Route Optimisation Emission Factors in 2026: Why the Factor You Choose Moves the Number More Than the Route You Plan

Route optimisation emission factors are the published conversion values a route planning system applies to distance and payload to turn a plan into a carbon figure. Choosing among them is not a technical detail: applying different but legitimate selections from the same official factor set to one unchanged distribution operation produces reported annual emissions ranging across 69%, while route optimisation itself moves the underlying number by roughly 21%. Any programme reporting routing-driven carbon savings without fixing its factor conventions first is measuring its own accounting choices rather than its routes. Locus, the world’s first Decision-Intelligent, Agentic TMS, holds factor selection, load basis and distance basis as versioned configuration alongside the route plan, so the number moves only when the operation does.

Why Factor Choice Beats Route Choice: The Business Case

The factors themselves carry enormous internal range. The UK Government’s 2024 conversion factors put an articulated vehicle above 33 tonnes at 0.07447 kg CO2e per tonne-kilometre at average load, against 0.17853 for the all-rigids average. Choosing a class label is therefore worth a factor of more than two before a single kilometre is driven.

Load convention multiplies that. The same set publishes separate values at 50% and 100% laden alongside an average, and the methodology states that the effect of vehicle loading on CO2 is linear with load, so factors can be interpolated where the actual load is known. Interpolating against measured load factors and taking the published average are both defensible, and they do not produce the same answer.

Scope decides the rest. Tank-to-wheel counts combustion only; well-to-wheel adds producing and distributing the fuel. The GHG Protocol Corporate Value Chain (Scope 3) Standard governs how transport emissions are reported in a corporate inventory, and the conversion factor set publishes well-to-tank values separately precisely so they can be added. A figure quoted without saying which scope it uses is not comparable to anything.

This is a routing problem specifically, rather than a general accounting one, because routing is the intervention whose result is most often claimed and least often isolated. A network redesign or a fleet renewal is visible in the asset register and nobody confuses it with a spreadsheet change. A sequencing improvement shows up only as a smaller number, and a smaller number is exactly what a looser convention also produces. The two are indistinguishable in the output unless the conventions were pinned before the programme started.

The publisher is candid that the underlying averages are coarse. The same methodology notes that traffic-, route- and freight-specific factors are not currently available, though these would present a more appropriate means of comparing modes. The conventions therefore matter more than the precision, because there is not much precision to be had.

How a Route Plan Becomes a Carbon Number

1. Distance is fixed from the plan or the execution record

Planned distance credits a route that may not have been run. Actual distance from telematics includes diversions, dead running and re-attempts. The gap between them is the first fork and it is invisible in the output.

2. A vehicle class is selected

The factor set offers rigid bands, articulated bands, and aggregate averages across all of them. Assigning the aggregate where the specific class is known is the single largest lever in the calculation.

3. A load basis is chosen

Either the published average-laden value is applied, or the 50% and 100% values are interpolated against the measured load factor for each leg. Both are supported by the source and they diverge substantially.

4. Refrigeration is recognised or ignored

The set publishes separate refrigerated factors. A chilled operation using standard values understates, and the understatement scales with the temperature-controlled share of the fleet.

5. A scope boundary is set

Tank-to-wheel alone, or tank-to-wheel plus well-to-tank for a well-to-wheel total. This choice is additive and consistent in direction, which makes it the easiest of the five to govern and the most commonly left unstated.

6. A factor vintage is pinned

The set is republished annually. Restating prior periods on the new vintage preserves comparability and changes history; not restating preserves history and breaks comparability. Either is defensible, and choosing neither deliberately is what produces unexplained movement.

What the Model Shows

The model applies published conversion factors to a stated operation rather than to observed customer data. The operation is fixed at 100 million tonne-kilometres a year: 55% on articulated vehicles above 33 tonnes at 70% load, 30% on rigids at 55% load, and 15% refrigerated articulated at 65% load. Nothing physical changes across any of the runs below. Only the conventions change, and every convention used is drawn from the same official factor set.

The range is 68.9%. Across the 32 combinations of vintage, class granularity, load basis, scope and refrigeration handling, reported emissions ran from 9,752 to 16,475 tonnes CO2e. The same trucks, the same routes and the same tonnes, reported 1.69 times higher at one end of the range than at the other, on choices made in a spreadsheet.

Three single choices each match or exceed the entire routing prize. Substituting the all-HGV average for correct vehicle classes moved the answer by 22.4%. Taking the published average-laden value instead of interpolating against measured load moved it by 20.9%. Adding well-to-tank to reach a well-to-wheel total moved it by 21.3%. Each is at or above the 20.7% of distance that route optimisation can remove in total, and roughly three times the 7.2% a strong first year of routing delivers. One caveat on that comparison is worth stating: the routing figure is a reduction in distance, which maps to a similar reduction in emissions at constant vehicle and load but slightly less where consolidation raises load factor and the payload penalty claws some of it back. The forks below move reported emissions directly. The comparison is therefore generous to routing, and the conclusion holds anyway.

The vintage update can move in either direction at once. Between the 2024 and the 2025 factor sets, the published average-laden value for all rigids rose 10.6%, from 0.17853 to 0.19748. Interpolated at an actual 55% load, the same class fell 13.6%, from 0.25058 to 0.21637. A fleet that updated its factor set and changed nothing else would report an increase or a decrease of similar magnitude depending only on which load convention it had adopted years earlier.

Sloppy method flatters, which is the uncomfortable part. The lowest figure in the whole range came from the coarsest available choices: the aggregate vehicle class, the published average load, and tank-to-wheel only. The highest came from correct classes, measured loads and the well-to-wheel scope that corporate reporting standards call for. The direction is not random, and an operation drifting towards convenience will show falling emissions without touching a route.

The order of repair follows from the size of the movements. Vehicle class is the largest distortion and the cheapest to fix, because the information already sits in the fleet master data and needs only to be joined to the calculation. Scope boundary is next and costs nothing at all, since it is a single documented policy decision rather than a data problem. Load basis is third and is the one that requires real capture work, because interpolation is only available to an operation that records load at each leg. Refrigeration handling and vintage restatement are smaller and can follow. An organisation that fixed those five in that sequence would remove most of the 68.9% range before touching a route.

Distance basis stacks on top of all of it. The five forks above all concern the conversion. Whether the distance entering that conversion is planned or actual, and whether it includes failed attempts and re-attempts, is a sixth choice applied before any factor is selected. It is not quantified here because it depends on an operation’s own execution data, which is exactly why it has to be measured rather than assumed.

The Six Forks Ranked: Key Differences

What to Look for in Route Planning Emissions Calculation

Factors held as versioned reference data

The factor set, its vintage and its identifier should be attributes of the calculation, not constants inside it. Without a version stamp, no period-on-period comparison can distinguish an operational change from a republished spreadsheet. The evaluation question is simple enough to ask in a demo: show the same month recalculated on two different factor vintages, side by side, without re-running the plans.

Load factor carried per leg, not per vehicle

Interpolating against measured load is only possible if load is known at each leg rather than averaged across a shift. This is the capability that separates a system that can use the linear interpolation the methodology permits from one that can only take the published average.

Scope boundary stated on the output

Every figure a platform produces should carry its scope on the face of it. A report that does not say whether it is tank-to-wheel or well-to-wheel is not a defensible number, and the difference here was 21.3%.

Distance traceable to a plan version

An operation that re-optimises during the day has several candidate distances for the same route. Unless the plan version is retained with the executed record, the distance basis is undefined and every downstream figure inherits the ambiguity.

Restatement handled explicitly

When a new vintage is published, the platform should be able to produce both the restated series and the as-reported series. Having only one of them forces a choice between comparability and history that nobody should have to make silently.

Route Emissions Data in Practice

A leading North American retailer. Ocean, rail and road ran through six separate legacy systems, so the same movement could be described differently depending on which system was asked. Consolidation delivered more than $1M in savings with exceptions resolved in under two hours. The reporting consequence is the quieter one: a single execution record is the precondition for a single defensible distance basis.

A Fortune 50 parcel and logistics network. More than a million freight shipments a year across 51 sites and a 4,500-strong driver pool, each site planning independently. Centralising raised weekly execution from 75% to 92% and exposed more than $14M in unused capacity, including $565K at a single site. Utilisation measured against each site’s own plan is the same failure mode as load factor averaged across a shift: the aggregate hides exactly the variance the calculation needs, and it hides it in a direction that looks like performance.

A beverage distributor with depot-based mixed fleets. Vans, trucks and motorbikes to thousands of small retail points, previously planned in spreadsheets, with many retail points holding no validated delivery location. Fuel consumption fell 37% and orders per trip rose 22%, with end-of-day reconciliation down 60% after address validation pinned each shop and trips were reconciled at close. Reconciliation at trip close is where load and distance stop being estimates and become records, which is the precondition for interpolating load rather than defaulting to the published average.

Common Mistakes in Route Emissions Reporting

Reporting a saving before fixing the conventions. If the factor conventions changed at any point in the series, the reported movement contains both the operational change and the accounting change, and the two cannot be separated afterwards because the inputs that would let you decompose them were never retained. Fix and version the conventions first, then start the baseline, then run the routing programme against it.

Using an aggregate vehicle class when the specific one is known. This was the largest single distortion in the model at 22.4%, and it always understates for a rigid-heavy fleet. It is also the easiest to fix, because the information is already in the fleet master data and the only thing missing is the join between the asset record and the calculation. Operations frequently discover the aggregate was chosen years earlier, by someone building a first estimate, and never revisited.

Leaving the scope boundary unstated. Tank-to-wheel and well-to-wheel figures differ by more than a fifth and look identical on a slide. A customer comparing your number against another supplier’s is comparing scopes as much as operations, and in a tender that is a competitive disadvantage handed over for free, because the supplier quoting the narrower scope looks cleaner without being cleaner.

Treating the annual factor update as a non-event. The republished set moved a rigid fleet by around 10% in this model, and in opposite directions depending on the load convention in use. An update applied without a restatement rule will show up as performance, and the year it flatters is the year nobody questions it. The rule should be written once, before anyone knows which way the next update will move.

How Locus Approaches Route Emissions Calculation

Locus, the world’s first Decision-Intelligent, Agentic TMS, treats the conversion conventions as configuration held against the plan rather than as constants inside the engine. The route planning and dispatch layer solves routing, load allocation and carrier selection together and retains load factor at each leg, which is what makes interpolation against measured load possible rather than forcing the published average. Six governance mechanisms covering explainability, traceability, evaluation, autonomy levels, execution sandbox and human-in-the-loop keep each plan traceable to the inputs and the version that produced it, so a distance basis remains defined in an operation that re-plans through the day. The Control Tower carries the executed record against the plan, which is what turns distance from an estimate into evidence and lets a restated series be produced alongside the as-reported one.

The platform reasons across more than 250 real-world operating constraints over 1.5B+ deliveries for 360+ enterprise customers in 30+ countries at 99.99% uptime, with $320M+ in aggregate logistics cost savings, 800M+ miles reduced and 17M+ kg of CO2 avoided. Locus has been recognised by Gartner for seven consecutive years, including the 2026 Gartner Hype Cycle for Supply Chain Execution and Logistics Technologies and the 2026 Gartner Market Guide for Multicarrier Parcel Management Solutions, where ShipFlex is featured as a Representative Vendor. Locus holds Leader designation in the QKS SPARK Matrix for Transportation Management Systems 2025 and the #1 position for Route Planning in G2’s 2026 Best Software Awards. In October 2025, Ingka Investments, the investment arm of Ingka Group, the world’s largest IKEA retailer, acquired Locus. Locus continues to operate independently.

Two deployments show what a defensible calculation depends on. A leading North American retailer ran ocean, rail and road across six legacy systems, so no single record described a movement end to end and no consistent distance basis existed. Consolidating onto Locus produced more than $1M in savings, 99%+ on-time store delivery, 95%+ route compliance, an 80%+ reduction in manual dispatch and break-even inside year one on a six to nine month go-live. A beverage distributor running vans, trucks and motorbikes from depots to thousands of small retail points cut fuel consumption 37% and raised orders per trip 22%, with route planning time down 35% and end-of-day reconciliation down 60%, after address validation pinned each retail point and trips were reconciled at close. Validated locations and reconciled trips are unglamorous, and they are what make a distance figure an observation rather than an assumption.

Route optimisation emission factors decide more of the reported number than routing does. Applying legitimate selections from one government factor set to a single unchanged operation produced a 68.9% range, with vehicle class, load basis and scope boundary each moving the answer by more than the entire 20.7% that better routing can remove. The practical sequence follows from that: fix and version the conventions, capture load and distance at the leg, state the scope on every output, then measure the routing programme against a baseline that holds still. Locus keeps those conventions versioned alongside the plan and the executed record, so a change in the number means a change in the operation. Request a Locus route emissions calculation review to see which conventions your current reporting depends on.

Frequently Asked Questions

Which emission factors should a route planning system use? For UK and European road operations the government conversion factor set is the standard starting point, applied at the most specific vehicle class available rather than an aggregate. The set also publishes separate well-to-tank values, refrigerated variants and values at 50% and 100% load, all of which should be selected deliberately and recorded.

How much does emission factor choice change a reported figure? In this model, 68.9% between the lowest and highest defensible configuration of one unchanged operation. Vehicle class granularity accounted for 22.4%, load basis 20.9% and scope boundary 21.3%, each individually at or above the total emissions reduction available from route optimisation.

Is route optimisation still worth doing for carbon reasons? Yes, and the point is sequencing. Better routing removes roughly a fifth of distance, which is real and permanent, but the reported figure will not reflect it until the conversion conventions are fixed and versioned. Otherwise the accounting noise is larger than the operational signal.

Should emissions be reported tank-to-wheel or well-to-wheel? Corporate value chain reporting expects the full fuel lifecycle, and the conversion factor set publishes well-to-tank values separately so they can be added to the combustion figure. The difference was 21.3% in this model, so the scope has to be stated on every output for a figure to be comparable.

What happens when the conversion factors are updated each year? The reported number moves without the operation changing. Between the 2024 and 2025 sets the all-rigids value rose 10.6% on the published average-laden basis and fell 13.6% at an interpolated 55% load. A restatement rule decided in advance is what stops that appearing as performance.

Why does load factor matter so much in route emissions? Because fuel burn is a function of mass, and the conversion factors reflect it linearly, so a vehicle at full load emits materially more per kilometre and materially less per tonne carried. Capturing load at each leg is what allows interpolation instead of the published average, and the two differed by 20.9% here.