Findings

_{FOR THE RAW DATA, PLEASE SEE THIS PAGE}

1

A VERY ROUGH ESTIMATE OF CO2e FOR SAMPLE PERIOD

• As discussed in point 3+4 of the findings, this is not only a vague, rough, estimated figure, but it’s also undoubtedly in no way accurate or representative.
• That said, to get a rough estimate energy usage of an email, I will use the widely cited 17g CO2e per email (Berners-Lee, 2020; Carbon Literacy, 2022; WholeGrain Digital, 2020; Griffiths, 2020) which is now understood to be an over-estimate (Malmodin, 2024; Carbon Trust, 2021), also see Figure 2.
• To provide a calculation which acknowledges the secondary research I have uncovered which suggests 17g is an over-estimate (Malmodin, 2024; Carbon Trust, 2021), I will reduce the widely quoted figure of 17g per email (see Figure 2) by 79.3% (see Figure 1) by using Malmodin’s (2024) study which shows the difference between using the average energy intensity model (the old calculations by Berners-Lee, Carbon Literacy, etc); and the current estimates using the power model (Malmodin, 2024).
• Using this calculation, there is between 25% and 16.3% reduction between models, as highlighted in Figure 1 below. I have calculated the reduction difference of my findings using the median of these points, 20.7%.
• This means the 17g CO2e per email figure would be reduced to:
  3.5g CO2e per mid-long email

.

Looking at Figure 2, I used the 17g CO2e ‘long email’, as I anecdotally perceive (after briefly looking through all 335 emails) that the vast majority of emails received were quite long and thought-through — I received only a literal handful of emails with just a few words in them.

Below are the final calculations for the estimated impact of my digital activity regarding emails:

.

TOTAL EMAILS RECEIVED:  335

Average energy intensity model:
17g × 335 emails = 5695g

Power Model:
5695g × 0.207 (20.7%)
= 1178.86g
= ~1.17kg CO2e

.

TOTAL EMAIL RECIPIENTS:  ~55,678

Average energy intensity model:
17g × 55,678 email recipients = 946,526g

Power Model:
946,526g × 0.207 [20.7%]
= ~195,930.88g
= ~195.93kg CO2e

.

2

BEING A MEMBER OF STAFF (OR STUDENT) AT UAL, WE RECEIVE A LOT OF EMAILS

• This is perhaps the most intriguing finding.
• Personally I received 335 emails in 10 days.
• I was included in emails sent out to roughly ~55,678 recipients*. That feels like a lot of information being sent around that I was part of.
• There is an obvious dip over the weekend, the Saturday 7 and Sunday 8 December 2024, but otherwise the distribution appears to show little correlation. But also how could it, with such a small sample size.
• 355 emails in 10 days isn’t very many compared to wild claims made in 2017, stating that the average office worker receiving 121 emails a day (Dubé, 2017; Wholegrain Digital, 2020), I’m sceptical of this figure, but all the same, 335 seems menial compared. Perhaps I could even say that I’m happy that I didn’t receive 121 a day, as that would be ~1210 in the sample period…
• The vast majority of new email threads I received were Moodle related (73%, or 165 email threads) therefore suggesting I generally received content I probably ignored.
• Students, on the other hand who are part of these Moodle pages as part of their learning or student experience, ‘should’ be reading these emails as they are ‘likely’ to be someway useful to them.
• Furthermore, in the 10-day sample period one MA course at LCF received 22 Moodle emails, which feels anecdotally quite a lot of emails to receive as a student. This would be on top of the direct emails from their course leader, unit leader, tutor, peers etc. See Figure 3 below, highlighting the most common course emails I was part of during the sample period.
• There is perhaps scope for further research here — do students actually read these emails?

.

* This figure does seem huge, predominantly because on 4/12/2024 all current 21,856 students (UAL, 2025) were sent an internal email, see Figure 3 below or row 78 of my raw data found here.

.

3

DATA TRANSFER ITSELF (sending emails) APPEARS TO CONTRIBUTE VERY LITTLE CO2e, WHEN COMPARED TO SENDING NO DATA (not sending emails)

• Network usage, as in amount of emails being sent and received, has very little (but still some) impact on the amount of energy used (Malmodin, 2024). This is because the network is always on and available.
• Permanent availability of the networked internet (routers always on, network always transferring data, etc) is the majority of the energy usage of internet use, and therefore ultimately the majority of the impact. “network equipment is always on and operates at a baseline power consumption regardless of activity.” (Malmodin, 2024)
• Networks cannot be ‘turned off’ in the traditional sense, there are too many moving parts. End-user devices such as phones and laptops can be turned off, but routers (such as at UAL, or in our homes) are almost never turned off. This means they have a latent energy, which means activity such as sending emails, downloading files, streaming video etc does not add much in the way of energy.
• The amount of devices, its ‘embodied carbon’, and the constant power draw from devices and components of the network is what uses the most energy.

.

4

WORKING OUT ACTUAL CO2e FOR SOMETHING SO SMALL (EMAILS) IS VERY DIFFICULT

• Calculations for the CO2e impact of sending/receiving an email from Berners-Lee (2020), Carbon Literacy (2022), WholeGrain Digital (2020), Griffiths (2020) seem to reference each other in perpetuity, and I was unable to find no real source of the calculation or model for how they worked out their rough calculations of emails energy usage.
• What can be deduced from Malmodin (2024) and The Carbon Trust (2021) is that these calculations used a model of average energy intensity, which means they worked out data transferred (gigabytes transferred) and multiplied this by a constant figure (energy used coefficient), which creates an assumed linear relationship of more data transferred = more power use.
• The recent seminal paper by Jans Malmodin (2024) who suggests an alternate power-model of measuring network energy usage which isn’t based on the linear assumption that energy usage = network usage. The power-model takes in to account the always-on aspect of the network (as discussed briefly above) which highlights actually sending emails, uploading files, streaming contributes very little the actual energy usage of the network (Malmodin, 2024)
• This means I have given a broad range for the actual CO2e of my data transference (network usage) during the sample period, see finding 1.

---

Reference List

Berners-Lee, M. (2020). How bad are bananas?: The carbon footprint of everything. London: Profile.

Carbon Literacy. (2022) The carbon cost of an email: Update!, The Carbon Literacy Project. Available at: https://carbonliteracy.com/the-carbon-cost-of-an-email/ (Accessed: January 13, 2025).

Carbon Trust (2021) Carbon impact of video streaming, The Carbon Trust. Available at: https://www.carbontrust.com/resources/carbon-impact-of-video-streaming (Accessed: January 11, 2025).

Dubé, D.-E. (2017) This is how much time you spend on work emails every day, according to a Canadian survey, Global News. Available at: https://globalnews.ca/news/3395457/this-is-how-much-time-you-spend-on-work-emails-every-day-according-to-a-canadian-survey/ (Accessed: January 23, 2025).

Griffiths, S. (2020) “Why your internet habits are not as clean as you think,” BBC, 6 March. Available at: https://www.bbc.com/future/article/20200305-why-your-internet-habits-are-not-as-clean-as-you-think (Accessed: January 23, 2025).

HESA (2024) Who’s working in HE?, Hesa.ac.uk. Available at: https://www.hesa.ac.uk/data-and-analysis/staff/working-in-he#provider (Accessed: 16 December 2024).

Malmodin, J., Mytton, D., Dag Lundén (2024). Network energy use not directly proportional to data volume: The power model approach for more reliable network energy consumption calculations. Journal of Industrial Ecology, 28(4). doi:https://doi.org/10.1111/jiec.13512.

UAL (2025) UAL ActiveDashboards, Arts.ac.uk. Available at: https://dashboards.arts.ac.uk (Accessed: January 7, 2025).

Wholegrain Digital (2020) Digital declutter toolkit from, Business Declares. Available at: https://www.wholegraindigital.com/digitaldeclutter/ (Accessed: November 18, 2024).