Written by Tsujita — Paddy & Water Inc. / Supply Chain

— Cross-Industry Supply-Demand Optimization: Detailed Requirements and the Ideal Core System —

Industry-Specific Characteristics Analysis / Detailed Requirements / Stable Parts Supply for Capital Equipment / Core System Requirements

July 2026

This paper explains, across industries, the requirements and ideal form of global PSI (Production, Sales, Inventory) planning spanning multiple countries and multiple sites. From consumer goods, where demand volatility is large and speed is paramount, to capital equipment, where a stable supply of parts determines competitiveness, it delves into the characteristics and requirements specific to each industry, and discusses what the core systems (ERP / planning platform) that support them “should be.” It also systematizes the detailed functional and non-functional requirements that should be defined at the time of construction and implementation. The primary intended readers are practitioners in SCM, production management, and IT planning.

Table of Contents

・Chapter 1 What Is Global PSI

・Chapter 2 Global PSI for Consumer Goods (A Representative Example of Demand-Volatile Industries)

・Chapter 3 Analysis of Global PSI Characteristics by Industry

・Chapter 4 What the Core System Should Be

・Chapter 5 Global PSI Platform Implementation Roadmap

・Chapter 6 Learning from Leading Cases: The Ideal Form of Global PSI

・Chapter 7 Case Studies of PSI/IBP at Global Companies (by Company and Industry)

・Chapter 8 Detailed Requirements for Global PSI (Functional and Non-Functional)

・Chapter 9 Conclusion

Chapter 1 What Is Global PSI

1-1 The Basics of PSI and the Implications of Globalization

PSI is the concept of supply-demand coordination that manages Production (production/receiving), Sales (sales/shipment), and Inventory in an integrated manner along a time axis. It establishes the identity “ending inventory = beginning inventory + receipts (P) − shipments (S)” across all periods, all SKUs, and all sites, thereby anticipating and preventing both excess and shortage. Global PSI is what results when this is run not at a single domestic site, but across a network spanning multiple countries, multiple sites, and multiple currencies.

・Differences from domestic PSI: long transportation lead times, tariffs and trade regulations, geographic separation between demand locations and supply locations, and multiple currencies and units of measure

・Additional elements: distribution requirements planning (DRP) across sites, unified visibility of global inventory, and preparedness for geopolitical and supply-disruption risk

1-2 Why Consumer Goods Are at the Forefront of Global PSI

Among the many industries, consumer goods is the field where the difficulty and value of global PSI appear most sharply. A large number of SKUs are sold through multiple channels worldwide, demand shifts suddenly due to promotions, seasonality, and trends, and product life cycles are short. Because this “fast, high-volume, and hard-to-read” demand must be synchronized with production and inventory that are geographically dispersed, consumer goods has become both the proving ground and the driving force for advanced global PSI. The global consumer goods market is estimated at approximately USD 3.4 trillion as of 2025.

? The global PSI methods established in consumer goods (demand sensing, omnichannel inventory, MEIO) serve as a “leading indicator” that spreads to other industries. This is also why this paper centers on consumer goods.

Chapter 2 Global PSI for Consumer Goods (A Representative Example of Demand-Volatile Industries)

2-1 Characteristics of the Consumer Goods Supply Chain

To understand global PSI for consumer goods, one must first grasp the characteristics of its demand and supply. Compared with general manufacturing, demand volatility is greater, and speed of decision-making is directly linked to competitiveness.

・Many SKUs, many brands: the number of items is enormous due to differences in flavor, capacity, and packaging. Forecasting units are fine-grained and management is complex

・Short product life cycles: new product introduction (NPI) and end-of-life (EOL) occur frequently. Inventory design for ramp-up and phase-out is important

・Promotion, seasonal, and trend volatility: demand can suddenly change several-fold due to sales campaigns, new commercials, or social media buzz

・Omnichannel: the same inventory must be optimally allocated across multiple channels such as stores, e-commerce, wholesale, and D2C

・Freshness and shelf life: for food and beverages, freshness management and waste reduction through FEFO (First-Expired, First-Out) are essential

2-2 Demand-Driven Planning and Demand Sensing

The starting point of global PSI for consumer goods is highly accurate demand forecasting. Beyond historical results alone, “demand sensing,” which incorporates real-time market signals, captures localized, short-term demand fluctuations.

・Internal data: shipments, orders received, promotional plans, and prices — the company’s own planning information

・External data: POS (point-of-sale) data, inventory, weather, macroeconomic indicators, and social media sentiment — signals close to real demand

・AI forecasting: machine learning integrates this data to forecast short-term demand by region, channel, and SKU with high accuracy

? PSI for consumer goods becomes more accurate the closer it moves from “shipment-based” to “consumption-based (POS/actual sales).” Sharing real demand data from downstream with upstream production and procurement directly suppresses the bullwhip effect.

2-3 Synchronizing Global Supply and Demand — Coordinating Centralized Production and Regional Demand

In consumer goods, a geographic and temporal gap arises between centralized production aimed at economies of scale (mother factories) and local demand in each country. Global PSI closes this gap through distribution requirements planning (DRP) between sites and the placement of strategic inventory.

・DRP (distribution requirements planning): plans how much to allocate and replenish to which site, from central warehouse → regional DC → customer

・Risk pooling: consolidating inventory upstream offsets variability, curbing total inventory while preventing regional stockouts

・Managing in-transit inventory: visualizing inventory in long-lead-time transportation such as ocean freight, and making replenishment decisions proactively

2-4 PSI for Promotions, New Product Introduction, and End-of-Life

A challenge unique to consumer goods is supply-demand management for promotions, new product introduction (NPI), and end-of-life (EOL). In each case, demand is non-stationary, and normal forecasting logic does not work well.

・Promotional PSI: demand forecasting that incorporates promotional effects, together with advance production increases and inventory build-up. Manages both the risk of leftover stock and the risk of stockouts

・New product introduction (NPI): forecasting demand with no track record via analogy (similar products), and designing initial inventory. Adjusting flexibly once the launch trajectory is observed

・End-of-life (EOL): a production-reduction and inventory-drawdown plan to sell off remaining stock. Exit design that minimizes write-off losses

2-5 Omnichannel Inventory and Freshness Management

・Omnichannel inventory: unifying visibility of inventory across stores, e-commerce, and DCs, and performing optimal cross-channel allocation (Available-to-Promise)

・Freshness and FEFO: allocating from lots closer to their expiration date to curb waste, and building pre-expiration discounting and diversion into the plan

2-6 Inventory Reduction Approaches for Consumer Goods

・MEIO (multi-echelon inventory optimization): optimizing the multi-tier network from central to regional to store as a whole, reducing total inventory by 10-30%

・Postponement: holding stock as common semi-finished goods and finalizing packaging and labeling after demand is confirmed, compressing SKU-level inventory

・Segment-specific inventory policy: using ABC/XYZ segmentation to keep fast movers thin and fast, minimize slow movers, and manage promotional items in a separate bucket

[Initiative example] A global consumer goods manufacturer: built a global PSI combining demand sensing that incorporates actual POS sales data, multi-echelon MEIO, and monthly S&OP. In categories with large promotional and seasonal swings, it reduced excess-inventory risk by more than 30% while curbing in-store stockouts (lost shelf presence).

2-7 Key KPIs for Consumer Goods PSI

KPI	Definition	Significance for Consumer Goods
Fill Rate / OTIF	Rate of immediate fulfillment and on-time delivery against requests	Prevents shelf stockouts and protects brand value and retailer relationships
Forecast Accuracy (MAPE/Bias)	Error in demand forecasting	The foundation of PSI. Especially important for promotions and NPI
Days of Supply (DOS)	Average inventory ÷ daily shipments	Freshness and working capital. The shorter, the more agile
Write-off Rate	Losses from expiration and obsolescence	Important for food. Curbed through FEFO and forecast accuracy
Cash Conversion Cycle (CCC)	Days of inventory + receivables − payables	Working capital efficiency. Directly tied to management indicators

Chapter 3 Analysis of Global PSI Characteristics by Industry

The global PSI methods refined in consumer goods are also applied to other industries, but because product characteristics, regulations, and demand patterns differ by industry, the “functions that must be prioritized” change. The major industries are compared side by side below.

Industry	Demand and Product Characteristics	Functions Emphasized in PSI	Key Inventory Considerations	Core System Focus
Consumer goods (CPG)	Many SKUs / promotions / seasonality / short life / omnichannel	Demand sensing / DRP / FEFO	Keep fast movers thin and fast, minimize slow movers	Demand planning + real-demand linkage + inventory optimization
Pharma / healthcare	Strict regulation / traceability / temperature control	Lot/expiration management / FEFO / supply responsibility	Thicker safety stock, stockouts strictly prohibited	Lot/serial management / GxP / regulatory compliance
Automotive	JIT/JIS / two-layer forecast and confirmed orders	Forecast planning / kanban / leveling (heijunka)	Minimize in-process inventory	EDI / production synchronization / multi-tier collaboration
Industrial machinery / equipment	Engineer-to-order (ETO) / long lead times / project-based	Multi-year forecasting / advance procurement of long-lead parts	Minimal finished-goods inventory; long-lead and service parts held as strategic stock	Project inventory / BOM complexity
Electronics	Extremely short life / difficult parts procurement / price decline	Securing parts supply / demand-supply response	Avoiding obsolescence risk; turnover is the top priority	Parts procurement / supply response / ATP
Chemical / process	Continuous/batch / equipment constraints / byproducts	Capacity-constrained planning / recipes / batch management	Balancing equipment utilization and inventory	Process orders / batch management / capacity planning

3-1 Pharma and Life Sciences — Regulation and Supply Responsibility

Pharmaceuticals and medical devices are premised on strict regulation, complete traceability, and cold-chain (temperature-controlled) logistics. Lot and expiration-date management, FEFO (First-Expired, First-Out), and serialization (unit-level identification) are mandatory, and because stockouts can directly affect patients’ lives, safety stock is set thicker than in other industries, prioritizing compliance over cost efficiency.

・Lot / expiration date / FEFO: ensures quality and freshness while curbing disposal due to expiration

・Supply responsibility (service level as top priority): maintains a high fill rate, since a stockout equals patient impact

・Cold chain: visibility of temperature excursions, and quarantine/tracking of inventory when excursions occur

・Regulatory compliance: GxP, serialization (anti-counterfeiting), and audit trails secured through the system

・Preparedness for high uncertainty: multiple scenario-based plans prepared on the premise of demand surges, raw-material shortages, and production capacity shortfalls

? Insight from case studies: As J&J (Chapter 7) shows, a PSI professional in pharma is not a “plan creator” but a “decision-support provider under uncertainty.” Combining multi-echelon inventory optimization (MEIO) with scenario-driven IBP, so as to achieve both regulatory compliance and supply responsibility, is the essence of PSI in this industry.

3-2 Automotive — Two-Layer Forecast/Confirmed Planning and Supplier Integration

Automotive is fundamentally based on JIT/JIS supply synchronized with the finished-vehicle maker’s production plan, requiring a PSI that handles medium- to long-term “forecasts” and short-term “confirmed orders” as two layers. Demand is leveled across the entire multi-tier supplier network, and in-process inventory is minimized through sequential replenishment via kanban. In recent years, as symbolized by the semiconductor shortage, achieving both “efficiency (minimal inventory)” and “resilience (preparedness for supply disruption)” has become a new point of discussion.

・Two-layer forecast and confirmed planning: long-term forecasts prepare capacity and materials, while short-term confirmed orders translate into execution

・Production leveling (Heijunka): smooths the peaks and valleys of demand, suppressing the propagation of variability (bullwhip) to suppliers

・Kanban and sequential replenishment: minimizes inventory between processes and between suppliers

・EDI integration: links order and kanban information in real time from OEM through Tier 1 to Tier 2 (a lifeline)

・Multi-tier visibility and risk management: visualizes Tier 2/3 supply risk, single-sourcing, and geopolitical risk (a lesson from the semiconductor shortage)

・Traceability: recall and quality tracking by batch/serial number

? Insight from case studies: what Toyota (Chapter 7) demonstrates is “speed of response to supply-demand change, rather than perfect forecasting.” The essence of automotive PSI lies not in holding inventory, but in demand linkage (pull) and rapid recovery capability in the event of an anomaly. Furthermore, as Cisco’s multi-tier supplier visibility shows, globally, “supplier integration” determines the success of PSI even more than management of one’s own processes.

3-3 Electronics and High-Tech — The Trade-off Between Extremely Short Life Cycles and Parts Procurement

Electronics and high-tech products have extremely short product life cycles, and prices fall quickly. Avoiding obsolescence risk and maximizing inventory turnover therefore become the highest priorities. At the same time, preparing for difficulty in procuring critical components such as semiconductors and their long lead times is also essential, creating the difficulty of managing the conflicting goals of “securing parts firmly while not overproducing finished goods” at the same time.

・Avoiding obsolescence: accelerates supply-demand response and avoids leaving inventory of old models (turnover as the top priority)

・Securing supply of critical components: long-term capacity reservations, dual sourcing, and strategic inventory for items such as semiconductors

・Supply response: immediate supply response to demand changes; controlling allocation during shortages through ATP/allocation

・Short-cycle PSI: replanning supply and demand on a weekly-to-daily basis to keep pace with market change

? Insight from case studies: Apple (Chapter 7) reserves semiconductor, display, and EMS capacity more than a year ahead of launch, and when supply is short, allocates not evenly but by profit and strategic importance. Samsung, though largely operating on a monthly cycle, conducts “weekly” replanning. The lesson from high-tech PSI lies in a two-pronged approach: “securing upstream capacity (long-term)” and “rapid downstream replanning (weekly).”

3-4 Chemical and Process Industries — Equipment Constraints, Batches, and Data Integration

The chemical and process industries have inherent constraints such as continuous or batch production, equipment capacity constraints, recipes, byproducts, and hazardous materials. Planning must balance equipment utilization with inventory while planning the multiple stages from raw materials through intermediates to finished products under capacity constraints. Beyond sophisticating demand forecasting, capacity planning and data integration with suppliers and customers are the key to overall optimization.

・Capacity-constrained planning: builds the supply plan with equipment and reactor capacity as constraints (incorporating the ceiling on producible volume)

・Recipe and batch management: planning by formulation, process orders, and lot/batch units

・Byproducts and co-products: managing, as a whole, the supply and demand of byproducts generated simultaneously with the main product

・Hazardous materials and regulatory compliance: regulations on storage and transportation, and SDS (Safety Data Sheet) compliance

・Data integration and logistics optimization: data linkage with suppliers and customers combined with JIT, and optimization of transportation and loading

? Insight from case studies: BASF (Chapter 7) achieves overall optimization and JIT inventory through big data, IoT, and data integration with partners, while Dow reduced logistics costs by approximately 18% through AI demand forecasting and load optimization. The lesson from chemical PSI is that, in addition to “precision in capacity-constrained planning,” “data integration with suppliers and customers” and “logistics optimization” substantially move the cost structure.

? The principle running through every industry is the same: capture real demand, anticipate supply and demand, and place inventory only where and in the amount needed. The difference lies in “which constraints and functions to reinforce,” and that difference appears as differences in core system requirements.

3-5 Capital Equipment and Industrial Machinery — Global PSI Centered on Stable Parts Supply

Global PSI for capital equipment (industrial machinery, plant equipment, semiconductor manufacturing equipment, and the like) has characteristics that are the polar opposite of consumer goods. Finished machines are engineered-to-order/made-to-order (ETO/MTO), produced in low volumes with long lead times, and demand is discrete, arising on a project (deal) basis. However, the real battleground is not the plan for the finished machine itself, but rather “how to keep supplying, without interruption, the many parts that make up the equipment — especially long-lead-time parts and dedicated parts.” Because a stopped piece of equipment halts the customer’s entire production line, the responsibility to supply maintenance parts (service parts) after delivery is also extremely heavy.

Requirements for PSI Aimed at Stable Parts Supply

・Part-level PSI: runs a part-level PSI separate from the finished-machine PSI. Common parts aggregate demand across multiple projects and models, offsetting variability

・Advance procurement of long-lead-time parts: based on demand forecasts and preliminary indications, even before a deal is confirmed, places advance orders for critical long-lead-time parts. Bottleneck parts in the BOM are identified in advance

・Visualization of and countermeasures for supply risk: visualizes the risk of single-source, dedicated, and end-of-life (EOL) parts, and prepares through dual sourcing, alternative designs, and lifetime buys (a final bulk purchase)

・Strategic inventory (buffer): holds long-lead-time or supply-uncertain parts as strategic inventory, consolidated upstream to gain a risk-pooling effect

・Maintenance parts planning: forecasts failures from the installed base (number of units in operation, model, and age), plans demand for maintenance parts, and positions them at regional service locations

・Collaboration with suppliers: stabilizes parts supply through forecast sharing, long-term contracts, and VMI. Unlike automotive JIT, this calls for collaboration that prioritizes “long lead time, low volume, and certainty”

・Linking design and procurement: reduces the number of parts through commonization and standardization, consolidating demand to make it easier to read (PSI starting from the design stage)

The essence of PSI for capital equipment lies in planning “two layers of demand” separately. That is, demand for parts for new equipment (deal-linked, discrete) and demand for maintenance parts (installed-base-linked, continuous) differ in nature, and so are forecast and inventory-designed along separate axes.

Point of Discussion	Difference from Consumer Goods	Response in Capital Equipment
Nature of demand	Continuous, high-volume → discrete, low-volume (deal-based)	Anticipated through the deal pipeline and forecasts
Where the battle is won	Avoiding shelf stockouts	Avoiding disruption in the supply of parts (especially long-lead-time parts)
The main role of inventory	Turnover of finished goods	Strategic inventory of long-lead-time and maintenance parts
Risk	Obsolescence and write-off	Supply interruption of single-source, EOL, or dedicated parts
Suppliers	High-speed replenishment linked to real demand	Long-term contracts, shared forecasts, and an emphasis on certainty

[Case image] An industrial machinery / semiconductor manufacturing equipment maker: incorporated advance procurement of long-lead-time parts, demand aggregation of common parts, lifetime buys of EOL parts, and maintenance-parts forecasting from the installed base into its global PSI, raising equipment on-time delivery rates while curbing maintenance-parts stockouts. It reflected in its PSI the philosophy that “equipment does not end when it is sold — value is only created by continuing to supply parts.”

? For capital equipment, global PSI is a matter not so much of “how to build the finished product” as of “how to keep supplying parts without interruption.” Advance procurement of long-lead-time parts, visualization of supply risk, and installed-base-based forecasting of maintenance parts are the three pillars of stable parts supply.

Chapter 4 What the Core System Should Be

To run global PSI, it is fundamental to view the core system as three layers — “execution ERP,” “planning,” and “visualization (control tower)” — with an architecture in which master data runs through all of them.

4-1 The Three-Layer Architecture — Execution, Planning, Visualization

Layer	Role	Representative Solution	Function in PSI
Planning	Supply-demand planning / inventory optimization	An APS such as SAP IBP	Demand forecasting / S&OP / MEIO / supply response
Execution (ERP)	Order management / inventory / production / cost	SAP S/4HANA	Execution of MRP, inventory, production, procurement, and shipment
Visualization (Control Tower)	Monitoring / alerts / decision-making	Supply Chain Control Tower	Overall visualization / anomaly detection / replanning linkage

4-2 Requirements for the Execution ERP (S/4HANA)

The execution ERP is what actually runs the supply and demand decided in planning. As the foundation of global PSI, the following capabilities are required.

・Single Source of Truth: centrally manages inventory, order, and production data, referenceable in real time

・Real-time MRP (MRP Live): performs requirements calculation at high speed on HANA, rather than through an overnight batch, responding immediately to supply-demand changes

・Advanced ATP (aATP): determines omnichannel inventory allocation and shippability in real time

・Demand-Driven Replenishment (DDMRP): achieves buffer-based, demand-driven replenishment that is resilient to variability

・Logistics execution (EWM/TM): integrates warehousing (EWM) and transportation (TM), supporting global logistics at the execution layer

・Multi-site, multi-currency, multi-language: the international support needed for global operations, together with trade (GTS) integration

4-3 Requirements for the Planning Layer (IBP / APS)

It is the planning layer (Advanced Planning System) that sophisticates PSI itself. Because the execution ERP alone cannot perform future supply-demand simulation or multi-echelon optimization, a dedicated planning platform is required.

・Demand planning and demand sensing: forecasts demand using statistics plus AI, incorporating POS and external data

・S&OP / IBP: integrates demand, supply, and finance into a single plan, involving management on a monthly basis to agree on supply-demand gaps

・Inventory optimization (MEIO): optimally positions safety stock across the whole network, reducing total inventory

・Response and supply planning: a supply plan that considers constraints, together with immediate response to demand changes

・Control tower: bundles the components of IBP together, visualizing and notifying of anomalies such as supply disruption

? The execution ERP (S/4HANA) and the planning layer (IBP) have different roles and function as two wheels of the same vehicle. ERP “runs the present accurately,” while the planning layer “reads the future optimally.” Loosely coupling these two layers is the essence of a global PSI platform.

4-4 Master Data — The Foundation of Everything

No matter how sophisticated the planning and execution systems are, they will not function if the quality of master data is low (garbage in, garbage out). It is no exaggeration to say that the success or failure of global PSI is, in fact, 80% determined by master data maintenance.

・Scope of maintenance: item master, BOM, sites/hierarchy, lead times and their variability, current and in-transit inventory, real-demand history, and ordering constraints (such as MOQ)

・Global unification: standardization of item codes, units, calendars, and currencies. Eliminates the situation where “the same item has different codes” across sites

・Governance: defines rules and responsibilities for registering and changing master data, continuously maintaining data quality (MDM)

4-5 Architectural Choices for a Global Core System

In global deployment, the placement policy for the core system is itself also a design challenge. The trade-off between single consolidation and distribution-plus-integration is chosen according to the business structure.

・Single Instance: consolidates the entire world into a single ERP. Data unification and visibility are excellent, but the difficulty of implementation and governance is high

・Distributed + integrated: each region has its own ERP, integrated through the planning layer (IBP) and a control tower. Flexible, but requires careful integration design

・Two-tier ERP: a practical compromise in which headquarters runs S/4HANA while subsidiaries/regions run a lightweight (cloud) ERP, bundled together at the planning layer

4-6 What to Prioritize in the Core System by Industry

Industry	Focus of the Execution ERP	Focus of the Planning Layer
Consumer goods	aATP, omnichannel inventory, FEFO	Demand sensing, promotional planning, MEIO
Pharma / healthcare	Lot/serial management, GxP audit trails	Expiration-date planning, supply responsibility, regulatory compliance
Automotive	EDI, production synchronization, kanban	Two-layer forecast/confirmed planning, leveling
Industrial machinery	Project inventory, complex BOM	Multi-year forecasting, long-lead-time parts planning

4-7 The Future — AI, Digital Twins, and Autonomous SCM

The future of the core system heads toward an “autonomous supply chain” driven by AI and digital twins. It will evolve toward a level where demand is sensed in real time, scenarios are simulated on a digital twin, and, within a certain range, replenishment and replanning happen automatically without human intervention. Consumer goods is at the forefront of this trend.

Chapter 5 Global PSI Platform Implementation Roadmap

Phase	Initiative	Priority Point for Consumer Goods	Core System
Phase 1: Visualization	Unified visibility of global inventory and supply-demand	Visualizing inventory across all channels	ERP integration, BI, Control Tower
Phase 2: PSI Integration	Integrate production, sales, and inventory into PSI; begin S&OP	Supply-demand agreement including promotions/NPI	IBP (Demand, S&OP)
Phase 3: Inventory Optimization	Statistical safety stock, MEIO	Differentiating fast movers/slow movers, FEFO	IBP (Inventory Optimization)
Phase 4: Autonomy	AI forecasting, demand sensing, automatic replenishment	Automatic replenishment linked to real demand	IBP + AI + Control Tower

5-1 Success Factors and Pitfalls

・Master data quality comes first: both planning and optimization are proportional to the quality of master data. Prepare the data before introducing tools

・Process and organization: the real work is building an operational model in which sales, production, SCM, and management agree within the same PSI/S&OP

・Phased implementation: do not aim for AI/MEIO right away; consolidate visualization and PSI integration as a foundation first, then sophisticate

・Avoiding dual management: clearly design which is authoritative — execution vs. planning, headquarters vs. region

⚠ The core system is a means, not an end. Even with a sophisticated planning engine, nothing functions without a supply-demand agreement process and maintained master data. Building the foundation in the order “mechanism (PSI/S&OP) → data → system” is, though it may look like a detour, the fastest route.

Chapter 6 Learning from Leading Cases: The Ideal Form of Global PSI

Chapter 1 described “why global PSI is necessary.” This chapter takes a further step and concretizes “then, what should it be,” by studying the initiatives of leading global consumer-goods and consumer-electronics makers. Analyzing the cases of leading companies, the ideal form of global PSI can be summarized into the following five design principles.

6-1 The Ideal Form — Five Design Principles

Principle	Content	Evidence Seen in Cases
① A Single Plan (One Plan)	All functions look at the same single plan	Major household-goods maker: consolidated 35 processes into 20
② Layering of Planning Levels	Links strategic (3-year), tactical (annual), and execution (weekly) levels	Major food maker: 3-year + 1-year + weekly buckets
③ A Regular Cadence (Rhythm)	Discipline of monthly S&OP/IBP plus weekly execution	Major food company: monthly operation of an 18-month MBP
④ Global Standards × Regional Execution	Balancing standard processes with local discretion	A Planning CoE plus regional teams
⑤ Sophistication of Forecasting and Scenarios	Anticipating the future through predictive analytics and digital twins	Major consumer-electronics maker: a digital twin of its logistics network

? These five principles are not independent but build upon one another. On the foundation of ① a single plan, ② layering and ③ a rhythm create a pattern of operation, ④ a CoE balances standardization with local execution, and ⑤ forecasting and digital twins strengthen the ability to anticipate.

6-2 Case Study ① Consolidation into a Single Plan — A Major Household Goods Maker

A leading global household-goods (CPG) manufacturer had, over years of expansion, been troubled by numerous siloed planning processes and overlapping roles. The company fundamentally redesigned this around the philosophy of “one system, one planner, one end-to-end plan,” consolidating 35 planning processes down to 20. It further shifted from periodic planning to “continuous” planning and execution that is constantly updated in response to supply-demand changes.

・Key point: PSI breaks down when each function holds a separate, disconnected plan. Consolidating into a “single plan, single source of truth” is the first foundation

・Clarifying roles: clarifies responsibility for demand, supply, and inventory on a single plan, eliminating duplication and turf

・Making it continuous: moves from monthly-batch planning to continuous planning and execution that keeps pace with real-demand changes

[Lesson] The starting point of global PSI is not advanced AI but “integration of planning and process.” Bundling into a single plan first, defining roles, and creating a pattern that runs continuously is the precondition for all subsequent sophistication.

6-3 Case Study ② Layering Planning Levels and a Regular Cadence — Major Food and Beverage Makers

Several major food and beverage makers operate their planning not at a single level of granularity, but layered by time horizon. One company links three layers — a three-year long-term plan, a one-year detailed plan, and a weekly-bucket forecast — while another rolls out an 18-month-horizon monthly business plan (MBP) worldwide, using predictive analytics as the main input to S&OP.

・Planning layers: changes granularity by horizon — broad strokes (strategy) for the distant future, detail (execution) for the near term

・A regular cadence: runs monthly S&OP/IBP (demand → supply → supply-demand reconciliation → management review) as a discipline

・Embedding predictive analytics: places statistical and AI-based demand forecasting at the foundation of the human consensus process

[Lesson] The ideal PSI is “a single plan” and, at the same time, “a layered group of plans.” Linking the strategic, tactical, and execution layers with monthly and weekly rhythms achieves both long-term direction and near-term responsiveness.

6-4 Case Study ③ Digital Twins and Control Towers — A Major Consumer Electronics and Logistics Equipment Maker

The leading edge of planning sophistication is digital twins and control towers. One major consumer-electronics maker has built a digital twin of its finished-goods logistics network spanning 188 sites worldwide, simulating and optimizing warehouse location, scale, service levels, and network configuration. A major logistics-equipment maker, meanwhile, has implemented an AI control tower on top of a digital twin, monitoring and controlling the flow of goods and orders in real time.

・Control tower: centrally monitors global supply, demand, inventory, and transportation, instantly detecting and notifying of anomalies

・Digital twin: replicates the network and supply-demand situation in a virtual space, running what-if simulations (site consolidation, supply disruption) before execution

・Forward-looking planning: moves from planning dependent on historical data to forward-looking planning that tests scenarios and adjusts dynamically

[Lesson] Mature global PSI does not “respond after something happens” but “tests it digitally before it happens.” Visualizing through a control tower and verifying scenarios through a digital twin builds resilience against supply disruption.

6-5 The Ideal Operating Model (Organization, Process, Governance)

Integrating these cases, the ideal operating model for global PSI can be described as follows.

・Organization: two layers — a “Planning CoE (Center of Excellence)” responsible for global planning standards, methodology, and talent development, and regional execution teams. Standards are set centrally, execution is local — a clear division of roles

・Process: links monthly IBP (demand review → supply review → supply-demand reconciliation → management review) with weekly PSI updates and exception handling

・Governance: clarifies the single source of truth, common KPI definitions, and decision-making authority when supply-demand gaps arise

・Talent: defines the roles of demand planners and supply planners, eliminating dependence on specific individuals and passing on knowledge

Layer	Horizon	Primary Owner	Primary Output
Strategic planning	3-5 years	Management / SCM strategy	Network design, capacity investment
Tactical planning (S&OP/IBP)	12-18 months	Planning CoE / business units	Supply-demand agreement, inventory policy, financial alignment
Execution planning (PSI)	Weekly to daily	Regional execution teams	Production, allocation, and replenishment by site

6-6 Maturity Steps Toward the Ideal Form

What the cases show is that the ideal form is not achieved in a single leap. Maturity proceeds in the following order.

・Step 1: integrate planning and process into a single plan and clarify roles (Case ①)

・Step 2: layer planning into strategic, tactical, and execution levels, and establish a rhythm of monthly S&OP plus weekly execution (Case ②)

・Step 3: establish a Planning CoE, balancing global standards with regional execution

・Step 4: raise foresight and responsiveness through predictive analytics, control towers, and digital twins (Case ③)

⚠ Do not skip the order. Digital twins and AI are the final stage of sophistication and do not function if introduced without the foundation of a single plan, a regular rhythm, and a CoE. The case companies, too, first solidified plan integration and operational discipline before advancing to leading-edge technology.

Chapter 7 Case Studies of PSI/IBP at Global Companies (by Company and Industry)

PSI at global companies has now evolved beyond mere supply-demand coordination into IBP (Integrated Business Planning), which integrates management, finance, sales, and SCM. Although the name differs by company, the integrated process of “demand forecasting → supply planning → inventory optimization → management decision-making” is common. This chapter organizes named, leading-company case studies from the consumer goods, healthcare, chemical, automotive/manufacturing, and digital domains.

7-1 Five Success Factors Common to Leading PSI/IBP

Success Factor	Content
① A Single Plan (One Number Plan)	Sales, production, and finance use the same single plan
② Scenario Planning	Immediately compares alternatives when supply and demand fluctuate
③ Digital Twin / Control Tower	Visualizes global supply and demand in real time
④ AI Demand Forecasting (Demand Sensing)	Reflects POS and market signals in forecasts
⑤ Direct Link to Management Meetings	Monthly PSI/IBP becomes the venue for management decision-making

? The destination reached by leading companies is not “PSI = inventory management” but “IBP = an integrated management process.” KPIs, too, have shifted from being inventory-centered to profit, cash, service level, and ROIC.

7-2 Consumer Goods (FMCG) — The Leading Edge of Demand Sensing and IBP

P&G — Profit-Driven Business S&OP

・Demand-Driven Supply Network: rather than relying on demand forecasts, directly incorporates POS, retailer inventory, and promotional information, connecting the process from the occurrence of demand through to supply. Simultaneously achieves inventory reduction, fewer stockouts, and improved cash

・Business S&OP: the PSI meeting is not an SCM meeting but a “management meeting,” discussing revenue, profit, service level, and cash

・Lesson: this has become “profit-driven PSI” rather than sales-driven PSI

Unilever — Connected Supply Chain and Digital Twin

・End-to-end integration: connects Demand Creation, Planning, Fulfillment, and Customer Service in a single, unbroken flow

・Digital twin: reproduces the impact of supply-demand changes on production, inventory, and logistics, and instantly simulates the effects

・Control tower: manages not “what is happening” but “what is about to happen”

・Lesson: the goal is not making a plan but “making plan changes at high speed”

Nestlé — Demand Sensing and SKU Concentration

・Demand sensing: reflects POS and market data in short-term demand forecasts, increasing responsiveness

・Focused SKU management: prioritizes core SKUs and concentrates production during logistics crises

・Supply Chain Control Tower: centrally manages transportation across an entire region (predicted ETA, carrier performance, visibility)

・Lesson: the effect of the management decision to “narrow down SKUs” is greater than that of improving forecast accuracy

Coca-Cola — CDSP and Service-Level Management

・CDSP (Customer Demand & Supply Planning): operates demand planning, S&OP, supply planning, and inventory optimization on a single platform

・AI demand forecasting plus scenarios: incorporates temperature changes and promotions into machine-learning forecasts, responding immediately to sudden demand shifts

・Digital twin: uses virtual models of logistics and warehouses to examine capital investment and capacity

・Lesson: PSI is not inventory management but “PSI = service-level management”

7-3 Healthcare and Life Sciences — Johnson & Johnson (J&J)

J&J, which operates IBP amid extremely high uncertainty across pharmaceuticals, medical devices, and consumer goods, is the most valuable reference model for management-integrated PSI in manufacturing. The company is pursuing a company-wide transformation, “APT 2.0 (Advanced Planning Transformation 2.0),” with OMP as its partner.

・The 3A framework: Automate (moving to exception-based planning, automating routine work) / Anticipate (forecasting demand shifts with AI/ML) / Accelerate (speeding decision-making through cross-functional end-to-end alignment)

・Scenario-driven IBP: compares multiple options in IBP meetings, premised on demand surges, raw-material shortages, and production capacity shortfalls

・Inventory optimization: optimally manages global inventory through multi-echelon inventory optimization (MEIO)

・Use of AI: equips itself with AI assistants, self-service analytics, and a control tower

[Lesson] At J&J, “a PSI professional is positioned not as a plan creator but as a decision-support provider.” Alongside raising service levels and reducing repetitive work, planning is tightly aligned with commercial and financial goals.

7-4 Chemical — BASF / Dow

Even in the chemical industry, with its constraints of equipment, continuous/batch production, byproducts, and hazardous materials, data-driven supply-demand planning is advancing.

・BASF: optimizes commercial flows through its Smart Supply Chain (big data, IoT, advanced logistics). It raises overall transparency through data integration with customers and suppliers, and compresses inventory and storage costs through a JIT inventory strategy

・Dow: analyzes historical data, market trends, and customer behavior with AI algorithms to improve demand-forecast accuracy, and has reduced logistics costs by approximately 18% through AI-based load optimization

・Lesson: in chemicals, “balancing equipment utilization with inventory” is essential. Beyond sophisticated demand forecasting, data integration with partners and JIT/logistics optimization are effective

7-5 Automotive and Industrial Machinery — Toyota / P&W / Cisco / Samsung

Toyota — Real-Demand Linkage and Recovery Speed

・TPS (Toyota Production System): pull-based, just-in-time, linked to real demand. Prioritizes actual demand over the sales plan

・Anomaly detection: emphasizes “speed of response to supply-demand change” over forecast accuracy. Rapid recovery matters more than a perfect forecast

Pratt & Whitney (P&W) — Capacity-Planning-Centered IBP

・Long lead times: parts-procurement lead times for aircraft engines run 12-24 months; monthly supply-demand, quarterly, and 3-5-year production plans are managed simultaneously

・Risk-based PSI: manages capacity constraints, supplier risk, and inventory risk for critical parts within PSI

・Lesson: in manufacturing, it is important to “build capacity planning, not just a sales plan, into PSI”

Cisco / Samsung — Supplier Integration and Weekly Replanning

・Cisco: virtual manufacturing (largely outsourced) plus multi-tier visibility (through Tier 2 and Tier 3). Supplier management matters more than managing its own factories

・Samsung: an integrated SCM center oversees global sales, production, and logistics, and although largely on a monthly cycle, conducts “weekly” replanning of supply and demand. Simply increasing meeting frequency substantially improves accuracy

7-6 Digital Players — Amazon / Apple / Inditex (ZARA) / Dell

・Amazon: AI-driven supply and demand (forecasting by product, customer, and region) and network inventory (managing worldwide inventory as a whole, rather than by warehouse). The unit of PSI is not a factory or a country but the “supply chain network”

・Apple: reserves semiconductor, display, and EMS production capacity more than a year before launch. When supply is short, allocates not evenly but by profit and strategic importance (Supply Allocation) — “maximizing corporate value” rather than “customer fairness”

・Inditex (ZARA): runs a near-daily PSI, sending in-store sales information to headquarters immediately. Deliberately tolerates stockouts to minimize “leftover stock.” The purpose of PSI is not to hold inventory but to reduce leftover stock

・Dell: minimizes finished-goods inventory through build-to-order (production after receiving orders), positioning supplier inventory around its factories — “optimal allocation of inventory responsibility” rather than “inventory optimization”

7-7 Global PSI Maturity and Implications for Japanese Companies

Surveying the above, leading companies can be divided into three groups by the pattern of their strengths.

Group	Representative Companies	Strength
FMCG group	P&G, Unilever, Nestlé, Coca-Cola	Demand forecasting, demand sensing, IBP
Manufacturing group	P&W, Cisco, Samsung, Toyota	Capacity planning, supplier integration, risk management
Digital group	Amazon, Apple	AI, scenario planning, network optimization

To grasp the overall picture of next-generation global PSI from a manufacturing management perspective, it is effective to combine the patterns of the following five companies.

・J&J (management-integrated IBP): scenario-based management decision-making

・P&G (One Number Plan): a profit-driven single plan

・Apple (capacity assurance): long-term capacity reservation and strategic allocation

・Amazon (AI plus network inventory): network-level optimization

・Toyota (real-demand linkage): response speed over forecasting

? Today’s leading companies operate not “PSI” but a “management-integrated IBP,” and their KPIs, too, have shifted from inventory to profit, cash, service level, and ROIC. For Japanese companies, AI demand forecasting that leverages the asset of past deal and performance data, together with an organizational design that turns PSI into a “management meeting,” is a realistic and highly effective first step.

Chapter 8 Detailed Requirements for Global PSI (Functional and Non-Functional)

This chapter concretizes, from the perspective of functional and non-functional requirements, what must be defined when actually building and implementing global PSI. It is intended to be used as a checklist for requirements definition and RFP creation. Weighting will differ by industry, but the framework of requirements is common.

7-1 Data and Master Data Requirements — The Premise for Everything

The accuracy of global PSI depends entirely on the quality of input data. First, the following master data must be maintained globally at a unified level of granularity and coding.

Data Item	Primary Use	Key Points of Maintenance
Item master	Unit of supply-demand planning	Globally unified codes, units, and classification (eliminating different codes for the same item)
BOM (E/M-BOM)	Requirements explosion / parts planning	Identification of common parts; explicit marking of long-lead-time parts for capital equipment
Sites / hierarchy	Distribution requirements planning (DRP)	Hierarchy and sourcing relationships among plants, DCs, and channels
Lead time	Safety stock / advance procurement	Retains procurement/production/transportation lead times and their “variability,” based on actuals
Inventory	Allocation / visibility	Centrally manages current stock, in-transit stock, and allocated stock as distinct categories
Demand history	Foundation for forecasting	Accumulates shipments, orders, and POS at the granularity of the point of consumption
Supply constraints	A feasible plan	Holds capacity, MOQ, calendars, and tariffs as constraints

7-2 Demand Planning Requirements

・Forecasting methods: can select and combine statistical forecasting (moving average, seasonal decomposition), machine learning, and demand sensing (POS, weather, external data)

・Forecast hierarchy: forecasts across the multiple layers of SKU × site × period, supporting both top-down aggregation and bottom-up disaggregation

・Non-stationary demand: can handle promotions, new product introduction (NPI), and end-of-life (EOL) separately from normal forecasting

・Consensus: incorporates manual input from sales and marketing, supporting consensus building on demand (One Number)

・Accuracy management: continuously measures forecast accuracy using MAPE, bias, and FVA (Forecast Value Added)

7-3 Supply and Inventory Planning Requirements

・Supply planning: can produce a supply plan that considers capacity, material, and transportation constraints (constrained), and can perform supply response to demand changes

・Safety stock: can statistically calculate safety stock (Z × σ × √LT) by service level, differentiated by item and site

・Multi-echelon optimization (MEIO): can optimally position inventory across the whole network, reducing total inventory through risk pooling

・Distribution requirements planning (DRP): can plan allocation and replenishment from central to regional to customer

・Long-lead-time and parts planning: can perform advance procurement of long-lead-time parts and installed-base-based forecasting of maintenance parts for capital equipment

・Scenarios: can simulate and compare what-if situations such as supply-demand shifts, supply disruptions, and tariff changes

7-4 Execution and Integration Requirements

・ERP integration: can bidirectionally integrate orders, inventory, production, procurement, and shipment with the execution ERP (such as S/4HANA)

・Synchronization method: can select between real-time (event-driven) and batch according to business requirements

・Inventory allocation: can determine omnichannel ATP/aATP (shippability and allocation) in real time

・Exception management: detects stockout risk, excess, and delay, automatically alerting and feeding into replanning

7-5 Visualization and Analytics Requirements

・Control tower: centrally monitors global supply, demand, inventory, and transportation, and can visualize and notify of anomalies

・Dashboards: can drill down into and analyze KPIs (fill rate, OTIF, days of supply, forecast accuracy)

・Digital twin: replicates the network and supply-demand situation in a virtual space, enabling scenario verification before execution

7-6 Non-Functional and Governance Requirements

・Global support: supports multiple countries, currencies, languages, time zones, and calendars

・Performance: processes large-scale SKU × site × period calculations within a practical time frame (speed of plan execution)

・Security and audit: equipped with access control, data access management, and change history (audit trails)

・Master data governance (MDM): defines rules and responsibilities for registering and changing master data, continuously maintaining data quality

・Organization and process: defines, as operational requirements, the rhythm of monthly S&OP/IBP plus weekly execution, and the division of roles between the Planning CoE and regional execution

? A pitfall in requirements definition is to focus heavily on “functional requirements (what it can do)” while neglecting “data, non-functional, and operational requirements (how it is run).” Global PSI is 80% determined by master data quality and operational process. Data maintenance and governance should be turned into requirements with the same intensity as the list of functions.

Chapter 9 Conclusion

The purpose of global PSI is to place inventory “in the necessary location, in only the necessary amount” under the uncertainty of supply and demand spanning multiple countries and multiple sites, thereby suppressing stockouts and excess at the same time. The forefront of this effort is consumer goods, where methods such as demand sensing, omnichannel inventory, and MEIO have been refined.

Supporting these are the three layers — the execution ERP (S/4HANA), the planning layer (IBP), and the control tower — together with the master data that runs through them all. The functions to be prioritized differ by industry, but the principle of “capturing real demand, anticipating supply and demand, and positioning it optimally” is common. The core system should be designed in three layers so as to be able to execute this principle, and should be unified through master data.

? Final message: global PSI is not something achieved simply by “putting in a system.” It functions only once the execution and planning core systems mesh together, on a foundation of real-demand data, a supply-demand agreement process, and well-maintained master data. Leading cases in consumer goods show that this integration itself is the source of competitive advantage.

(This paper was prepared based on general methodology for global SCM/PSI, practical trends in consumer goods (CPG/FMCG), core-system functionality such as SAP S/4HANA and SAP IBP, and publicly available literature from 2025-2026. The figures presented are representative estimates; actual effects will vary by industry and product characteristics.)