8 recoveries. And what each one does not prove.

The position

Full commercial and operational accountability for an 11-account defence and transport portfolio: 40 sites, 25 people, a £29.4m plan target.

The flagship BAE Systems partnership was at risk ahead of the extension decision. Delivery ran at 86% on-time in-full.

A £263k credit exposure sat open with no agreed cause, and nobody could state the portfolio’s slow-moving stock position. That absence is usually the tell: where the number does not exist, every decision above it is a guess.

What the analysis proved

• Roughly 97% of the £263k exposure sat in 2 products, driven by calendar-based replenishment - a cause internal teams had looked for and missed.
• Across 17,000+ SKUs, fewer than 200 drove half of all sales. £23.3m of slow-moving and obsolete exposure was classified for the first time.
• £57,567 of margin leakage across 181 products, plus £40k of annual surcharge exposure.
• Rejected deliveries traced to labels carrying the internal purchase order instead of the customer’s - a transaction-logic fault no inspection could fix.

What changed

Service first. SLAs redesigned around what sites actually needed, a weekly performance forum, and headline review replaced with exception management - cause separated from ownership, named owners, next actions.

Then the stock. The operating model was rebuilt on ABC and XYZ segmentation, with a 3-tier governance model setting what was allowed back into the range.

The commercial recovery ran alongside, moving the engagement from transactional supply to strategic partnership at prime level. The FY27 forecasting architecture was built from nothing.

What it returned, and what stayed

Left behind: a 16-metric KPI tracker the customer and board could both read, automation removing an estimated 100 hours a year, a workforce atlas covering 200,000+ employees across nearly 400 sites, and a 25-person team on an embedded cadence. None of it needed me to operate it.

What this does not claim

The credit analysis converted an undifferentiated £263k exposure into a defined, concentrated target. It does not claim a cash recovery, because the reconciliation outcome was not verified. Contract retention was a board decision with many inputs; the recovery programme was one of them, not the sole cause.

The position

Air, land, maritime and naval operations each making site-level decisions in isolation. No common rhythm linking usage, service performance, stocking levels, underwritten workwear, forward demand and back orders.

Without that link, teams could report a shortage indefinitely without ever establishing whether the cause was inventory policy, shifting demand, an exiting product or an inadequate supply route.

What changed

S&OP was reframed from a retrospective review into an exception-led operating process. ABC segmentation set consumption, stocking settings, service exposure and forward demand against each other.

Trust was built by separating immediate continuity controls from the reversible, longer-term policy decisions - so people could agree to the urgent without conceding the strategic.

What it returned

What this does not claim

Delivery-to-promise improved by roughly 6 percentage points and availability by nearly 3 over the recovery period. These are measured portfolio outcomes across a period in which several things changed at once; the evidence does not isolate the S&OP redesign as the sole cause, and it is not presented as doing so.

The position

A framework exit ahead of the contract end date, exposing the supplier to branded and bespoke stock, non-cancellable inbound orders and uncertain service-credit obligations.

Hundreds of items, allocation thresholds, run rates and notice clauses all had to be reconciled. The conversation had started as a write-off discussion.

What changed

Sole commercial lead. An 8-scenario, line-level liability model separated on-hand stock, non-cancellable inbound, dedicated items and shared lines - then tested specification, allocation, recoverability, demand and exit horizon as variables rather than assumptions.

Crucially, the analysis was wired back into replenishment controls, so ordering reflected the planned exit rather than historic settings. The exposure stopped growing while it was being argued about.

What it returned

What this does not claim

This is a reduction in the classified liability, not a claimed saving. The shared stock did not disappear; it was directed into consumption, transfer, redeployment or disposal. The value was in making the exposure defensible line by line before a settlement negotiation, not in making it smaller on paper.

The position

Quality management across an ISO 9001 site during a board-mandated turnaround. The defect rate stood at 32% and was being managed as something the process did, rather than something wrong with it.

Warranty information existed but was dispersed, with nobody accountable for turning customer failures into manufacturing learning. Cases could be closed individually without a recurring failure mode ever surfacing.

What changed

A dedicated warranty function established as a central quality-capture point, combining recurring field patterns with internal defects through Pareto analysis, SPC, capability studies and operator knowledge.

Production and engineering were brought into PFMEA to locate the failure mechanisms, then controls implemented at source - standard work, visual management, Kaizen, error-proofing - rather than by adding end-of-line inspection.

What it returned

What this does not claim

The 15% in-process reduction is attributable to the PFMEA and critical-control work specifically. The move from a significant baseline to zero was delivered by the wider quality programme over the period, not by any single intervention.

The position

Delegated authority across operations, quality, engineering and environment during corporate restructuring. An integrated ISO 9001, 14001 and 45001 system in name only - 3 standards that had grown separately and duplicated each other’s paperwork.

DMAIC analysis showed the constraint was queue time, not work time: claims accumulating at the handoffs between teams. Energy sat outside the system entirely, treated as an overhead rather than a process variable.

What changed

A DMAIC-led redesign aimed at the queues rather than the work. Document control and CAPA moved into a single 5S-based system with real-time non-conformance visibility.

Energy was brought inside the system as a Six Sigma programme that tested whether each unit of consumption supported a defined production, quality or compliance requirement before proposing any change. Operator and Quality challenge was used as a control, not an obstacle.

What it returned

What this does not claim

The energy reduction is measured against the site’s own established baseline. The clean audit record is contextual evidence that controls remained intact during the change - not a claim that the energy programme caused the audit result, nor that a specific sustainment period has been verified.

The position

A factory running batch-oriented flow with an inefficient layout, creating bottlenecks, queues and delays between manufacturing stages.

Value-stream analysis showed much of the elapsed cycle time was not productive processing at all. The task was to move work through the process differently, not to ask operators to work faster.

What changed

The initial cell design was treated as a hypothesis to be tested, not a solution to be installed. More than 20 Kaizen events with production, maintenance and shop-floor employees iteratively adjusted layouts, standard work and visual controls.

Tools, materials and work-in-progress were repositioned around the manufacturing sequence with the operators, then written into the 5S standard so the arrangement could not quietly revert.

What it returned

What this does not claim

These are site-level outcomes over a period of sustained Lean deployment, not the result of a single event. 5S audit scores improved materially and weekly results held in the 70s and 80s, which is the evidence the change stuck - but no claim is made beyond the period observed.

The position

A supplied component producing a documented scrap rate of roughly 30% at the boundary between supplier quality and internal production.

Because the scrap was being caught internally, the pull was to manage it internally - adjust, sort, absorb. That is the quiet failure mode in a regulated operation: a problem that is contained is a problem nobody escalates.

What changed

Fishbone analysis kept material, method, machine, measurement, people and supplier inputs all in scope, then 5 Whys traced the failure upstream. The component was compared before and after entering production to locate where the condition was created.

The supplier was worked with through the evidence rather than blamed, which is what made the corrective action land at source. Effectiveness was verified using the same downstream measure that exposed the problem - not by confirming a form had been completed.

What it returned

What this does not claim

The scrap result was verified in the production environment rather than merely closed on paper, but the evidence does not establish a specific long-term sustainment period. The downtime and equipment-integration figures are separate pieces of work from the same role, listed for context.

The position

Key UK accounts facing recurring supply disruption, supplier-performance gaps, procurement bottlenecks and cost pressure across multiple sites.

The pattern underneath was familiar: the lowest purchase price was masking higher operational cost - long lead times, unreliable availability, replenishment failures and product complexity the price comparison never showed.

What changed

Current-state flow mapped with Lean Six Sigma and value stream mapping, then Total Cost of Ownership applied to establish whether each constraint came from supplier performance, lead time, replenishment, product complexity or the commercial structure.

Only then was the intervention selected, rather than applying one lever to every problem. Outcomes were measured specifically to confirm the constraint had been removed, not moved somewhere else.

What it returned

What this does not claim

The £4m is a programme total contributed to, not delivered alone. What was personally owned was the diagnosis, the intervention design, the cross-functional implementation and the measurement. The £200k per-site rationalisation saving is the documented, directly attributable component.