Logistics and Fulfilment Execution Systems

Introduction

Logistics and fulfilment execution systems convert supply-chain plans into physical and financial actions. They determine what should be picked, packed, shipped, routed, tracked, delivered, returned, and paid for. Whereas enterprise resource planning (ERP) and advanced planning systems establish broader commercial and planning requirements, execution systems coordinate the operational detail required to move individual inventory units, orders, shipments, vehicles, and parcels.

The principal systems considered in this chapter are the warehouse management system (WMS), transportation management system (TMS), order management system (OMS), yard management system (YMS), fleet management system, parcel-management platform, and last-mile delivery system. These applications are increasingly connected through application programming interfaces (APIs), event streams, mobile devices, sensors, electronic proof of delivery, and artificial-intelligence-supported decision tools.

The managerial challenge is not simply to purchase the system with the greatest number of features. It is to design an execution architecture in which decisions are made at the correct level, data remains synchronized, exceptions are visible, and operational controls support customer service, cost, compliance, and resilience. A technically sophisticated system can still fail if master data, process ownership, integration, workforce capability, or governance is weak.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Distinguish the roles and boundaries of WMS, TMS, OMS, yard, fleet, parcel, and last-mile systems.
2. Explain the principal warehouse, transportation, delivery, settlement, and returns workflows.
3. Evaluate execution systems using service, cost, control, data, integration, and scalability criteria.
4. Apply basic logistics-performance measures and interpret their managerial significance.
5. Analyze an omnichannel fulfilment problem and recommend an integrated solution.
6. Identify governance, cybersecurity, privacy, safety, regulatory, and ethical risks in logistics software.

1. The Execution-System Landscape

A useful distinction is between the system of record, the system of decision, and the system of execution. An ERP may be the system of record for financial transactions and product masters. An OMS may be the system of decision for order allocation and fulfilment promises. A WMS, TMS, or delivery platform is primarily a system of execution: it sequences and controls operational work.

These categories overlap. For example, an OMS may decide that an order is fulfilled from a particular facility, but the WMS decides the precise pick task and the TMS selects a carrier service. Effective architecture therefore requires explicit ownership of each data object and decision.

The architecture should be designed around physical and informational events. Examples include order released, inventory reserved, pick completed, carton closed, shipment tendered, trailer arrived, parcel scanned, delivery attempted, delivered, and return received. Event-based integration reduces dependence on batch files, but it also makes event definitions, sequencing, duplication, and failure recovery critical design issues.

2. Warehouse Management Systems

2.1 Core warehouse workflow

A WMS typically supports the following flow:

Inbound appointment → receiving → inspection and discrepancy handling → put-away → replenishment → allocation → picking → packing → staging → shipping → inventory reconciliation.

At receiving, the system compares the expected advance shipping notice or purchase order with the physical receipt. It may capture serial numbers, lot numbers, expiry dates, country of origin, temperature data, or quality status. Put-away rules then select a suitable location according to product dimensions, compatibility, velocity, handling constraints, and available capacity.

Inventory in a WMS is more detailed than a simple quantity balance. It may be represented by SKU, location, lot, serial number, status, handling unit, owner, and inventory condition. A useful conceptual identity is:

This relationship assumes that all quantities use the same unit of measure and that allocations and holds are not already excluded from on-hand inventory. In practice, systems may distinguish available, damaged, quarantined, blocked, consigned, and in-transit inventory.

2.2 Allocation and picking

Allocation links customer demand to inventory. Rules may prioritize promised delivery date, customer class, shelf life, lot sequence, distance from the packing area, or order consolidation. Common picking methods include discrete picking, batch picking, zone picking, cluster picking, wave picking, and goods-to-person automation.

The choice is contingent on order profile. Batch picking can reduce travel for many small orders but requires accurate sortation. Zone picking can improve specialization and throughput but introduces handoffs. Wave planning synchronizes work with carrier cut-off times, labour availability, and dock capacity.

A basic warehouse productivity measure is:

The unit is picks per labour hour. The measure should specify whether a “pick” means a line, unit, case, or handling unit. Comparing facilities without a common definition can produce misleading conclusions. Productivity should also be interpreted with error rate, safety, overtime, congestion, and order complexity; increasing speed while increasing errors is not necessarily an improvement.

2.3 Packing, shipping, and inventory accuracy

Packing systems select cartons, dunnage, labels, documents, and carrier services. They may validate the contents by barcode scan, weight, dimension, or image recognition. Shipping confirmation should create an auditable link between the customer order, inventory decrement, parcel or freight identifier, and carrier handoff.

Inventory accuracy is commonly expressed as:

The formula assumes that each audited position is classified as correct or incorrect according to a defined tolerance. A unit-based variant may compare recorded and counted quantities, but the organization must define whether a small quantity variance counts as an error.

Cycle counting is generally superior to relying exclusively on an annual physical count. Risk-based counting assigns greater frequency to high-value, high-velocity, regulated, or error-prone items. The WMS should record count causes, adjustments, authorizations, and investigation outcomes rather than treating adjustments as routine data cleaning.

2.4 Warehouse automation and human work

Automation may include conveyors, automated storage and retrieval systems, autonomous mobile robots, sorters, robotic picking, voice systems, and computer vision. The WMS remains essential because automation requires task orchestration, inventory identity, exception management, and integration with controls systems.

Automation is not automatically economical. A sound evaluation includes demand volume, profile variability, peak capacity, maintenance, system availability, labour alternatives, safety, integration cost, and exit options. Human-centred design is equally important: poor interfaces, excessive walking, unrealistic engineered standards, or opaque productivity monitoring can reduce safety and retention.

3. Order Management and Omnichannel Fulfilment

An OMS coordinates demand across channels such as stores, web marketplaces, mobile applications, call centres, and business-to-business orders. Its central functions include order capture, payment status, inventory visibility, sourcing, promise calculation, split-order management, customer communication, cancellation, and return orchestration.

3.1 Available-to-promise and order promising

A simple available-to-promise calculation is:

ATP is meaningful only when timing, location, inventory status, reservations, and unit of measure are consistent. More advanced order promising considers future receipts, production capacity, transport lead time, safety stock, fulfilment rules, and alternative locations.

The OMS may use a sourcing hierarchy, such as:

1. fulfil from the customer’s preferred or nearest eligible node;
2. protect scarce inventory for higher-priority demand;
3. minimize split shipments;
4. respect delivery promise and carrier cut-off;
5. minimize total fulfilment cost subject to service constraints.

These objectives can conflict. Shipping from the nearest facility may reduce delivery time but increase transportation cost if the order is fragmented. Store fulfilment may increase geographic coverage but disrupt store labour and inventory accuracy. A mature OMS makes trade-offs explicit instead of embedding them in undocumented rules.

3.2 Omnichannel operating models

Important models include ship-from-warehouse, ship-from-store, buy online and pick up in store, curbside collection, endless aisle ordering, drop shipment, marketplace fulfilment, and store returns of online purchases. Each model changes inventory ownership, picking responsibility, customer promise, tax treatment, and reverse-logistics requirements.

The design question is not whether every channel should access every node. It is whether the organization can provide reliable inventory availability and consistent customer treatment across channels. “Real-time inventory” is valuable only if transactions are captured promptly and inventory states are trustworthy.

4. Transportation Management Systems

A TMS supports planning, procurement, execution, visibility, settlement, and performance analysis. Its scope may cover parcel, less-than-truckload, full truckload, intermodal, ocean, air, and private fleet movements.

4.1 Transportation planning

Planning begins with shipment requirements: origin, destination, weight, cube, handling constraints, delivery window, equipment type, and service level. The TMS may consolidate compatible shipments, select modes, build loads, assign carriers, and sequence stops.

A simplified utilization measure is:

The measure assumes compatible units and a defensible estimate of usable volume. It does not indicate whether weight, axle limits, product incompatibility, loading sequence, or legal restrictions make the load feasible. Weight utilization should be calculated separately.

Routing decisions are constrained optimization problems. A vehicle-routing formulation may seek to minimize total cost while satisfying vehicle capacity, customer time windows, route duration, driver rules, and service requirements. Exact optimization becomes computationally difficult as the number of stops and constraints increases, so commercial systems often combine mathematical programming, heuristics, historical patterns, and human intervention.

4.2 Tendering and execution

After planning, the TMS tenders freight to selected carriers through portals, APIs, electronic data interchange, email, or transportation exchanges. It records acceptance, rejection, appointment status, pickup, in-transit events, exceptions, and delivery confirmation.

Tendering logic should include fallback carriers and escalation thresholds. A plan that depends on one carrier accepting every load is operationally fragile. The system should also distinguish a planned event from a verified event. A scheduled pickup is not evidence that freight was collected.

4.3 Freight settlement

Freight settlement compares expected charges with carrier invoices. A freight invoice may include line-haul, fuel surcharge, accessorials, detention, lumper charges, taxes, and other adjustments. Automated audit rules identify duplicate invoices, incorrect rates, unauthorized accessorials, mismatched weights, and service failures.

A simple freight-cost-per-unit measure is:

The numerator and denominator must use the same scope and period. If returns, accessorials, or interfacility transfers are excluded from one side but included in the other, the indicator is not comparable.

A rate engine should preserve effective dates, lanes, zones, equipment, minimum charges, currency, fuel-index rules, accessorial conditions, and contract version. Settlement is therefore both a financial-control process and a data-governance process.

5. Yard and Fleet Execution

5.1 Yard management systems

The yard is a buffer between transportation and warehouse operations. A YMS manages gate appointments, driver check-in, trailer identification, spot moves, dock assignment, dwell time, seal status, and trailer inventory. Without yard visibility, a WMS may show a shipment ready while the correct trailer is inaccessible, or a TMS may show a carrier arrived while the facility cannot locate the vehicle.

Yard performance commonly includes trailer dwell time, dock-to-stock time, appointment adherence, gate processing time, and percentage of moves completed without manual escalation. These metrics require precise event timestamps and consistent definitions. For example, dwell may mean gate-in to gate-out, arrival to dock departure, or loaded arrival to release; each definition answers a different question.

5.2 Fleet management systems

Fleet platforms support vehicle assignment, driver work, fuel, maintenance, telematics, safety, compliance, and utilization. Telematics can supply location, mileage, harsh-braking events, engine diagnostics, temperature, and geofenced arrival or departure.

Preventive maintenance can be triggered by time, mileage, engine hours, condition, or manufacturer requirements. The system should not encourage unsafe extension of maintenance intervals merely to improve asset utilization. Fleet management also requires attention to driver privacy, collective agreements, fatigue, and the distinction between safety coaching and punitive surveillance.

6. Parcel and Last-Mile Systems

Parcel systems address high-volume, small-shipment execution. They commonly support carrier selection, rate shopping, label generation, manifesting, customs documentation, tracking-number assignment, and delivery-event normalization. A parcel label is not merely a document; it is an identifier connecting the physical package to the order, carrier, service, and tracking history.

Last-mile systems plan routes, assign stops, manage delivery windows, provide driver navigation, capture proof of delivery, and process exceptions such as customer unavailable, address failure, damage, or refusal. Delivery planning must consider vehicle capacity, geographic restrictions, time windows, stop service time, driver hours, traffic, and promised delivery dates.

A useful service measure is:

The organization must define the promised window, treatment of customer-requested changes, time-zone rules, failed attempts, and orders lacking a valid promise. A high on-time percentage calculated only after excluding difficult deliveries may be operationally meaningless.

Proof of delivery may include signature, photograph, geolocation, timestamp, recipient identity, or locker confirmation. Each method has evidentiary, privacy, and accessibility implications. Photographs should avoid unnecessary capture of people, private interiors, or sensitive information.

7. Returns and Reverse Logistics

Returns begin with authorization and continue through transportation, receipt, inspection, disposition, refund, replacement, repair, resale, recycling, or disposal. An OMS or dedicated returns-management system may create a return merchandise authorization, while the WMS receives and evaluates the item.

A robust process distinguishes:

• customer remorse from product defect;
• sellable, open-box, repairable, quarantined, recyclable, and waste inventory;
• refund eligibility from physical receipt;
• carrier damage from manufacturing failure; and
• fraud risk from legitimate service recovery.

Returns should not be measured only as a cost. They provide product-quality signals, customer-experience evidence, and demand information. However, generous return policies can create transport emissions, handling cost, inventory ageing, and fraud exposure. The correct policy depends on product economics, customer expectations, regulation, and environmental objectives.

8. Integration, Data, and Control Architecture

Execution systems exchange master data, transactional data, and event data. Critical master data includes SKU dimensions, weights, units of measure, location attributes, carrier services, rates, calendars, customer addresses, hazardous-material classifications, and handling requirements. Errors in dimensions or addresses can propagate into cartonization, rating, routing, labour planning, and customer promises.

Integration should address five questions:

1. Who owns the data?
2. What event or transaction triggers the exchange?
3. What is the required latency?
4. How are duplicates, omissions, and out-of-order messages handled?
5. How can the transaction be reconciled and repaired?

A reliable interface requires idempotency: replaying the same valid message should not create duplicate inventory movements, shipments, or charges. It also requires monitoring for message failure, schema changes, latency, and reconciliation differences.

Control design should include role-based access, segregation of duties, approval thresholds, audit trails, encryption, backup, disaster recovery, vendor continuity, and incident response. A user who can alter carrier rates and approve freight invoices presents a different risk from a user who can only print labels. Cybersecurity is operational resilience because ransomware or identity compromise can stop physical movement.

9. Evaluation Framework for Selecting Execution Systems

A graduate-level evaluation should combine functional fit with total operating consequences. Suggested dimensions are:

A business case should not rely on licence price alone. Benefits may arise from fewer touches, reduced travel, improved load utilization, lower freight leakage, fewer errors, better inventory availability, faster returns, and improved customer retention. Benefits should be linked to measurable baselines and implementation assumptions.

10. Artificial Intelligence and Contemporary Developments

AI and advanced analytics are increasingly used for demand prediction, slotting, labour forecasting, ETA prediction, address correction, image-based damage detection, route adjustment, fraud screening, and conversational support. Their value depends on data quality and operational integration. A highly accurate ETA model that is not used in customer communication or dispatch decisions produces little benefit.

Important governance questions include:

• Is the model’s output explainable to planners and customers?
• Does it disadvantage particular regions, customers, drivers, or workers?
• Can a human override it, and is the override recorded?
• What happens when demand patterns shift or data is missing?
• Are personal, location, biometric, or image data collected lawfully and proportionately?

Digital twins, control towers, event-driven architecture, robotics-as-a-service, locker networks, electric delivery vehicles, and crowdsourced delivery are also reshaping execution. These developments create new dependencies: charging infrastructure, battery range, cyber-physical security, data-sharing agreements, and responsible work design.

11. Illustrative Case: Northstar Home Goods

Background

Northstar Home Goods is a fictional omnichannel retailer selling small appliances and home accessories through stores, a website, and a marketplace. It operates two distribution centres and eighty stores. Online orders are promised within two days in major cities. The company uses an ERP, a basic WMS, separate carrier portals, and spreadsheets for store fulfilment.

Problem

On-time delivery is 86%, inventory accuracy is 93%, and online returns take an average of twelve days to reach disposition. Customers frequently order items shown as available but later receive cancellation notices. Distribution-centre trailers wait in the yard, while stores experience unexpected picking work. Freight invoices contain frequent accessorial disputes.

Analysis

The central problem is not one defective application. It is the absence of a coordinated execution model. The ERP records inventory, but store and distribution-centre updates are delayed. The OMS function is incomplete, so promises are made without reliable node-level availability. The TMS function is fragmented across carrier portals, and no YMS records trailer location or dock status. Returns are handled as customer-service cases rather than as a controlled reverse flow.

Approach

Northstar establishes an OMS as the orchestration layer and retains the WMS as the warehouse execution system. It introduces store inventory rules that exclude display units and damaged stock, then connects inventory events through APIs. A TMS is implemented for parcel and regional freight, with a rate engine and invoice audit. A lightweight YMS manages appointments and trailer status. The returns process creates authorization, carrier label, receipt scan, inspection disposition, and refund events.

The company defines a common event dictionary and performance dashboard. Measures include promise accuracy, on-time delivery, inventory accuracy, split-order rate, freight cost per order, return cycle time, and percentage of invoices automatically matched. A pilot begins with one distribution centre, ten stores, and one metropolitan delivery region.

Outcome

After the pilot, Northstar reports improved inventory confidence and fewer cancellations. Store fulfilment is restricted to eligible products and operating periods, reducing disruption. Carrier invoices are matched against contracted rates before payment. The company does not eliminate all exceptions; instead, it makes them visible and assigns ownership. The results are illustrative rather than empirical forecasts, and the case does not claim a universal performance improvement.

Learning points

1. Omnichannel capability depends on inventory discipline, not only on an OMS licence.
2. Execution visibility requires common event definitions and reliable timestamps.
3. Financial settlement should be connected to operational evidence.
4. Pilot scope should include process, data, integration, training, and governance.
5. Exception ownership is as important as automation.

12. Challenges and Limitations

Execution systems can create false confidence when dashboards display precise but inaccurate data. Integration can reproduce errors faster, and automation can make an incorrect rule operate at scale. Customization may preserve obsolete processes and increase upgrade risk. Cloud systems reduce some infrastructure burdens but create dependency on connectivity, vendor availability, contractual service levels, and data portability.

There are also organizational limitations. Warehouse, transport, merchandising, finance, store operations, and customer service may optimize different objectives. A warehouse may prefer large waves, while transportation needs early shipment visibility. Customer service may promise exceptions that undermine route economics. Governance must therefore establish cross-functional decision rights and escalation paths.

Ethical and regulatory issues include worker surveillance, algorithmic management, discrimination in service allocation, accessibility for customers, safety, hazardous-material compliance, customs documentation, consumer privacy, retention of delivery images, and environmental reporting. Legal requirements vary by jurisdiction; organizations must obtain specialist advice rather than assuming that a software configuration guarantees compliance.

Best-Practice Principles

1. Design the operating model before configuring software.
2. Assign explicit ownership to every major data object and decision.
3. Standardize event definitions, units of measure, time zones, and exception codes.
4. Protect customer promises with feasible inventory and transport logic.
5. Measure service, cost, quality, safety, sustainability, and resilience together.
6. Build reconciliation, auditability, and recovery into every integration.
7. Pilot with representative peak conditions rather than only average demand.
8. Preserve human authority for safety-critical and high-impact decisions.
9. Treat returns and settlement as core execution processes.
10. Review algorithms, vendors, access rights, and data practices continuously.

Chapter Summary

Logistics and fulfilment execution systems coordinate the movement of inventory, orders, vehicles, parcels, information, and money. The WMS controls warehouse tasks and inventory states; the OMS orchestrates order sourcing and customer promises; the TMS plans, tenders, tracks, and settles transportation; the YMS coordinates trailers and docks; fleet systems manage owned assets and drivers; parcel systems manage small-package execution; and last-mile systems control final delivery.

Effective execution depends on more than functional software. It requires accurate master data, event-driven integration, explicit decision rights, exception management, reliable metrics, cybersecurity, workforce capability, and governance. Omnichannel fulfilment increases the need for synchronized inventory and transparent trade-offs among service, cost, speed, and sustainability. AI and automation can improve decisions, but only when their data, assumptions, accountability, and human controls are sound. The appropriate evaluation question is therefore not “Which system has the most features?” but “Which integrated operating architecture can execute the organization’s promises safely, economically, transparently, and resiliently?”

Key Terms

WMS: A system that controls warehouse inventory, locations, tasks, and workflows.

TMS: A system that plans, executes, tracks, and settles transportation.

OMS: A system that captures, orchestrates, allocates, and manages customer orders.

YMS: A system that coordinates trailers, gates, appointments, and dock activity.

Fleet management system: Software for managing owned or dedicated vehicles, drivers, maintenance, and telematics.

Parcel-management system: Software for rating, labelling, manifesting, and tracking small shipments.

Last-mile system: Software for planning and executing final delivery to the customer.

Available-to-promise: The supply quantity that can be committed to future demand under defined timing and inventory rules.

Wave picking: The coordinated release of warehouse work according to time, carrier, labour, or operational constraints.

Slotting: The assignment of products to warehouse locations to support efficient handling.

Yard dwell time: The elapsed time an asset remains in a defined yard process or area.

Proof of delivery: Evidence that a shipment reached the intended recipient or location.

Freight settlement: The process of validating and paying transportation charges.

Reverse logistics: The movement and disposition of returned, repaired, recyclable, or end-of-life products.

Event-driven integration: Integration in which systems exchange messages when defined business events occur.

Idempotency: The property that repeating the same valid transaction does not create an additional unintended effect.

Omnichannel fulfilment: Coordinated fulfilment across multiple sales and service channels.

Exception management: The identification, prioritization, assignment, and resolution of deviations from plan.

Control tower: An integrated visibility and decision-support capability spanning supply-chain activities.

Review Questions

5 Short Answer Questions

1. What is the primary distinction between an OMS and a WMS?
2. Identify three activities commonly supported by a TMS.
3. Why must cube utilization and weight utilization be measured separately?
4. What is the purpose of a yard management system?
5. Define idempotency in an execution-system interface.

5 Long Answer Questions

1. Explain how a WMS supports the process from receiving through shipping, including the role of inventory status and task sequencing.
2. Compare the strategic and operational roles of OMS, TMS, and WMS in omnichannel fulfilment.
3. Discuss the benefits and limitations of automation in warehouse and last-mile operations.
4. Explain why freight settlement is both a financial-control process and an operational-data process.
5. Evaluate the governance risks associated with AI-supported routing, delivery prediction, or workforce management.

3 Analytical / Application Questions

1. A retailer has accurate distribution-centre inventory but delayed store updates. Customers are promised two-day delivery from the nearest eligible node. Analyze the likely failure mechanisms and recommend an OMS, inventory, and integration design.
2. A carrier’s invoices show rising detention and accessorial charges, while trailers frequently wait outside the distribution centre. Develop an investigation using TMS, YMS, appointment, dock, and settlement data.
3. Design a pilot for an omnichannel fulfilment transformation involving one warehouse, ten stores, parcel carriers, and returns. Specify process scope, data standards, performance measures, governance controls, and criteria for scaling.

Multiple Choice Questions with Answers

Question 1

Which system is primarily responsible for sequencing warehouse picking and put-away tasks?

A. Warehouse management system

B. Transportation management system

C. Order management system

D. Yard management system

Correct Answer: A. Warehouse management system

Question 2

A retailer wants to decide whether an online order should be fulfilled from a store or distribution centre. Which capability is most directly involved?

A. Carrier invoice auditing

B. Order orchestration and sourcing

C. Trailer spot-move control

D. Vehicle preventive maintenance

Correct Answer: B. Order orchestration and sourcing

Question 3

If on-hand inventory is 1,000 units, allocated inventory is 250 units, and inventory on hold is 50 units, what is available inventory under the chapter’s definition?

A. 700 units

B. 750 units

C. 800 units

D. 1,300 units

Correct Answer: A. 700 units

Question 4

Which measure best evaluates whether deliveries met the customer’s promised time window?

A. Cube utilization

B. Inventory accuracy

C. On-time delivery rate

D. Pick productivity

Correct Answer: C. On-time delivery rate

Question 5

What is the principal purpose of a yard management system?

A. Forecasting seasonal demand

B. Managing trailers, gates, appointments, and docks

C. Calculating product selling prices

D. Inspecting returned merchandise

Correct Answer: B. Managing trailers, gates, appointments, and docks

Question 6

Why should a transportation plan measure cube utilization separately from weight utilization?

A. Volume and weight are different physical constraints

B. Cube utilization measures customer satisfaction directly

C. Weight utilization applies only to parcel shipments

D. Cube utilization eliminates the need to check legal restrictions

Correct Answer: A. Volume and weight are different physical constraints

Question 7

Which feature prevents replaying the same valid interface message from creating a duplicate shipment?

A. Slotting

B. Idempotency

C. Wave picking

D. Geofencing

Correct Answer: B. Idempotency

Question 8

Which activity is part of freight settlement?

A. Comparing carrier invoices with contracted rates

B. Assigning products to warehouse slots

C. Determining customer payment credit

D. Creating a driver navigation map only

Correct Answer: A. Comparing carrier invoices with contracted rates

Question 9

Which situation is an example of reverse logistics?

A. Tendering a new outbound load

B. Moving a returned product for inspection and disposition

C. Assigning a trailer to an inbound dock

D. Calculating a vehicle’s fuel consumption

Correct Answer: B. Moving a returned product for inspection and disposition

Question 10

A system reports 950 deliveries within the promised window out of 1,000 eligible deliveries. What is the on-time delivery rate?

A. 5%

B. 50%

C. 95%

D. 105%

Correct Answer: C. 95%