Overton Blog

Calculating our carbon footprint: a work in progress

As we grow we've been thinking about the kind of company we'd like to be and what kind of practical steps we can take to get there. As well as joining excellent initiatives such as the Living Wage Foundation and the FSB Good Business Charter, we recently set out to calculate our organisation’s carbon footprint. Given that we’re a small team and that we don’t have a physical product, we assumed this would be relatively straightforward…but it turned out to be more complicated than we initially thought!

We wanted to share our method with you for a few reasons. Firstly, we want to be transparent about our working-out. Secondly, we struggled to find much info about how other tech companies attempted to do this so we thought our journey might provide some useful learnings for other startups. And thirdly, we would welcome any feedback to improve the method!

The GHG Protocol is the world’s most widely used greenhouse gas accounting standard so we chose to base our calculations around their structure:

  • Scope 1 are direct emissions caused by any process or activity by the company that causes greenhouse gas emission, eg burning fuels onsite or using company car;
  • Scope 2 are indirect emissions caused by your company purchasing energy (from sources you do not own or control). This is usually in the form of electricity, heat, or steam;
  • Scope 3 emissions are all the indirect emissions that occur because of your business activity. For example: waste (decomposing waste emits greenhouse gases), emissions related to any goods or services purchased (for example any fuel emitted transporting them), employees commuting to and from work.

Reporting on scopes 1 and 2 is essential under the GHG Protocol, whereas reporting on Scope 3 is usually optional because it is very difficult to access the required data. However, we’ve given it a go!

 

Scope 1: Direct emissions

We don’t produce any Scope 1 emissions as we don’t have any manufacturing operations, company vehicles etc. 

=0 tonnes of CO2 per year

 

Scope 2: Indirect emissions from electricity we have purchased 

Our only Scope 2 (indirect) emissions would be from energy used to heat and power our workspace but as we have a coworking space and staff working remotely, we do not purchase electricity directly – therefore this is discussed under Scope 3.

=0 tonnes of CO2 per year

 

Scope 3: Indirect emissions that occur as a result of our business operations

 

Servers

Although we don’t have a physical product and therefore we have no manufacturing or post-use recycling to consider, our platform’s online so we knew our server use was going to be a significant contributor to our carbon footprint. 

Like many other web businesses we rent servers belonging to server hosting companies, which physically sit in very well-connected data centres (most of the ones we use are in northern France) designed specifically for running large numbers of computers. Usually we’re the only company using a particular server, but sometimes we’re sharing them with other businesses.

Calculating the carbon footprint of a server is really quite complicated as it turns out, and even the server hosting companies themselves struggle to accurately measure it. 

It depends on how much renewable energy powers the national grid in the server’s home country, the climate of the surrounding area (warmer locations need more cooling), the efficiency of the data centre and server technology, among other things. We have used the UK Energy Research Centre’s estimate of energy usage per server, which they say is between 500 and 1000 watts an hour depending on its specification. 

Then there’s the challenge of mapping that onto the usage of an individual company such as Overton. Individual servers use more power under heavy load, and less power when they’re idling, and there is no reliable way – at least at a distance – to know exactly how much you’re using in a given period. 

So we have collected all the information we could find, but there are inevitably some unknowns.

We use two different server providers so we calculated our footprint separately for each. 

 

OVH servers

We have 15 dedicated (i.e. they’re not shared) mid-sized servers with OVH. OVH runs its own data centres and therefore has visibility and control over its environmental impact. They publish their global renewable energy share (79%) and their average carbon emissions (0.18T CO2e / MWh) and have publicly committed to ambitious sustainability goals

However, as we know that all of our OVH servers are located in France, we can be more accurate in our estimates.  It turns out that being in a French data centre makes a very big difference in our carbon emissions because the French national grid has a much lower carbon output of 0.054T / MWh (compare that to 0.309T / MWh if we had UK-based servers – if our servers were in London we’d potentially be responsible for 5-6x more CO2).

To calculate the MWh used by Overton’s OVH servers each year:

15 servers

800 Wh used by each server (using UKERC’s mid-range estimate)

8760 hours in 1 year

1,000,000 to convert Wh to MWh

=

105.12 MWh per year

 

Then to calculate the resulting carbon emissions per year:

105.12 MWh per year

0.054T CO2e / MWh (from the French national grid estimate, as of 2019)

= 

5.68 tonnes of CO2 per year

 

Linode servers

Overton also uses 25 1GB “virtual” (i.e. shared) servers provided by Linode. Linode has given an estimate of their server energy consumption as 64 Wh for a 4GB virtual server, which equates to 16 Wh for the 1GB servers we use. As Linode use data centres run by third-parties around the world their ability to report on the % of renewable energy used is limited so we’ve had to make do with the UK government estimate of 0.309kg CO2e / KWh (or 0.309T CO2e / MWh), despite different countries producing different proportions of renewable energy. 

To calculate the MWh of energy used by Overton’s Linode servers each year:

25 servers

16 Wh used by each 1GB server (using Linode’s estimate)

8760 hours in 1 year

1,000,000 to convert Wh to MWh

=

3.5 MWh per year

 

Then to calculate the resulting carbon emissions per year:

3.5 MWh per year

0.309T CO2e/MWh (from the UK government)

=

1.08 tonnes of CO2 per year



Laptop usage

Similarly, we’re reliant on our laptops for almost every aspect of our work so we wanted to include this too. We have 4 full time staff, a typical laptop uses 0.05 KWh (or 0.00005 MWh), according to Smarter Business, and the UK government carbon emission estimate is 0.309T CO2e / MWh. To calculate the footprint of our laptop usage:

4 full time staff

8 hours per day

256 working days per year

0.00005 MWh used by a typical laptop (from Smarter Business)

0.309T CO2e / MWh  (from the UK government)

=

0.13 tonnes of CO2 per year



Office space

As we currently have hot desk seats in a coworking space, we don’t have access to any of the energy bills to figure out the building’s energy usage or the proportion which is renewable. In addition, the number of people who share the space changes on a daily basis, we don’t know how to calculate our proportion of the building’s energy usage. 

We will enquire with the office provider to see if they have any data available or if they have plans to use more renewable energy in future, but as it stands this is likely to be pretty small as we only have 3 people working from the office a few days per week.

= Unknown but likely very small

 

Home working

Although we don’t have access to exact figures for our team’s household energy consumption, we can make an estimate of how much energy may be used by staff to heat and light their homes during work hours. Not all of this will be directly attributable to work – for example if other people share the house, if energy is used for cooking or laundry or if other rooms are being heated/lit – and we have already included laptop usage above – so we’ve assumed a quarter of average household hourly energy use could be attributed to work.

Roughly half our staff time is spent working from home, which equates to 2 FTE staff, and all are UK-based. According to Ofgem, the average household in the UK uses 2,900 KWh of electricity and 12,000 KWh of gas, totalling 14,900 KWh per year. This equates to 1.7 KWh used per hour (14,900 KWh / 8760 hours per year). We used the UK government carbon emission estimate of 0.309T CO2e / MWh.

To calculate the energy used by Overton per year:

2 FTE staff

x

8 hours per day

256 working days in a year

x

1.7 KWh average energy used per hour in a UK household (from Ofgem)

/

4 (to remove non-work energy usage and laptop power) 

/

1000 to convert KWh to MWh

1.74 MWh per year

 

Then to calculate the carbon emissions from this energy:

1.74 MWh per year

x

0.309T CO2e / MWh (From the UK government)

=

0.54 tonnes CO2 per year



Staff travel

All our office-based staff travel to work by public transport – the impact of this is very hard to estimate but likely to be very small so we’re leaving it out for now. At the moment we don’t do very much business travel; however, it’s possible we’ll fly to a couple of international conferences so we’ll include an allowance for 2 seats on a shorthaul flight (London-Rome 234 kg CO2 per passenger) and 1 seat on a longhaul flight (London-New York 986kg of CO2 per passenger).

((2 x 234kg for shorthaul flights) + (1 x 986kg for longhaul flights))

/

1000 to convert kg to tonnes

=

1.5 tonnes of CO2 per year



Waste

Our product is virtual, we are virtually paperless in our operations and our office space doesn’t use single use cups so we don’t send any waste to landfill.

= 0 tonnes CO2 per year

 

Other suppliers

The vast majority of our supply spend is with our server providers and coworking space, which have been explored above. We use a few subscription tools like Hubspot, Slack and Google suite, but calculating our footprint for these is very difficult as we don’t have the environmental impact figures from these providers and we don’t know our proportion of the supplier company’s overall impact.

= unknown but likely very small



Totalling our carbon footprint

This is very much the first step on our journey to understand our environmental impact as an organisation, and as we grow we’ll need to revisit these assumptions and calculations. As we’ve mentioned above there are limitations to the available data on Scope 3 emissions; where this is the case, we’ve over-estimated a bit, and we’ve also added an extra 2 tonnes, just to be on the safe side.

 

Scope 1 = 0 tonnes

Scope 2 = 0 tonnes

Scope 3 = 8.93 tonnes

5.68 tonnes - OVH servers

1.08 tonnes - Linode servers

0.13 tonnes - laptop use

0.54 tonnes - home working

1.50 tonnes - staff travel

0.00 tonnes - waste

SUBTOTAL = 8.93 tonnes of CO2 per year

+ 2 tonnes to account for unknowns (eg office space, other suppliers, additional hires)

TOTAL = 10.93 tonnes of CO2 per year




What next?

Obviously the best option would be to eliminate the carbon emissions we produce altogether, but given that the majority of our emissions are linked to third party server companies there’s a limited amount we can do ourselves. Our OVH servers account for almost two thirds of our footprint but they are ahead of the competition in terms of improving energy efficiency and increasing their use of renewable energy, so we’re satisfied that our aim to reduce our carbon footprint is best served by sticking with them for the time being.

The next best option is to offset our emissions. We know this can be controversial as many companies use it instead of genuinely trying to cut their own emissions and to ‘greenwash’ their brand, and because many players in the carbon offsetting market have been fraudulent or unscrupulous. So we will commit to only using offsetting as a genuine last resort where we cannot reduce our emissions, prioritising suppliers who have a good environmental record and offsetting using a reputable provider.

In order to achieve carbon neutrality, we chose to buy carbon offset credits with Gold Standard as it was founded by WWF and other environmental NGOs and tracks the credits linked to the projects supported so they cannot be re-sold elsewhere. We chose to support two projects: ​​Solar Cooking for Refugee Families in Chad and the Nicaforest High Impact Reforestation Program in Nicaragua.

 

We’d love to hear what you think of our approach and if you have any suggestions for how we can improve our method! Get in touch at support@overton.io

What is Overton

We help universities, think tanks and publishers understand the reach and influence of their research.

The Overton platform contains is the world’s largest searchable policy database, with almost 5 million documents from 29k organisations.

We track everything from white papers to think tank policy briefs to national clinical guidelines, and automatically find the references to scholarly research, academics and other outputs.