Commutes, trains and flights: the carbon maths of working from anywhere

Hybrid work is often presented as a climate win: fewer people driving to the office, fewer lights on in half-empty buildings. It can be. But the footprint of how a company works depends on three things that pull in different directions — commutes, business travel and office energy — and the numbers are rarely where intuition puts them.

Here is how we calculate them in Welyne WorkOne, with the factors we use. All figures are kilograms of CO₂-equivalent per passenger-kilometre, rounded from ADEME Base Empreinte and the UK government's DESNZ conversion factors (2024 editions).

What a remote day saves

A remote day avoids one round-trip commute. The saving is the distance multiplied by the factor for the way that person travels:

• Petrol or diesel car: 0.192 kg per km
• Electric car: 0.052 kg per km
• Public transport (average): 0.045 kg per km
• Bike or walking: zero

Take someone who drives 18 km each way. One home day avoids 36 km of driving, or about 6.9 kg CO₂e. Two home days a week over 44 working weeks avoid roughly 600 kg a year. The same person commuting by train or metro would avoid about 140 kg; a cyclist, nothing at all.

This is why commute profiles matter. A company-wide average hides an enormous spread: remote days for suburban drivers save many times more than remote days for city cyclists. Welyne WorkOne asks each person for their distance and usual mode once, and applies it to the days they actually work from home.

What a business trip costs

Air travel is where the numbers become large. Per kilometre, short flights are the worst, because take-off and climb burn a disproportionate share of the fuel. The factors we apply include the additional warming effect of emissions at altitude:

• Short-haul flight (under 1,000 km): 0.258 kg per km
• Medium-haul (1,000 to 3,500 km): 0.187 kg per km
• Long-haul (over 3,500 km): 0.152 kg per km
• High-speed train in France: 0.012 kg per km

A return flight from Paris to Amsterdam — about 430 km each way — comes to around 220 kg CO₂e per passenger. The same trip by train is around 10 kg.

Long-haul is on another scale. A return trip from Paris to Montreal, roughly 5,500 km each way, is about 1.7 tonnes per passenger — close to three years of that same driver's remote days.

The office side of the ledger

Buildings consume energy whether or not they are full. Welyne WorkOne attributes a factor of 1.4 kg CO₂e to each occupied desk-day, covering heating, cooling, lighting and services. On its own, that makes a single office day look cheap next to a car commute. The real lever is at building level: a floor that stays heated for a handful of people on Fridays costs almost as much as a full one. Consolidating quiet days onto fewer floors, and switching the rest to standby, often saves more than any individual behaviour change.

What we don't count, and why we say so

Working from home isn't zero-carbon. Heating a home that would otherwise be empty during the day adds emissions, especially in winter and in homes heated by gas. The effect varies enormously with season, climate and heating system, and there is not yet a factor we consider robust enough to apply per person. We show it as a caveat on the dashboard rather than hide it in a number that looks precise and isn't.

Train factors also vary by country. The low figure for French high-speed rail reflects a largely low-carbon electricity grid; rail in countries with more fossil-fuelled grids is several times higher, though still far below flying. Where a country-specific factor is available, it replaces the default.

Where to act first

• Make the train the default under 800 km, and show its footprint next to the flight's when people plan.
• Group trips: one visit that covers three clients instead of three separate flights.
• Encourage remote days where they save most — long car commutes — rather than setting a single target for everyone.
• Close underused floors on quiet days instead of heating them for a few people.

None of these require perfect data. They require data that is honest about its assumptions, visible to the people making the choices, and good enough to show which way the numbers move.