Case file
- What happened: In February 2018, KFC UK switched its logistics contract from Bidvest (a food-specialist operating multiple distribution depots) to DHL (operating a single new depot). Within days, deliveries collapsed. Stores ran out of chicken.
- Scale: Roughly 600 of about 900 UK restaurants closed — two thirds of the estate — for periods ranging from hours to days.
- Root cause: An unvalidated process change. The consolidation from a multi-depot food-grade network to a single, unproven depot created a single point of failure with no demonstrated operational readiness at full volume.
- The bill: Estimated tens of millions of pounds in lost sales, brand damage, and crisis-recovery costs. The exact figure was never formally disclosed, but press estimates converged on £20–30 million in lost revenue alone.
The situation
KFC UK operated roughly 900 restaurants supplied through a mature, multi-depot distribution network run by Bidvest, a company whose core business was food logistics. Multiple depots meant geographic redundancy: if one depot struggled, others absorbed the slack. Routes were proven. Drivers knew the stores. Cold-chain integrity was established practice, not aspiration.
The decision to consolidate logistics under DHL — operating from a single new depot alongside a partner called QSL — was a cost play. One depot, one contract, one streamlined operation. On a spreadsheet, the savings are obvious. On a PFMEA, the risk column lights up red immediately. Or should.
How it unfolded
The cutover happened in mid-February 2018. Within days, the single DHL-managed depot was overwhelmed. Deliveries to restaurants were delayed, incomplete, or simply absent. Chicken — the one product KFC cannot trade without — sat in warehouses or in transit beyond shelf-life windows. Stores that ran dry had no fallback. There was no second depot to reroute through, no Bidvest contract held in parallel as a safety net during transition.
Roughly two thirds of the UK estate closed. The brand published a full-page apology ad rearranging its own initials to spell FCK. Honest. Human. Self-aware. The ad was the response. The collapse was the preventable part.
Root-cause anatomy
Technically, the failure is textbook. You took a distributed, food-specialist network with built-in redundancy and replaced it with a single node operated by a partner whose core competence is parcel logistics, not fresh and frozen food distribution at this complexity. Cold-chain management, route density, store-specific delivery cadence, volume surge handling — all untested at full national scale on day one.
Organisationally, this is a procurement decision that was never stress-tested by the operations and quality functions with the authority to stop it. The cost case moved forward. The operational-readiness case was not made forcefully enough — or it was, and got overruled. We know the outcome. Nobody with veto power asked the question that actually matters: what happens when the single depot fails on day three at full volume?
Consolidation doesn't eliminate risk — it concentrates it. A single point of failure is just a cost saving that hasn't failed yet.
Where the quality system failed
The failure sits in process change management and PFMEA discipline. A logistics network change of this magnitude is a process change. Under any credible quality framework — IATF 16949, ISO 9001, a private standard — a process change of this scale demands a failure-mode analysis before approval, not after go-live.
The PFMEA should have identified, at minimum:
- Failure mode: single depot unable to meet delivery volume at full network load.
- Effect: restaurant stockouts, forced closures, perishable inventory loss.
- Severity: 9 or 10 — the business cannot trade.
- Cause: unproven depot capacity, cold-chain inexperience, no volume qualification, no parallel-run validation.
- Current controls: none credible — no fallback depot, no phased ramp, no Bidvest retention during transition.
When severity is 10 and occurrence controls are unproven, the gate does not open. APQP discipline says you qualify the new process at volume before you cut over. You run parallel. You phase. You hold the old contract until the new one has demonstrated it can carry the load for weeks, not hours. Change control was bypassed by commercial pressure.
What would have caught it
Three concrete actions, any one of which would have prevented or contained this:
- A credible PFMEA with the single depot named as a catastrophic failure mode. Severity 10, unproven occurrence controls. That document, signed by operations and quality leadership, should have been a gate condition for contract approval — not a post-incident artefact.
- A phased cutover with parallel running. Transition regions sequentially. Keep Bidvest active on a residual contract for a defined period. Run DHL at increasing volume over weeks. Measure delivery accuracy, cold-chain integrity, and store fill rates against threshold before each expansion.
- An operational-readiness audit before go-live. DHL and the depot partner running full-volume simulation for a sustained period, with independent observers verifying that the depot can actually handle 900 stores' worth of chicken, packaging, and consumables on a daily cycle without degradation.
My take
I have stood in rooms where a commercial decision collided with an operational-readiness reality, and the commercial side won because the cost case was concrete and the risk case was, to the untrained ear, hypothetical. At SNOP, building a 900-employee greenfield plant from nothing, we ran process validation for weeks before we let a single serial part ship to a customer. A new unproven depot is a new unproven process. You do not go live on faith.
In aerospace, where I work now, a process change of this magnitude without a PFMEA and a validated qualification plan would not pass a gate review. It would not survive an EASA or AS9100 surveillance audit. The cost of a single point of failure in aviation is catastrophic, so the discipline is enforced. In food and beverage, the cost showed up differently — not in safety, thankfully, but in £20–30 million of lost revenue and an apology ad that will outlive most marketing careers. The mechanism is identical. The discipline was absent.
What this means on your floor
- Any single-supplier or single-node consolidation is a PFMEA exercise first and a procurement exercise second. If the failure-mode analysis doesn't survive scrutiny, the business case is irrelevant.
- Operational readiness must be proven at full volume before go-live — not assumed, not promised, not projected. Run it. Measure it. Gate it.
- Phased transitions with parallel running cost more upfront and are the cheapest insurance you will ever buy. Hold the old contract until the new one earns the load.
- If severity is catastrophic and controls are unproven, change control does not approve. Full stop. That is the line quality leadership exists to hold.
The FCK ad was honest, human, and effective. It was also unnecessary. The collapse behind it was not bad luck or supplier incompetence exposed by circumstances — it was a process change approved without a credible failure-mode analysis, implemented without operational-readiness proof, and paid for by 600 stores with no chicken to sell. Consolidation without validation does not save money. It manufactures the single point of failure that eventually collects the bill.
