Carbon-aware network design

A network designed only for cost often isn’t the lowest-carbon one. Fewer, larger warehouses can mean longer truck journeys; air freight saves days and burns far more fuel than sea or rail. Carbon-aware design puts emissions into the same model as cost and service.

First, measure. Emissions are usually reported under the GHG Protocol:

• Scope 1: direct emissions, such as your own trucks and boilers.
• Scope 2: purchased electricity for plants and warehouses.
• Scope 3: everything else in the value chain, including outsourced transport, suppliers and product use. For most companies this is the largest share.

Freight emissions are commonly estimated as tonne-kilometres × an emission factor for each mode, which is far higher for air and road than for rail and sea.

Then design. There are two common ways to add carbon to the optimization:

• As a cost: put a price on each tonne of CO₂, so the model trades money against emissions directly.
• As a constraint: cap total emissions, or require a cut such as 20%, and find the cheapest network that meets it.

Running the model with different caps gives a trade-off curve: often the first chunk of emission cuts is cheap (better truck loads, shifting long lanes from road to rail, solar on warehouses), while deeper cuts cost much more. That curve is what leadership needs to set a sensible target.

Freight emissions, step by step

Emission factors depend on the mode, the vehicle, how full it is and the fuel, so use published defaults (such as those in the GLEC Framework) or your carriers’ measured data, and make sure every lane uses the same basis.

That break-even, the cost per tonne avoided, is the abatement cost of the option. Ranking every option by its abatement cost gives a marginal abatement cost curve: the cheap cuts on the left, the expensive ones on the right.

Putting carbon in the network model

• As a price: add (carbon price × emissions) to the cost objective. Each lane, mode and site is chosen as if carbon were a real cost.
• As a cap: add a constraint that total emissions stay below a limit, and minimize cost under it. Tightening the cap step by step traces the cost–carbon trade-off curve.
• As a reported metric only: the weakest option, but a necessary first step to build the data.

Levers a network model can choose

Mode shift (road to rail or sea), fuller vehicles and better consolidation, shorter distances through better-placed sites, cleaner energy at warehouses and plants, and electric vehicles for urban deliveries.

The Scope 3 data problem

For most companies, the largest emissions sit in Scope 3: purchased goods, outsourced transport, product use and disposal. Measuring them depends on data from suppliers and carriers, often unavailable or inconsistent. Many companies start with spend-based estimates (rupees spent times an average factor per rupee), which are easy but flawed: if a supplier raises prices, measured emissions rise even though nothing physical changed. Moving to activity-based data (tonne-km, kWh, kg of material) and then to supplier-specific data improves accuracy and makes real improvements visible.

Standards and rules

• GHG Protocol: the corporate and Scope 3 standards define what to count.
• ISO 14083 and the GLEC Framework: methods for calculating emissions from transport chains, including hubs and multiple legs.
• Science Based Targets initiative (SBTi): a widely used route for setting targets aligned with climate goals.
• Disclosure rules: in India, SEBI requires the largest listed companies to publish a Business Responsibility and Sustainability Report (BRSR), with assurance of key indicators phased in for the biggest firms. Exporters to the EU also face its Carbon Border Adjustment Mechanism (CBAM), which puts a carbon cost on imports of goods such as steel, aluminium and cement.

Trade-offs that can backfire

• Fewer, larger warehouses cut building energy and inventory but may lengthen truck journeys.
• Faster delivery promises push towards air freight, part-empty vans and more nodes.
• Shifting to rail can need more inventory to cover longer and less flexible transit times.
• Intensity versus total: emissions per unit can fall while total emissions rise with volume growth. Targets should say which they mean.

A network model with carbon included makes these interactions visible instead of leaving each team to optimize its own piece.

Credible targets

Base targets on measured baselines, separate quick wins (load factors, mode shift on long lanes, renewable electricity) from structural changes (network redesign, supplier changes), and tie them to the abatement cost curve so the business knows what each extra cut will cost. Report progress with the same method every year, and restate the baseline if the method changes.