

What correction would actually look like

Correction is boring, and boring is the problem. It means recalculated PFMEAs with honest detection rankings for every historical escape; reaction plans with named owners and dates; escape rate on the board agenda with the same weight as warranty cost; stop-ship authority that sits at the line instead of in a committee. Layered audits that sample the process rather than the paperwork. Read-across to every part family sharing the process, because a corrected PFMEA that stays on one part number is a hobby. None of it photographs well next to a press release about fresh leadership. All of it changes the Monday-morning meeting between the failure mode and its next incumbent.

Executives are cheaper to replace than control plans, which is why control plans so rarely get replaced.

Ford's yard will empty. The trucks will be reworked, shipped and booked as the cost of doing business. The question that decides the next 582 recalls is whether anyone reworks the system with the urgency applied to the org chart: the detection paths, the audit layers, the escalation logic. Two-thirds of the leadership is gone. The failure mode still has tenure. And tenure beats a new hire every time.