Supply Chain Frequently Asked Questions
Answers to the most common questions about supply chain management, aimed at developers building AI/RAG applications in the supply chain domain and professionals looking to deepen their understanding of key concepts.
What is the difference between supply chain management and logistics management?
Logistics covers the physical movement and storage of goods from one point to another - transport, warehousing, distribution. Supply chain management has a broader scope: it includes logistics but also procurement, demand planning, production, supplier relationships and returns management. In short, logistics is a subset of SCM. This distinction matters for AI system design: a supply chain model must understand a broader domain than just transport optimisation.
How do you calculate safety stock?
The most common statistical formula is: SS = z × σLT, where z is the z-factor corresponding to the desired service level (e.g. 1.65 for 95%) and σLT is the standard deviation of demand during the lead time period. When both demand and lead time are variable: SS = z × √(LTavg × σD² + Davg² × σLT²). For AI applications generating stock recommendations, it is essential to handle historical demand data at sufficient granularity (daily or weekly).
What is the bullwhip effect and how is it mitigated?
The bullwhip effect is the progressive amplification of demand variability as one moves upstream along the supply chain from retailer to manufacturer. A 10% increase in retail sales can translate into order swings of 40% or more at the manufacturer level. Key causes: local forecasting based on orders (not POS), lot-sizing policies, promotions and lack of information sharing. Mitigation approaches include: sharing POS data, frequent small orders, CPFR and VMI systems.
When should you use MRP versus EOQ?
MRP is appropriate for dependent demand items (components whose demand derives from planned finished goods production): it calculates requirements based on the product structure (BOM) and MPS. EOQ is better suited for independent demand items (finished goods, spare parts) with relatively stable demand: it optimises lot size by balancing ordering and holding costs. In many ERP systems, MRP manages components while finished goods use ROP/EOQ-based policies.
What changed between Incoterms 2010 and Incoterms 2020?
The main change in Incoterms 2020 (effective 1 January 2020) is the replacement of DAT (Delivered at Terminal) with DPU (Delivered at Place Unloaded). DPU broadens the scope: the delivery place can be anywhere, not just a terminal. Another change: the minimum insurance level for CIP was upgraded to Clause A ("all risks" coverage), while CIF retains Clause C. In document systems and supply chain chatbots, correctly handling this terminology change is critical to avoid confusion with contracts still referencing DAT.
How does the S&OP process work?
The monthly S&OP cycle typically has five stages: (1) data collection and cleansing; (2) demand review - the commercial team updates the statistical forecast with market inputs; (3) supply review - the operations team checks production and logistics capacities and identifies constraints; (4) pre-S&OP - resolution of demand-supply gaps and preparation of alternatives; (5) executive meeting - selection of a single agreed company plan among scenarios. An AI assistant for S&OP typically supports stages 1-2 with predictive analytics and stages 3-4 with constraint optimisation.
What is the difference between make-to-stock, make-to-order and assemble-to-order?
In Make-to-Stock (MTS) finished goods are produced ahead of demand forecasts and held in inventory; the customer waits only for delivery time. In Make-to-Order (MTO) production starts only after a customer order is received; inventories are of raw materials and WIP. Assemble-to-Order (ATO) is a hybrid: components and sub-assemblies are built to forecast but final assembly occurs on the specific order, enabling high customisation with short lead times. The choice among MTS/MTO/ATO depends on the positioning of the Customer Order Decoupling Point (CODP).
What are procurement lead times and how do they affect inventory?
Procurement lead time is the total elapsed time from purchase order issuance to goods receipt in the warehouse. It includes order transmission time, supplier processing time, production time (if applicable), transport time and receiving/inspection time. Long or variable lead times require higher safety stocks and greater capital lock-up. Reducing and stabilising lead time is one of the highest-impact levers for inventory reduction.
How do you measure supplier performance?
Typical supplier evaluation considers four dimensions: quality (defect rate, non-conformances, PPM), delivery (OTIF, adherence to confirmed date), flexibility (ability to absorb volume or mix changes) and cost (TCO, price trends, transparency). A Supplier Scorecard consolidates these KPIs into a weighted score. For strategic suppliers (critical or leverage in the Kraljic matrix), financial, geopolitical and continuity risk indicators are added.
What does "end-to-end visibility" mean in supply chain?
End-to-end visibility is the ability to monitor in real time the status of orders, shipments, inventory and production processes across all nodes of the supply chain - from tier-2 supplier (supplier's supplier) to end customer. It requires data integration between ERP, WMS, TMS and partner systems (EDI, API). Benefits include faster response to disruptions, prevention of stockouts and optimisation of in-transit inventory. For RAG systems, end-to-end visibility produces structured and semi-structured data that updates at high frequency.
How does warehouse management differ from inventory management?
Warehouse management (WMS) handles the physical operations inside a warehouse: receiving, putaway, picking, packing, shipping and slotting. Inventory management handles stock levels and policies at a strategic and tactical level: how much to order, when to order, where to stock. The two systems overlap in tracking physical stock positions (WMS counts units) but have different objectives: WMS optimises operational productivity, inventory management optimises working capital and service level.
What is the inventory holding cost?
Inventory holding cost (or carrying cost) is the total cost of holding one unit of stock for one year. It includes: cost of tied-up capital (typically 8-15% of value), warehousing cost (space, energy, labour), obsolescence and deterioration risk, insurance and taxes. On average, holding costs range from 20% to 40% of inventory value per year. Ignoring holding costs leads to sub-optimal lot sizing.
How does cross-docking work and when is it beneficial?
In cross-docking, incoming goods from a supplier are transferred directly - without storage - to the shipping area, for specific destinations. The product passes through the DC for hours, not days. It is beneficial when: volumes are high and consistent, products are perishable or fast-moving, delivery windows are tight and demand is predictable. It requires precise synchronisation between inbound and outbound and real-time visibility.
What are Incoterms and what are they used for?
Incoterms (International Commercial Terms) are commercial terms standardised by the International Chamber of Commerce (ICC) that define, in an international sale contract, who bears: transport, insurance, customs clearance, and risk of loss or damage at each stage of transport. They do not govern transfer of ownership or payment terms, which remain in the sales contract. The current version is Incoterms 2020, effective 1 January 2020.
How do you implement a demand sensing system?
Demand sensing is implemented by replacing aggregate statistical forecasts with high-frequency demand signals: daily POS (Point of Sale) data, distributor inventory levels, e-commerce browsing data. The process includes: data ingestion from multiple sources (EDI, retailer API, IoT sensors), short-term forecasting model (1-30 days) typically ML-based (Random Forest, LSTM), integration with the planning system to automatically update replenishment plans. The horizon is too short for classical MRP; demand sensing operates at the execution level.
What is the difference between inbound and outbound logistics?
Inbound logistics manages material flows from suppliers to the company: transport, receiving, inspection and storage of raw materials and components. Outbound logistics manages flows from the company to customers: picking, packing, shipment and distribution of finished goods. The two have different cost drivers and KPIs: cost per tonne/km and replenishment punctuality are typical of inbound; cost per delivery and OTIF are typical of outbound.