Here is an uncomfortable observation – the easiest part of any recall is announcing it. Ford is currently recalling vehicles faster than its dealer network can repair them, and the trade press has started publishing the gap between recall velocity and repair throughput the way it once published defects per unit. Elsewhere, a maker has had registrations suspended over an axle defect before any remedy existed at all. Every quality director should recognise the mechanism, because this is not a Ford problem. A field action is a production system: takt, parts, bays, prioritisation rules. Most organisations design it three weeks after the press release, by which point the backlog is already compounding – in grounded stop-sale stock and in customer patience. I have run this arithmetic at smaller scale, in plants and across supplier networks. It is unforgiving at any volume.

The maths nobody runs before the press release

Take a campaign covering 400,000 VINs at 1.5 technician-hours per unit, diagnostics and paperwork included. Assume the certified dealer network absorbs 30,000 units a week – optimistic, since those same service bays earn their living on routine maintenance and ordinary warranty work. Thirteen-plus weeks to zero at full dedication. Closer to half a year at a realistic fifty per cent. The weeks-to-zero number never appears in the announcement, yet it is the only number the customer ever experiences.

Then there is the stop-sale inventory. 12,000 grounded new vehicles at €35,000 apiece is €420 million of frozen stock. Floorplan financing at seven per cent burns roughly €565,000 a week in interest alone – before depreciation, and before the buyer who simply walks across the street. At SNOP we cut defect costs by seventy per cent, and none of it came from heroics. It came from putting the weekly cost of failure on a single A3 where nobody in the leadership review could un-see it. A recall backlog deserves the same page, updated every Friday, weeks-to-zero curve drawn in ink.

The parts fight the line

Remedy kits draw on the same supply base that feeds the line building new cars. When both need the same harness, valve or control module, someone must set the allocation split – and that split is a containment decision, not a logistics detail. It needs a named owner and a daily review. The line has a schedule, takt penalties and a planner shouting at it every morning; the field has a parts distribution list. Left unowned, the allocation defaults to whoever is louder, and louder is always the line.

The second problem is worse. A field fix is a running change executed by dealer technicians – thousands of them, trained to different levels, working to a service instruction instead of a control plan. In the plant, a running change goes through trial builds, PPAP, torque verification and a reaction plan. In the field it gets a PDF. No poka-yoke at the dealer bay. No andon when a fastener crosses threads. No reaction plan beyond the comeback. At WITTE Automotive we ran unplanned process changes through QRQC discipline – containment owner, dated actions, a 24-hour loop – and tracked the failure-cost curve on a Q-Wall until it bent. The field version demands that rigour with a far worse span of control. I have run containment across a multi-site operation of more than 2,000 people, and the engineering was rarely the hard part.

A recall is a promise priced in someone else's patience – and patience, unlike parts, cannot be expedited.

Manage the backlog like a line, not a lawsuit

The legal instinct during a field action is to go quiet, minimise statements, wait. The production instinct is the opposite: publish the curve and burn it down. The backlog is work-in-process, so manage it like WIP – a daily stand-up on the constraining VIN families, an A3 for the capacity plan, a burn-down rate reviewed with the same seriousness as the daily build. It is the daily build now.

Triage at VIN level. Rental and fleet units piling on mileage fastest go first, then geographic clusters where heat or corrosion accelerates the failure mode, then the rest by age and duty cycle. Rationing by queue order alone is an abdication disguised as fairness.

And give customers dated commitments. Not a second apology – a date. I once went a full quarter with zero critical customer escalations, and not because problems stopped arriving. Every affected customer held a dated commitment that we kept, reviewed daily against the burn-down. An apology absorbs one failure; a kept date absorbs a campaign. Treat the comeback rate – the vehicle repaired twice – as seriously as an escape from your own plant: open an 8D, capture the dealer technician's findings, feed them back into the service instruction, close the loop the way you would an andon pull.

Key takeaways

  • Run the takt arithmetic before the press release: VIN count divided by weekly certified capacity gives weeks-to-zero – the number customers actually live with.
  • Name an owner for the parts split between the production line and the field, and review it daily as the containment decision it is.
  • Treat the field fix as an uncontrolled running change: give dealer technicians a control plan, poka-yoke where the fix allows, and an 8D loop on every comeback.
  • Manage the backlog as production WIP – daily burn-down, VIN-level risk triage, dated commitments to customers that are kept, not repeated.

If your PFMEA carries a severity-10 failure mode, the remedy capacity plan belongs in the same file, beside the severity rating – not in the crisis folder where it surfaces after the announcement. Regulators close files; customers open calendars. A recall ends with the last repaired VIN, not the acceptance letter, and the warranty clock runs the whole way down.