TL;DR:
Not all empty miles are the same. Some are avoidable waste caused by weak visibility, poor data and fragmented execution. Others are structural, set by geography, specialist vehicles, imbalanced freight flows and customer rigidity.
A good Transport Management System reduces the avoidable portion and exposes the structural portion. It does not abolish physics, market structure, or the fact that a fridge lorry returning empty from the Highlands is not a software bug.
The article below is followed by a simplified causal loop diagram of that system. Adjust the force strength on each driver to see how the balance between avoidable and structural empty miles shifts.
Empty Running in UK Logistics: Why the Problem Is Bigger Than Bad Planning
Empty running is one of those logistics phrases that sounds harmless until you stop and think about what it means. A heavy goods vehicle moves across the country, consumes fuel, uses road capacity, creates emissions, occupies driver time, adds wear to the vehicle, and earns no direct revenue from the space inside the trailer. It is transport activity without transported goods. That sounds inefficient because it is inefficient.
In UK logistics, this is not a marginal issue. The Department for Transport’s 2025 domestic road freight statistics report that GB-registered HGVs moved 162 billion tonne-kilometres of goods and travelled 19.0 billion vehicle kilometres in 2025. Industry reporting of the same figures puts empty running at around 5.9 billion kilometres, equivalent to roughly 31% of total loaded and empty HGV kilometres. So, roughly one in three HGV kilometres may be travelled without a load.
Even that does not tell the full story, because a vehicle can be technically “loaded” while still being badly underutilised. A lorry could travel 100 miles with one pallet on board and still count as loaded. Operationally, that may still be a poor use of capacity, driver time, fuel, road space and network opportunity. The empty-running statistic is useful, but blunt. It separates loaded from empty, but it does not tell us whether the loaded journey was efficient. A full trailer and a trailer carrying one lonely pallet are both “loaded” in the basic measurement. One is probably productive. The other is merely avoiding the embarrassment of being officially empty.
The real issue is broader than empty running alone. It is about capacity utilisation, flow, service design, planning flexibility and decision quality. The tempting response is to say, “Operators should plan better.” That is true, but only partly true. Empty running is not simply the result of poor planning or weak execution. It is also a symptom of the structure of the logistics system itself.
Not all empty miles are equal. Avoidable empty miles are caused by poor visibility, late decisions, weak master data, fragmented systems, missed consolidation opportunities, poor communication, inefficient execution or limited planning control. These are the miles that better systems, better processes and better operational discipline can reduce. Structural empty miles are different. They are caused by the shape of the network: freight flows are not symmetrical, demand is not evenly distributed, customers do not all want goods moved at convenient times, vehicles are not all interchangeable, and depots are not always in the right place. Then there is a third category: underutilised loaded miles. These are the movements that are technically not empty, but still make poor use of available capacity.
Lean Thinking gives us a useful way to understand this. In Lean terms, empty miles are a form of waste: movement, effort, fuel, time and capacity that do not directly create value for the customer. But Lean also warns us not to treat visible waste as the whole problem. Waste is usually a symptom of the system that produces it. If a vehicle runs empty because the planner lacked visibility, that is a process problem. If it runs empty because customer delivery windows are too rigid, that is a service-design problem. If it runs empty because freight flows are geographically imbalanced, that is a structural problem. If it runs technically loaded but with one pallet on a full-size trailer, that is a utilisation problem. Lean is useful here because it moves the conversation away from blame and towards flow, value and system design.
The convenience loop makes this harder. Modern consumers expect speed, choice, availability and convenience. That expectation flows upstream. Retailers and manufacturers respond with stronger service promises: shorter lead times, narrower delivery windows, higher availability, more frequent replenishment and more flexible fulfilment options. Those promises may create customer value, but they also create operational complexity. A tighter delivery window reduces the planner’s ability to consolidate. A same-day or next-day commitment reduces the time available to find a backload. More fragmented orders create more complex networks. More complex networks reduce planning flexibility. Nobody in this loop is necessarily behaving badly. The customer wants convenience, the retailer wants to compete, the haulier wants to serve the customer, and the planner wants to make the work executable. Yet the system can still produce waste.
This is also where competitive strategy matters. Businesses compete on price, service, reliability and responsiveness, but logistics-heavy businesses often compete by making stronger promises than their competitors: faster delivery, later cut-offs, more precise time windows, better availability and less friction. Each promise may be commercially sensible, but when every firm competes by tightening the service offer, the whole market can move towards a higher-complexity operating model. It becomes a strategic trap. If one business offers a more convenient service, others may feel compelled to match it. If they do, they inherit the operational complexity. If they do not, they risk losing customers. The result is a market where everyone is forced to run harder just to stand still.
The tragedy of the commons is a useful comparison. Each actor makes a decision that is rational from their own position, but the aggregate result is worse for everyone. A customer wants a specific delivery time because it suits their operation. A retailer wants a tighter promise because it improves customer satisfaction. A haulier accepts a suboptimal movement because the vehicle has to reposition anyway. A planner prioritises the urgent load because failing service today is more visible than creating waste tomorrow. Each decision can be defensible, but collectively they consume shared resources: road capacity, driver availability, fuel, carbon budget, depot space, planning bandwidth and network flexibility. The commons here is not a village field. It is the national logistics network.
This is where a good Transport Management System has real value. It does not reduce empty mileage by magic; it reduces it by improving the quality, timing and visibility of decisions. Planners need to see orders, vehicles, depots, delivery windows, driver availability, subcontractor options, vehicle types, customer rules and constraints in one decision space. If that information sits across disconnected systems, emails, portals, phone calls and someone’s memory, matching work to capacity becomes slower and weaker. A mature TMS improves load consolidation by identifying orders that can sensibly travel together, taking account of geography, timing, vehicle type, delivery sequence, customer restrictions and available capacity. It improves backload identification by showing where vehicles will become empty, what work is nearby, whether the timing works, whether the equipment is suitable and whether the commercial return is worthwhile.
A good TMS also reduces under utilisation by making fill visible. Empty mileage is only part of the problem. The better question is not simply, “Is the vehicle loaded?” but “Is this movement genuinely efficient?” That means understanding pallet fill, cube utilisation, weight utilisation, vehicle suitability, revenue per mile, route quality and whether the journey makes good use of time and capacity. A weak system records empty running after it has happened. A good system helps reduce avoidable empty running before it happens. A better system also highlights underutilised loaded journeys, because the aim is not simply to avoid empty trailers. The aim is to improve the flow and utilisation of the whole transport network.
From a product perspective, this is the key point. A TMS should not just be a transaction system where users enter orders, build loads and allocate vehicles. It should be a decision-support system. The product challenge is not merely to display more information, but to improve decision quality. A planner does not just need a list of theoretical backloads. They need usable options that account for time, geography, compatibility, customer rules, vehicle type, driver hours, depot constraints, commercial acceptability and confidence in the data. The best product design helps users understand trade-offs before the plan has already gone sideways.
The companion model should be read in that spirit. It is not intended to be a precise forecasting engine. The percentages are indicative rather than definitive, and the aim is to make the structure of the problem visible. It separates avoidable empty miles from structural empty miles because that is the critical management distinction. In practice, a fuller operational model would also need to consider loaded utilisation: pallet fill, cube, weight, revenue per mile, route compatibility, driver hours and network opportunity. The empty-running statistic tells us something important, but not everything important. It is better to be approximately right about the structure of the system than precisely wrong about a single number.
This is also where Logan’s Run is a surprisingly useful reference. In the film, the system appears clean, controlled and rational, right up until you realise the rules are hiding a much darker reality. Logistics metrics can work the same way. A simple loaded-versus-empty measure may look tidy, but it can hide underutilised movements, poor flow and structural waste. The number may say the vehicle was loaded. The operation may tell a different story.
Empty running in UK logistics is therefore not simply a failure of planning. It is a visible symptom of a deeper system: consumer expectations, competitive pressure, service promises, operational complexity, Lean waste, weak utilisation measurement, technology maturity and structural constraints. Some empty running is avoidable and should be attacked through better systems, better visibility, better planning processes and better execution control. Some of it is structural and needs to be understood, priced, redesigned or accepted as the cost of a particular operating model. Some loaded running is still inefficient and needs to be measured through better utilisation metrics.
Or, to borrow the spirit of Life of Brian, the problem is not that nobody is trying to help. It is that everyone is pulling in slightly different directions while insisting they are part of the same movement. In logistics, that gets expensive very quickly.
How to use the model
Apologies the model will not render well on small screens
Reading the loops
- R1 — Convenience loop. Consumer expectations increase service promises, which increase network complexity and reduce planning flexibility, reinforcing pressure on the freight network.
- R2 — Competition loop. Competition on speed and availability creates tighter commitments and less operational slack, which can increase avoidable empty running.
- B1 — TMS control loop. TMS maturity improves visibility and matching opportunities, reducing avoidable empty miles and giving planners better options.
- B2 — Structural ceiling. Structural constraints limit what optimisation can remove. Some empty miles are system-driven and set a floor below which execution alone struggles to go.
The core message is the same one that runs through Theory of Constraints and systems thinking generally: improving the wrong part of the system, or expecting software to remove waste that is structural, leads to disappointment. The useful question is not “can we get to zero?” but “which empty miles are ours to remove, and which are the system’s?”
Click a loop button, a loop badge, or any variable to highlight the related causal path and dim everything else. Drag the sliders in the Force Controller to change the strength of each driver. The output variables — avoidable, structural and total empty miles — are calculated rather than set directly, and update live in the KPI panel.