Comparing the carbon footprint of IT companies is ridiculously easy on the one hand but totally misleading, or it is difficult but much more informative, relevant and useful.
The Carbon Disclosure project lists many technology company's carbon footprint. By this we may find out that IBM in 2006 said it produced 2.451 million tonnes of carbon whereas Dell produced about 400,000. Therefore Dell is less harmful to global warming.
This is certainly true in absolute terms but it is an almost useless finding. Dell may emit 1,000 tonnes annually for each one of 400 employees whereas IBM's 100,000 employees emit 245 tonnes each, making IBM vastly better on a per-employee basis. Microsoft might emit 100 tonnes per employee so it is more efficient still, only it has no hardware manufacturing and therefore IBM is at a profound disadvantage in the comparison.
Carbon horses for emitting courses
The size of a business and what it does are hugely important in calculating how relatively good or bad it is compared to its peers in the matter of carbon emissions. There is also the issue of where it gets its electrical energy from.
Normalising business carbon footprints
We might decide that the best way to measure the carbon-emitting aspect of a business is to divide the number of employees, both permanent and part-time, plus the number of contractors by the total carbon emissions for a year. This would give us the carbon footprint per head and provide a valid comparison between businesses involved in the same type of operation.
Another way of doing this would be to use a business' annual revenue instead of its headcount to take size into account. The revenue for a year would be divided by the carbon emissions for that year to generate what Dell calls a carbon intensity number.
If £10 million of revenue generates 10,000 tonnes of carbon then the carbon intensity is 100. If it generates 20,000 tonnes then the carbon intensity is 50, a lower number but a worse carbon intensity because the higher the intensity number the better.
We could take the reciprocal to reverse this effect. One over 100 is 0.001 whereas one over 50 is 0.002 with the larger number now reflecting the larger emissions.
However there can be some businesses with very high revenues and very low emissions, a city financial hedge fund for example. Just because a £10 million hedge fund with ten employees and a 100 toones/year carbon footprint has a carbon intensity of 0.000001 doesn't mean that a factory-based business is totally evil from a carbon footprint point of view.
Sectorising business into different carbon footprint levels
It is unfair to compare a hardware manufacturer, with intrinsically higher power needs for industrial processes, with a software manufacturer which has no industrial process needs. It is probably not valid to compare a chip foundry company with a metal chassis assembly company like Dell.
Let's essay a carbon footprint categorisation of IT suppliers with these points in mind:
a) Chip foundry - like Intel
b) IT manufacturer - like Dell or EMC
c) Hardware and software IT systems company - like IBM, HP and Sun
d) Software-only inventor - like VMware
e) Services company - like Accenture and any people-at-a-desk business.
Categories D and E can probably be combined, both being people-at-a-desk-based business types.
Should we include the carbon footprint of bought-in components? I would say not; use the value-added tax (VAT) principle and count only the carbon emissions a business adds during its operations.
We note of course that most businesses don't directly emit carbon at all. They merely buy an electricity supply and some or all of the generation of that causes carbon emissions. We'll return to this point later.
In order to compare the relative carbon emissions intensity of Intel, Dell, IBM, and VM we must take into account the different base level of carbon emissions to be expected from companies in different sectors, rather like a golf handicapping scheme which enablers golfers of different skill levels to play in the same group and have a meaningful contest.
Creating such a relative weighting scheme is completely beyond the wits of the writer of this feature and is best left to a cross-industry consortium such as the Green Grid.
Not all electricity generation is causative of global warming. The green environmental friends sing the praises of renewable energy generation using wind, waves and sunlight. They may also sing the praises of wood-burning power generation and loathe nuclear power stations.
Yet, from a strict global warming standpoint, any energy generation that involves burning, that is coal-fired, gas-fired, oil-fired or wood-fired, is bad. Any power generation that involves no burning is good and, therefore, nuclear power generation is beneficial.
We ignore here, for the purposes of our discussion, the carbon emission costs of building any power station. This is not to wipe them out of existence but merely to focus on our discussion of IT suppliers' carbon footprints specifically.
Most business's carbon footprint is calculated by counting up the annual electricity consumption and converting that, using an assumed ratio, into carbon emissions.
To the extent that any business has some or all of its energy generated by non-burning methods of power generation then its carbon footprint needs reducing in proportion.
To do this we have to be able to measure the proportion of dirty, i.e. adding to global warming, and clean, i.e. reducing or being neutral to global warming, energy consumed.
Where does this leave us?
I hope that it is now clear that simply comparing absolute carbon footprints of different businesses is almost totally misleading on its own. We need to compare the relative carbon footprints, which take into account:-
1. The size of the business in headcount or revenue terms,
2. The sector of the business in industry type terms, and
3. The proportion of clean and dirty energy supplied to the business.
Until we have an agreed standard for calculating relative carbon footprints then we will not be able to compare and contrast IT suppliers and their carbon footprints.