Ford fired managers this week for what it called a "good enough" mentality. Good instinct, wrong tool. In two decades of running and repairing quality in automotive and aerospace plants, I have never walked a floor where "good enough" was an individual defect. It was always an output – the reliable product of a KPI sheet that pays for shipping and budgets nothing for stopping. Fire the manager and you have swapped a component in a machine that builds the replacement at the same rate it builds door panels.

'Good enough' is manufactured, not hired

Put yourself in a shift manager's chair at 16:45 on a Friday. Stopping the line costs €9,000 an hour, lands on Monday's operations report and arrives with their name on it. Letting a marginal part flow costs – what, exactly? Nothing that anyone reads. The scrap is a logistics cost, the sort-at-customer sits in someone else's budget, the warranty claim lands eighteen months later in a ledger nobody reconciles back to the decision that created it. Output, OTIF and hours-per-unit get reviewed weekly. Cost of poor quality gets totalled never, or quarterly in a form so aggregated it cannot embarrass anyone. Under that arithmetic, tolerance is not a character flaw. It is rational behaviour, and the manager exhibiting it is the one reading the dashboard most accurately.

The deviation workflow finishes the job. A temporary deviation, valid thirty days, gets extended once, then twice, then someone stops writing the extensions down. The containment becomes the process. I have audited plants where the most stable element of a control plan was a temporary 100% sort that had outlived three process engineers.

A plant's real quality standard is the worst defect it has agreed to ship twice.

Leading quality initiatives across a 2,000-plus workforce on multiple sites, I watched one system hire "good enough" managers, promote them for output, then fire them for tolerance – three verdicts, one system. The organisation itself never went on trial. That is the part of the Ford story nobody discusses: dismissals are popular because they are legible. Everyone understands a firing. Almost nobody wants to read a deviation log.

What firing actually buys

The signal is real. When a board treats quality culture as a personnel matter, the tolerance ceiling moves, and people notice ceilings. A dismissal communicates faster than any policy. It also teaches people to hide. I have spent years de-escalating customers, and every zero-escalation quarter I have run was built on problems surfacing early – loudly, cheaply, before they boarded a truck. Fear reverses that. The defects continue, the escalations go quiet, the dashboard improves for two quarters. Then the customer calls. I have taken that call.

The successor inherits the escalation matrix untouched. The same ship-anyway path that needs three signatures at 16:45 on a Friday and one signature at 09:00 on Monday. The same quiet arithmetic. Nobody plans to become tolerant; it happens one signature at a time, each individually defensible. By week twelve the new manager has signed their first deviation under delivery pressure. By week twenty-four they cannot tell you which of their standards they have already traded away. The system relearns them faster than they can reform it.

Designing the tolerance out

Price the tolerance in €

At WITTE Automotive we attacked this with QRQC, A3 and a Q-Wall. QRQC forced failure into the open within twenty-four hours, in the hall, with the people who made the part. A3 gave each failure an owner, a mechanism and a countermeasure aimed at cause rather than symptom. The Q-Wall did the political work: every deviation, every sort, every containment hour priced in € and visible daily. Once tolerance carried a price tag on a wall senior management walked past, "good enough" had nowhere to hide. The failure-cost reduction that followed was substantial. None of it came from hiring braver people. It came from changing what the arithmetic rewarded.

Give 'no' structural backing

Then, at SNOP, I built the alternative from scratch. Standing up the QA/QC department of a greenfield plant – 900-plus employees, empty halls, no legacy habits – meant writing the rules before the habits existed. Stop authority went in on day one: a quality engineer could halt a shipment, the escalation travelled upward, and whoever stopped the line paid nothing for it in career terms. That is the entire trick: when "no" is structurally protected, you rarely need to fire anyone for saying "yes" too easily. Within a quarter we ran zero critical customer escalations; defect cost fell by 70% and customer satisfaction settled at 98%. People did not become heroes. The system stopped billing them for heroism.

Key takeaways

  • Audit what your KPIs pay for. If shipping is rewarded weekly, stopping is costed hourly and cost of poor quality is totalled never, tolerance is the rational output. Change the dashboard before the org chart.
  • Age your deviation log. Any "temporary" containment older than ninety days is a de facto process change that skipped approval – treat it as one.
  • Price tolerance in € on a wall leadership physically walks past. "Good enough" dies of arithmetic, not attitude.
  • Write stop authority into job descriptions with an upward escalation path, so halting a shipment is a duty, not an act of courage.

Fire for cause, yes. Some people choose the easy signature every time, and when they do, act. But first requalify the system that trained the person you just fired: audit the KPIs for what they pay, age the deviation log, total the containment spend in €. Otherwise you are swapping components in a process you never validated – which is precisely what a decent auditor would write you up for. Firing the "good enough" manager is quality management by organ removal: the surgery is decisive, the crowd applauds, and the disease is still on the payroll two doors down.