Fleet Execution, Control and Service Performance

Introduction and context

Fleet operations convert a transport plan into physical movement, customer service, and financial results. Planning determines which vehicles and drivers should perform which work; execution determines whether that work is performed safely, legally, economically, and as promised. The execution environment is dynamic: traffic conditions change, loading takes longer than expected, customers alter delivery windows, vehicles fail, and documentation may be incomplete. Consequently, fleet management is not merely dispatching vehicles. It is a continuous control problem involving sensing, decision-making, intervention, communication, and learning.

A mature operation combines standard operating procedures (SOPs), a transport management system (TMS), vehicle and driver data, a control tower, exception management, and customer-service routines. Its purpose is not to eliminate all variation—an unrealistic objective—but to detect material deviations early and recover them at the lowest total cost while protecting safety, compliance, and customer trust.

Learning objectives

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

1. Explain the daily transport-execution cycle from release to settlement.
2. Design SOPs for dispatch, loading, delivery, proof of delivery (POD), and exception handling.
3. Evaluate the role and governance of a fleet control tower.
4. Calculate and interpret utilization, empty mileage, on-time delivery, and cost per stop.
5. Distinguish leading, lagging, and balancing performance indicators.
6. Apply structured communication and escalation methods during disruptions.
7. Assess technology, ethical, regulatory, and managerial issues in fleet control.

1. The execution system

Fleet execution is a closed-loop system. A plan creates expected events; operational data reveals actual events; controllers compare the two; interventions correct significant deviations; and performance data improves future plans. The loop has six stages:

1. Release: orders, routes, vehicle requirements, and time windows are confirmed.
2. Prepare: vehicles, drivers, equipment, documents, and cargo are checked.
3. Dispatch: the route is started with a clear manifest and communication protocol.
4. Monitor: milestones, location, delay risk, temperature, fuel, and safety events are observed.
5. Recover and communicate: exceptions are assessed, actions assigned, and stakeholders informed.
6. Close and learn: POD, empty returns, costs, incidents, customer feedback, and root causes are recorded.

The system should distinguish three types of information. Planned data describes what should happen; event data records what has happened; and decision data supports an intervention. Confusing these categories creates false precision—for example, treating a GPS position as proof that a customer accepted the shipment.

1.1 Daily operating rhythm

A disciplined operating rhythm normally includes:

• a pre-shift readiness review;
• a dispatch meeting or digital release;
• milestone monitoring during the shift;
• scheduled exception reviews;
• end-of-shift reconciliation; and
• a daily or weekly performance review.

The review should be based on exceptions rather than a lengthy discussion of every trip. A useful escalation rule specifies the threshold, owner, response time, and communication audience. For example, a projected delivery delay exceeding 30 minutes may require the dispatcher to contact the customer, while a safety incident requires immediate escalation to the safety manager regardless of delay duration.

2. Standard operating procedures as control architecture

An SOP is a controlled description of the sequence, responsibility, decision rights, records, and acceptance criteria for a recurring activity. It is more than a checklist. A good SOP reduces process variation while allowing authorised discretion when conditions change.

Each SOP should identify:

• purpose and scope: what activity and shipment types are covered;
• roles: dispatcher, driver, warehouse, customer-service agent, control-tower analyst, and manager;
• inputs: orders, route, vehicle, legal documents, customer instructions, and hazard information;
• critical steps and controls: activities that prevent safety, quality, or service failure;
• exception rules: when normal work stops and escalation begins;
• records: timestamps, signatures, photographs, temperature readings, and incident reports;
• revision control: owner, approval date, version, and review frequency.

2.1 Pre-dispatch controls

Before departure, the dispatcher should verify order completeness, vehicle suitability, driver qualification, route restrictions, delivery windows, loading sequence, and required equipment. The driver should conduct a vehicle inspection covering tyres, lights, brakes, mirrors, fluid leaks, safety equipment, and visible damage, subject to local law and company policy. Cargo controls include count, packaging condition, weight distribution, restraint, seal number, and special handling requirements.

A vehicle should not be released merely because it is available. It must be fit for the assigned work. For example, a refrigerated vehicle with an unverified temperature condition may be operationally available but not serviceable for a chilled load.

2.2 Loading and handover

Loading should follow the route sequence, weight and axle constraints, segregation rules, and securement requirements. The handover should reconcile the physical shipment with the electronic manifest. Any discrepancy should be resolved or explicitly accepted by an authorised person before departure.

The loading record should normally include shipment identifiers, piece count, seal number where relevant, hazardous-material status, temperature requirement, and departure time. These controls create traceability and reduce disputes.

2.3 Delivery and POD

POD is the evidential record that a delivery event occurred under specified conditions. It may include recipient identity, signature, timestamp, geolocation, quantity accepted, condition, photographs, seal status, and exceptions. A GPS arrival event alone is not POD; it indicates location, not acceptance or condition.

POD quality has three dimensions:

• completeness: all required fields are present;
• accuracy: the record corresponds to the actual event; and
• timeliness: it is available soon enough for invoicing, customer visibility, and dispute resolution.

Electronic POD (ePOD) improves speed and legibility, but it introduces cybersecurity, device availability, user-authentication, and data-retention requirements. Offline capture and later synchronisation are important where connectivity is unreliable.

2.4 Closeout and settlement

At route completion, the operation should reconcile delivered, refused, damaged, short, returned, and undelivered quantities. The vehicle’s ending mileage, fuel, time, tolls, and waiting time should be recorded. Unresolved exceptions should have an owner and due date. Finance should not treat an incomplete delivery record as a routine completed stop, because this can distort revenue recognition, claims, and KPI reporting.

3. Control towers and exception management

A fleet control tower is a coordinating capability—not simply a room containing screens. It integrates data, people, procedures, and decision rights to provide end-to-end operational visibility and intervention. It may be centralised, regional, outsourced, or hybrid.

A control tower typically performs four functions:

1. visibility: consolidate orders, routes, vehicle positions, milestones, and alerts;
2. prediction: identify probable lateness, capacity failure, temperature excursion, or missed connection;
3. orchestration: allocate actions to dispatchers, carriers, depots, drivers, or customer-service teams;
4. learning: analyse causes and improve plans, standards, and supplier performance.

A useful operating model separates monitoring from decision authority. A dashboard may identify a risk, but a named role must decide whether to resequence stops, use a subcontractor, notify a customer, or accept a service failure.

3.1 Exception taxonomy

Exceptions should be classified by cause and consequence. Common categories include:

• capacity: no vehicle, driver, or loading slot;
• asset: breakdown, tyre failure, refrigeration fault, or telematics loss;
• network: congestion, road closure, weather, border delay, or depot backlog;
• cargo: damage, shortage, incorrect documentation, or temperature deviation;
• customer: closed site, refusal, changed appointment, or unloading delay;
• compliance and safety: hours-of-service risk, dangerous-goods breach, collision, or injury;
• data: incorrect address, duplicate order, missing milestone, or device failure.

A practical prioritisation matrix considers severity, probability, time to impact, and recoverability. A low-probability collision is still a high-priority event because its potential harm is severe. Conversely, a minor delay with substantial recovery time may be handled through routine monitoring.

3.2 The exception response cycle

The response cycle is: detect, validate, assess, decide, communicate, execute, and close. Validation is essential because automated alerts can be false—for example, a stationary vehicle may be waiting at a customer rather than experiencing a breakdown. The record should capture the original alert, diagnosis, decision, person responsible, time of action, and outcome.

The aim is not to maximise interventions. Excessive intervention creates driver distraction, route instability, and contradictory instructions. Control-tower performance should therefore include the quality and outcome of interventions, not merely the number of alerts processed.

4. Communication and customer service

Transport communication is an operational control. It aligns expectations, reduces uncertainty, and enables recovery. Communication should be timely, factual, role-appropriate, and proportionate to impact.

A disruption message should state:

1. what happened and when;
2. which shipment or route is affected;
3. the current operational impact;
4. the best available estimate of the revised outcome;
5. the action being taken;
6. the next update time; and
7. the responsible contact.

Avoid false certainty. “The vehicle is delayed by an estimated 45–60 minutes because of a road closure; the next update will be at 14:00” is more credible than an unsupported promise of exact arrival.

Customer service should distinguish between information, recovery, and resolution. Information explains the status. Recovery changes the operational plan. Resolution closes the commercial or service issue, including claims, credits, redelivery, or root-cause action. Service recovery should be governed by authority limits so that employees can act quickly without creating inconsistent commitments.

Drivers also require communication discipline. Messages must not encourage unsafe phone use or pressure to violate driving-time, speed, loading, or rest requirements. Hands-free systems and permitted communication windows should be defined by policy and law.

5. Measuring service and fleet performance

KPIs are measurement conventions, not self-executing truths. Each KPI requires a precise numerator, denominator, data source, time period, inclusion rule, and owner. Performance should be segmented by lane, customer, vehicle type, depot, carrier, and exception category where sample size is adequate.

5.1 Vehicle utilisation

Utilisation has several valid meanings. A common operational definition is the proportion of available vehicle capacity used. For weight-based freight:

For volume:

The symbols and units must match: tonnes with tonnes, or cubic metres with cubic metres. A mixed shipment may be constrained by weight, cube, pallets, axle limits, or loading geometry. Therefore, high weight utilisation does not necessarily mean good overall utilisation. Time utilisation—productive operating hours divided by available hours—is a different measure and should not be substituted without disclosure.

5.2 Empty mileage

Empty mileage measures movement without revenue-generating or operationally justified cargo. A basic ratio is:

where total distance equals loaded distance plus empty distance for the defined trip set. Units may be kilometres or miles, but they must be consistent. Some empty travel is structurally necessary, such as positioning to a remote pickup or returning a vehicle to its base. Analysis should therefore separate avoidable empty distance from strategically required repositioning.

5.3 On-time delivery

On-time delivery (OTD) is commonly:

The time-window definition must be explicit. A delivery at 10:05 is on time for a 10:00–10:30 window but late for a strict 10:00 appointment. Cancelled orders, customer-caused closures, and force-majeure events require documented inclusion rules. Reporting multiple views—gross OTD, carrier-controlled OTD, and customer-caused exceptions—can support fairer diagnosis, but the published headline measure must remain stable.

5.4 Cost per stop

A useful operating measure is:

Relevant cost may include driver labour, vehicle cost, fuel, tolls, subcontractor charges, and allocated dispatch cost. It must be stated whether fixed overhead, failed deliveries, returns, and accessorial charges are included. Cost per stop can fall when stops are consolidated, but excessive consolidation may increase lateness or handling risk. It is therefore a balancing measure, not a standalone optimisation target.

5.5 Worked example

An urban fleet completes 240 eligible deliveries in one day. Of these, 222 occur within the agreed windows. Vehicles travel 3,600 km, of which 720 km are empty. Relevant operating cost is 18,000 currency units, and 240 stops are completed.

OTD is:

Empty-mile ratio is:

Cost per stop is:

Thus, the reported results are 92.5% OTD, 20% empty mileage, and 75 currency units per completed stop. These values do not explain causation. A further analysis might examine whether empty mileage arose from depot imbalance, customer geography, failed backhauls, or route design.

6. KPI governance and managerial interpretation

A KPI hierarchy should connect strategic outcomes to operational drivers. Customer promise and safety are outcome dimensions; OTD, damage, complaints, and incidents are outcome indicators; dispatch readiness, route adherence, dwell time, and POD completeness are process indicators.

Leading indicators provide early warning, while lagging indicators report results. A balanced dashboard may include:

Metrics can create dysfunctional incentives. If drivers are rewarded only for stops per hour, speeding, skipped inspections, unsafe loading, or premature delivery confirmation may increase. This is why safety and compliance should be non-compensable thresholds, and why incentive systems need multiple measures and audit checks.

7. Technology, data, and artificial intelligence

Core execution technology includes TMS, fleet-management systems, GPS, electronic logging, ePOD, barcode or RFID scanning, geofencing, temperature sensors, mobile applications, and customer portals. Integration should be evaluated by data quality and decision usefulness, not by the number of devices installed.

Important data controls include master-data ownership, address validation, time synchronisation, identity and access management, audit trails, retention rules, and cybersecurity incident procedures. A missing scan can be a process failure; a duplicated scan can create false inventory and billing outcomes.

AI can support ETA prediction, anomaly detection, image-based damage assessment, driver coaching, and demand or capacity forecasting. However, AI outputs are estimates, not authority. Governance should address training-data bias, explainability, model drift, human override, privacy, and accountability. For example, a model trained on historical traffic may perform poorly during a new road closure or extreme weather. High-impact decisions—disciplinary action, customer refusal, or safety intervention—should retain meaningful human review.

8. Applied case: NorthStar Home Delivery

This is an illustrative case. NorthStar Home Delivery operates 42 vans from two depots and delivers appliances to households within scheduled two-hour windows. Its headline OTD has fallen from 94% to 86% over three months. Customer complaints report late arrivals, incomplete installation evidence, and inconsistent updates.

Background and problem

The company’s dashboard uses driver-reported completion times but does not distinguish customer-caused waiting from route lateness. ePOD photographs are optional, and dispatchers monitor vehicle locations without a documented intervention policy. Empty mileage is 24%, partly because the two depots serve overlapping areas.

Analysis

A four-week review separates planned arrival, geofence arrival, customer acceptance, and route completion. It finds that 38% of late deliveries originate in late depot departures, 27% in excessive installation dwell time, 21% in address or appointment errors, and 14% in traffic and vehicle incidents. The previous OTD measure also counted several customer-requested reschedules as fleet failures.

Approach and solution

NorthStar introduces a readiness gate 30 minutes before dispatch, mandatory address and appointment validation, standard installation-time allowances, and a two-level exception protocol. A projected delay over 20 minutes triggers customer contact; a safety or vehicle event triggers immediate escalation. The control tower publishes a revised ETA and next-update time rather than waiting for a complaint.

The company makes ePOD mandatory, requiring recipient confirmation, appliance condition, installation completion status, and photographs where permitted. It redesigns depot territories and uses a backhaul or repositioning plan for selected routes. KPIs are redefined: gross OTD remains visible, but customer-caused reschedules and controllable fleet lateness are separately reported.

Outcome

After eight weeks, illustrative results show OTD rising to 93%, empty mileage falling to 18%, and POD completeness rising from 71% to 98%. Cost per stop initially increases because of additional installation buffers and control-tower staffing, then declines as failed deliveries and repeat visits fall. The improvement is not attributed to technology alone: process ownership, data definitions, and disciplined communication were equally important.

Learning points

1. A headline KPI can conceal multiple operational causes.
2. Better classification may initially appear to worsen or fragment reporting but improves managerial action.
3. Customer communication is part of service execution, not an afterthought.
4. Reducing empty mileage must not compromise promised windows.
5. ePOD creates value only when its fields are meaningful, complete, and governed.

9. Challenges, ethics, and regulatory considerations

Fleet execution operates within road-transport, employment, data-protection, dangerous-goods, vehicle-safety, and accessibility requirements that vary by jurisdiction. Managers must verify applicable local law rather than rely on generic procedures. Records should be retained for legally required periods and protected from unauthorised access.

Driver monitoring raises legitimate privacy and dignity concerns. Location data should have a defined purpose, proportionate collection, restricted access, transparent notice, and retention limits. Constant surveillance can damage trust and encourage gaming. Fatigue-management systems should support safe scheduling, not simply create punitive alerts.

Commercial pressure must never justify overloading, falsifying POD, exceeding driving limits, bypassing inspections, or concealing damage. Governance requires segregation of duties where practical, audit trails, protected reporting channels, incident investigation, and leadership reinforcement that safety outranks short-term service metrics.

Best-practice design principles

• Define every KPI before publishing it.
• Treat the SOP as a living control document and review it after incidents.
• Give every exception one accountable owner and a due time.
• Use thresholds that trigger action, not dashboards that merely display problems.
• Separate planned, actual, estimated, and verified events.
• Design POD for evidential value, not just electronic convenience.
• Balance service, cost, productivity, quality, safety, and compliance.
• Test technology in offline, low-connectivity, and failure conditions.
• Audit incentives for unsafe or misleading behavioural effects.
• Convert recurring exceptions into corrective and preventive actions.

Chapter Summary

Fleet execution is a closed-loop managerial and operational system that turns transport plans into safe, compliant, and customer-visible outcomes. Effective execution depends on readiness controls, loading and delivery discipline, reliable POD, structured closeout, and clear ownership of exceptions. A control tower integrates visibility, prediction, orchestration, and learning, but technology cannot replace decision rights or professional judgement.

Utilisation, empty-mile ratio, OTD, and cost per stop are useful only when their definitions, denominators, data sources, and inclusion rules are explicit. They must be interpreted together because improving one can damage another. Communication is a control mechanism: accurate disruption messages and timely recovery decisions protect customer trust. AI and digital systems can improve prediction and evidence, but privacy, cybersecurity, model limitations, human oversight, and legal compliance remain essential. The central managerial task is to improve reliability and productivity without transferring hidden cost or risk to drivers, customers, or society.

Key Terms

Control tower: A coordinated capability for transport visibility, intervention, and performance learning.

Dispatch: The authorised release of a vehicle, driver, cargo, and route for execution.

Exception management: The structured detection, prioritisation, response, and closure of deviations.

Electronic proof of delivery (ePOD): Digital evidence of delivery, acceptance, quantity, condition, and related details.

Empty mileage: Distance travelled without revenue-generating or operationally justified cargo.

Geofence: A virtual geographic boundary used to trigger or validate events.

KPI: A defined performance measure linked to an operational or strategic objective.

On-time delivery (OTD): The percentage of eligible deliveries completed within the agreed time window.

Operational readiness: The condition in which vehicle, driver, cargo, documents, and systems are fit for assigned work.

Proof of delivery (POD): Evidence that a delivery occurred under specified conditions.

Route adherence: The degree to which actual route execution follows the authorised plan.

Service recovery: Actions taken to reduce the impact of a service failure and restore customer confidence.

Standard operating procedure (SOP): A controlled description of roles, steps, controls, records, and exceptions for a recurring process.

Utilisation: The proportion of available capacity or time used productively, under a specified definition.

Dwell time: Time spent waiting or being serviced at a location rather than travelling.

Exception owner: The person accountable for resolving a defined operational deviation.

Leading indicator: A measure that provides early warning of future performance.

Lagging indicator: A measure that records an outcome after it has occurred.

Review Questions

5 Short Answer Questions

1. Why is GPS arrival data not, by itself, sufficient proof of delivery?
2. Identify the six stages of the fleet execution closed loop.
3. What information should a disruption message contain?
4. Define empty-mile ratio and state its denominator.
5. Distinguish a leading indicator from a lagging indicator.

5 Long Answer Questions

1. Explain how an SOP functions as a control architecture for daily fleet execution.
2. Discuss the roles, capabilities, and decision rights required in an effective fleet control tower.
3. Critically evaluate OTD and cost per stop as performance measures, including their limitations.
4. Design a communication and escalation process for a serious delivery disruption.
5. Evaluate the benefits and governance risks of ePOD, telematics, and AI-supported ETA prediction.

3 Analytical / Application Questions

1. A fleet completes 320 deliveries, with 288 within their agreed windows. It travels 5,000 km, including 1,250 empty km, and incurs relevant cost of 22,400 currency units. Calculate OTD, empty-mile ratio, and cost per stop. Interpret the results and identify one additional KPI needed for diagnosis.
2. A dispatcher receives an alert that a vehicle has been stationary for 35 minutes. Develop a validation, prioritisation, communication, and closure sequence, distinguishing a customer unloading delay from a breakdown.
3. A carrier’s OTD improves from 90% to 96%, but damage claims and driver hours-of-service violations increase. Analyse why the KPI improvement may be misleading and recommend a balanced performance-governance response.

Multiple Choice Questions with Answers

Question 1

Which statement best describes the role of a fleet control tower?

A. It replaces all dispatchers with automated routing software.

B. It integrates visibility, prediction, orchestration, and learning.

C. It records only vehicle locations after delivery.

D. It focuses exclusively on reducing fuel consumption.

Correct Answer: B. It integrates visibility, prediction, orchestration, and learning.

Question 2

Which item is not, by itself, sufficient proof of delivery?

A. Recipient confirmation with timestamp

B. Documented quantity and condition

C. GPS arrival at the customer geofence

D. A signed delivery record

Correct Answer: C. GPS arrival at the customer geofence

Question 3

A fleet travels 4,000 km, including 800 empty km. What is the empty-mile ratio?

A. 5%

B. 20%

C. 25%

D. 80%

Correct Answer: B. 20%

Question 4

A fleet completes 200 eligible deliveries and 186 are within the agreed windows. What is OTD?

A. 7%

B. 86%

C. 93%

D. 107.5%

Correct Answer: C. 93%

Question 5

Which practice most improves KPI comparability?

A. Changing the denominator each week

B. Publishing definitions and inclusion rules

C. Removing all exceptions from the data

D. Using only financial outcomes

Correct Answer: B. Publishing definitions and inclusion rules

Question 6

What is the correct first response to an automated long-stoppage alert?

A. Immediately discipline the driver

B. Ignore it until the route is complete

C. Validate the alert and determine the cause

D. Automatically reroute every remaining stop

Correct Answer: C. Validate the alert and determine the cause

Question 7

If relevant transport operating cost is 12,000 currency units and 160 stops are completed, cost per stop is:

A. 60

B. 75

C. 90

D. 120

Correct Answer: B. 75

Question 8

Which is a leading indicator of possible future service failure?

A. Completed customer complaint count

B. Final monthly OTD

C. Late depot departure before route execution

D. Quarterly claims paid

Correct Answer: C. Late depot departure before route execution

Question 9

Why should safety and compliance accompany productivity KPIs?

A. To make dashboards longer

B. Because productivity-only incentives can encourage unsafe behaviour

C. Because safety is unrelated to fleet performance

D. To eliminate the need for operational data

Correct Answer: B. Because productivity-only incentives can encourage unsafe behaviour

Question 10

Which governance measure is most appropriate for AI-supported ETA prediction?

A. Allowing the model to make every disciplinary decision automatically

B. Removing human override to ensure consistency

C. Monitoring model drift and retaining human review for high-impact decisions

D. Using predictions without recording their data source

Correct Answer: C. Monitoring model drift and retaining human review for high-impact decisions