Decisions architected across 14+ enterprises.

Lead-time variability priced into safety stock, not into the planning solve. Expedite freight running 18% over budget. Senior planners attriting at 22% annually.

Architected a probabilistic supply substrate that priced lead-time variability into the planning solve and surfaced supplier risk as a learned distribution per lane.

Reduced expediting spend; freed working capital.

Sequential plant-by-plant scheduling that lost the network optimum at every handoff. Sub-optimal sequencing absorbed in changeover costs and rework.

Replaced sequential plant-by-plant scheduling with concurrent multi-horizon optimization across the network. Sequence-stability as a measured cost the policy weighed.

Network-level throughput up; changeover scrap down; planner confidence restored.

Safety stock sized by formula; seasonal and promotional cycles never priced in. Stockouts on Tier-A customers; markdowns on tail SKUs at quarter-end.

Built driver-based safety stock policies that adapted to seasonal-promotional cycles. Multi-echelon optimization sized buffers to actual variability per node.

Working capital reduced; Tier-A service up; markdown exposure down.

Critical-care SKUs running thin under demand spikes the deterministic plan never anticipated. Regulator-grade traceability tangled with manual workarounds.

Calibrated demand distributions per molecule and channel. Set policy-driven percentile commits per criticality tier. Pegged supply to demand at the line for full traceability.

Critical-care stockouts cut materially; audit posture defensible; planner load reduced.

Standard-cost margin overstated for high-touch customers. The customers consuming most of the operating margin were also consuming most of the changeovers and expedites.

Activity-based costing applied through the pegging cascade. ABC-adjusted margin became the commit objective; the picture of customer profitability changed.

Margin reallocated to defensible customers; contracts renegotiated on evidence.

Component shortages forcing allocation calls in spreadsheets. Strategic customers de-prioritized; gut-feel calls revisited in board reviews.

Allocation as a first-class policy. Tier weights, contractual penalties, downstream revenue, and customer LTV all priced into the same solve.

Allocation decisions defensible to the board; strategic accounts protected.

Promotional uplift modeled in marketing, never reaching planning in time. Stockouts on promo, glut after.

Causal demand modeling fed the planning solve. Promotional uplift as a learned response surface — modulated by elasticity, distribution, and stock-on-hand.

Promo stockouts down; post-promo overhang down; uplift conversion up.

Hundreds of suppliers; commit confidence opaque. Lead-time master data months stale.

Bidirectional supplier collaboration — buyer sends request-for-commits, suppliers return commit confidence and counter-proposals. Reliability distributions learned per supplier.

Lead-time master data healed; supplier risk priced in; expedite spend down.

Tier-2 and Tier-3 supply opacity. Single-source dependencies surfacing only after disruption.

Modeled Tier-2 and Tier-3 supply as a multi-echelon network with named risk events per node. Diversification investments justified by quantified risk-adjusted EVA.

Multi-tier exposure reduced; diversification roadmap defensible at the board.

Expiry exposure absorbed at month-end; FEFO discipline manual; regulatory exposure on near-expiry product.

Shelf-life and perishability as first-class constraints in the planning solve. FEFO + expiry-aware allocation. Driver-based safety stock adapted to lot-level signal.

Expiry write-offs down; FEFO compliance up; regulator posture strengthened.

Sequence-dependent setup costs absorbed plant-by-plant. Energy and scrap losses on every avoidable changeover.

Sequence-dependent setup, changeover scrap, and energy as first-class variables in the optimization. Campaign-style sequencing where the chemistry allowed.

Effective throughput up; changeover energy and scrap down; sequence stability up.

Contract manufacturers modeled as black-box suppliers. Allocation decisions to CMs run on contract minimums rather than on EVA. CM disruptions surfacing as procurement, not planning, problems.

Modeled each CM as a first-class production source with allocated capacity, component inventory, and routing alternatives. Make-vs-buy evaluated against EVA in the same solve.

CM allocation defended by math; multi-CM redundancy converted into measurable resilience.

Capable-to-promise from stale snapshots. Tier-A customers losing dates the system never knew it could not hold.

Continuous capable-to-promise calibrated to actual supply distributions. Demand-supply matching segmented by commercial value, not by SKU master.

Promise reliability up; manual rescheduling down; Tier-A service measurably improved.

Replan cycles touched every plan when one event broke. Plan thrash absorbed in expedite and overtime.

Event-triggered decision emission within policy bounds. Surgical replan sized to the minimum impact radius — the rest of the plan stays committed.

Plan thrash eliminated; expedite spend reduced; planner trust in the system restored.