Gauthier Roussilhe

Back to articles

Digital sustainability, the fog to come

Gauthier Roussilhe, August 20 2020

What is known as "digital sustainability", i.e. designing relevant digital services while reducing their environmental footprint, has started to gain momentum over the last two years. In France, this field is accelerating rapidly and new players are arriving every day. This acceleration is linked to the regulatory and social progress of the last two years on the subject. New public funds and aids are dedicated to it, the specifications are evolving in this direction and offers are being structured. In short, this sector of activity is reaching a new phase of maturity. Thus, the older players are going on the defensive to protect their "market share" as "broader-shouldered" players (digital companies, consulting agencies, etc.) come in. What was yesterday a small market and above all a field of research and practice is becoming a more or less important market and, in my opinion, we are witnessing a movement to close off knowledge, fuelled by the fear of seeing one's activities phagocyted. This fear is legitimate in the sense that the organisations and people in demand (public administration, local authorities, companies, etc.) do not have the necessary background to determine whether a provider is really competent or not. In a movement contrary to the one I am observing today, I wish to open up my methodology and, at the same time, make a critical assessment to improve demand and give some advice on how to better choose a service provider.


Table of Contents

What is available?
Markets and methodologies
Is optimising a service digital sustainability?
Digital sustainability is not a miracle solution
The training gap
Advice for the years to come

What is available?

In France, when you start to take an interest in the subject of digital sustainability, you quickly come across a few practical documents such as the 115 good practices of GreenIT and Frédéric Bordage, or the good practices reference framework of the "Institut du Numérique Responsable" (INR), or, more recently, the guide of the association "Designers Éthiques". The first is very oriented towards programming practices, the second and third include other fields such as UX design, strategy, etc. However, before looking at these technical documents it is important to understand the eco-design framework. This is normalised by ISO standards such as 14062:2002 (IEC 62430:2019 now), but this standard is general to eco-design and is not specific to digital. This is to say that the design of a digital service can be viewed, from an environmental point of view, as the creation of any other service. This assertion is partially true but also poses real limits to the practical application as we shall see.


Markets and methodologies

Over the last four years, I have been fed with a lot of existing literature and I have mostly piloted sustainable digital projects in "real life conditions" with different types of clients. From this digestion and practice I have derived my "tailor-made" methodology which corresponds to my knowledge and vision of things. This methodology aims to lay the foundations of my digital sustainability practices and to highlight the crucial points for trade-offs throughout the project.


As a first step, I have selected 3 conditions (or starting points) that are not exceptional or new:


  1. 1 → The environmental footprint of the service, whether digital or not, must be reduced ;

  2. 2 → The needs expressed by users must be met with relevance ;

  3. 3 → It must be assumed that digitisation is not necessarily the best option to address the first two points.


I would like to press on the third point and thus formulate my first advice: if the person who helps you on digital sustainability does not sincerely and often question the digitisation of your service, or the digitisation of certain functions of it then there is a strong chance that he or she has not fully understood the goal of digital sustainability. I'm not saying that this questioning is easy, I myself failed to work in depth on this issue on a large project I'm working on, but that doesn't prevent me from regularly questioning the project team on this. Eventually, the profession will have to be better equipped to support this fundamental questioning.


In a second step, when the conditions are met, I lay down seven pillars that will guide the digitisation process:


  1. → The service must encourage the lifespan of the equipment ;

  2. → The service should reduce resource consumption (both environmental and IT) in absolute terms ;

  3. → The service should support its own lifespan by meeting relevant medium and long-term needs and facilitating maintenance and evolution ;

  4. → The service must be optimised for the most difficult access conditions (old equipment or with limited computing power, little network access, paid data) ;

  5. → Digital sustainability is only part of a virtuous circle that includes accessibility, privacy, open data, free software, etc;

  6. → Opening and documenting work done should be the norm, not the exception ;

  7. → The work done must be measured and must be part of a pre-existing environmental strategy.


All of these pillars respond to each other and must therefore be taken altogether to hold the project together. Similarly, behind each pillar there are many good practices, indicators, trade-off methods that I cannot explain here for lack of time. Once again, I will rely on one pillar, the fifth, to provide a second piece of advice: digital sustainability can be seen in an additional layer to an ever more complex specification. For me it is the opposite, it is an additional pretext to mutualise the effort and to make a high-quality digital environment happen with a single movement (accessibility, open source, security, etc.). I don't do digital sustainability projects without bringing in web accessibility or without trying to take out GAFAM services from the project as much as possible. If your service provider doesn't fight on this point then they are not doing you any favours. For example, proposing an AWS (Amazon Web Services or other Google or Microsoft cloud services) as part of a digital sustainability mission is a contradiction, except for exceptional needs. Adding a layer that will complicate the issues of RGPD and sovereignty under the pretext that Amazon would announce "green" data centers is a strategic error. All virtue comes together or nothing happens.


Summary of the 3+7 methodology

This is a brief overview of my personal methodology, it is not fixed and will continue to evolve (new pillars will appear). I would not advise using it because I have adjusted it specifically to my knowledge and my vision of the subject. In any case, this introduction has made it possible to highlight two pieces of advice that seem to me to be key in the choice of a service provider and in the understanding of qualitative support.


Is optimising a service digital sustainability?

When I work on a project, I use the methodology presented above to do the scoping. In addition, I use various technical indicators to monitor the evolution of the product. It may seem counter-intuitive but I do not integrate environmental indicators into the project (gCO2e, cl of water or others) because they are approximate orders of magnitude obtained from technical indicators that I integrate into the monitoring. Environmental indicators are, in my opinion, useful for a posteriori communication but are not relevant in the conduct of a project. If one follows the pillars and indicators provided well then one arrives at a very lean and optimised digital service. For example, I recently helped redesign the Commown website with Timothée Goguely and Derek Salmon. With the new version we have divided the average page weight by 22, the number of requests by 4, reduced the First Contentful Paint (FCP) time by 5 and the Time to Interactive (T2I) by 11.


Monitoring of technical indicators for the redesign of the Commown website

This is good, but is digital sustainability really about that? If we remove the questioning stage on what we digitise or not, then we have only done optimisation. Indeed, the Before/After comparison does not show the digital sustainability approach, it's more like a photo at the finish line than a video of the race. We worked with Commown upstream to remove at least half of the pre-existing content, we systematically questioned the needs expressed and the interest of digitising to meet them. This process was all the more enjoyable as Commown wanted to comply with the RGAA (French Web Accessibility Guidelines) and to take GAFAM services out as much as possible. Everything came together so the project was successful. In the end, the digital services I helped design have near-maximum scores on popular web performance tools. So is digital sustainability just about optimising a service?


Most of the web business consists of optimising things that are sometimes heavy, sprawling and the optimisation process can itself lead to a certain voracity in computing resources. If we are inspired by the Pareto principle, digital sustainability consists in a way of using only 20% of the resources given to meet 80% of the needs expressed by the users. This is where the idea of digital sobriety is expressed. This idea concerns both the design of the service, the devices and material flows necessary for its operation. The less the service calls on external resources (external dependencies and libraries, third parties, etc.) and computing resources (CPU/RAM power, network throughput), the more resilient the service will be in the event of a reduction in flows (saturated or restricted network) or in the event of exceptional events (natural disasters, cable cuts, server failure, etc.). Similarly, digital sustainability and sobriety approaches aim to reduce dependence on monopolistic services which will later have complete freedom over their pricing policy and access to their services.


It should therefore be remembered that digital sustainability must be framed by principles from which indicators are derived. Indicators without the above-mentioned principles in no way guarantee the quality of an digital sustainability approach. Moreover, this approach never comes alone, it must be accompanied by questions on accessibility, sovereignty, privacy, security, etc. One last point seems particularly important to me: designing a sustainable digital service also means questioning the IT equipment purchasing policy of the organisation in which we are working. It is probably more important to convince a large company to extend the life of its digital equipment than to design a sustainable service (unless it is an essential public service or a service with a very large audience). Ideally, you should do both, but these tasks are usually beyond the scope of a web agency.


Digital sustainability is not a miracle solution

Digital sustainability carries with it all the elements to bring about greenwashing practices: poor knowledge of the subject by clients, low level of skills of service providers, pressure to implement actions or to align to existing offers (see the new positioning of GAFAMs on the topic). Moreover, between: 1) designing a sustainable digital service and communicating about it, or 2) implementing a real environmental strategy and fundamentally rethinking its business model, the first option is much easier. The fact that Volkwagen Canada is communicating about the creation of a green website that promotes electric SUVs (1Mb for one page including 700Kb of JS) is, on the one hand, the starting shot of this greenwashing cycle and, on the other hand, the height of cynicism because this site is in no way designed in a sustainable way or even optimised (in the same way that electric SUVs are not the answer to the energy and ecological transition). This is not a critique of Wholegrain Digital as I know they didn't be involved directly in the design and development of this website, yet this is a problematic project.


Given this context, it seems to me that digital sustainability will be a lever for greenwashing policies. I have seen various campaigns that aimed at pushing large companies to design sustainably their websites (in a broader definition, this corresponds to reducing the site's power consumption and carbon emissions). In response, it would be extremely easy for Total and co. to design sustainably their digital services, but without changing their operations at all. This process would even allow them to improve their corporate communications at little cost as websites are the main access point with a company now. Thus, I do not support the principle that large companies, which only change their actions at the margin, should be actively proposed to change their digital services in a sustainable fashion.


For me, it is clear that digital sustainability only makes sense within an ambitious and coherent corporate environmental strategy and a willingness to apply it internally (via top management and/or employees). As a result, I have worked for companies that are already aligned in their environmental policy. In that case, digital sustainablility and sobriety are just following this movement. In my experience, this pair lacks the strength to drive an environmental strategy (for many reasons I cannot explain here) and are much easier to implement when they are safe "in the pack". To check the strategy of a company that contacts me, I certainly look at the CSR reports but I especially observe where the company's investments are directed, it's much more efficient to see the real strategy beyond the veneer of communication. In the end, digital sustainability does not commit a company to change its activities, which is why I prefer to refuse to work with the fossil fuel or industrial food sectors, and to concentrate on companies with a sincere and established strategy, and public administration.


The training gap

One of the problems we will have to face quickly is the lack of training on the subject. I'm pretty sure that a master's degree in environmental sciences of the digital sector would be very successful. However, to be relevant you need to be multi-disciplinary: a good knowledge of environmental sciences (LCA, MFA, ...), of the hardware infrastructure of the digital sector, and a good general knowledge of software/web development and digital design are required. Such a training requires an important mental plasticity to follow a project at macro/meso/micro level. As I do not cover all the required knowledge myself, I very often work with many colleagues to have a complete approach. Also, a trade-off is not only settled from a technical or design point of view, everything has to be thought together and this makes the exercise extremely complex.


In France, people working on the subject have generally been trained through short courses at GreenIT, INR (via MOOC), Ecoinfo or DDemain. However, 3 days are not enough to understand the subject. On the client side, there are very few people who know how to write relevant specifications or who simply know what to ask because the subject is vast and complex. Without structuring the training of future professionals in the subject, the vagueness will continue for years to come.


Advice for the years to come

Today, the couple called digital sustainability and digital sobriety is beginning a period of greenwashing, at least from my perspective. This does not mean that there will not be very competent people to meet this need, but simply that they will be more or less hidden by less competent and less relevant actors but with the means to be visible. Similarly, competent actors will refuse, out of ethics, to work with certain companies with a lot of financial means and little desire to change their unsustainable activities. Then the less competent but visible actors will respond to their demand and distort the practice. This is of course a worst-case scenario dictated by a dynamic of the lowest bidder, but it remains a strong risk, which is why I wish to state it. To make my point, I once attended, by proxy, a presentation by the international consulting firm Cybercom on "Digital Sustainability", it was simply their powerpoint to sell cloud solutions (AWS / Azure) but with a few slides on the renewable energy powering of servers. I'll leave you to admire this slide show which implies that the cloud brings emissions to 0.


Screenshots from Cybercom's presentation "Digital Sustainability - How to do IT with net-positive approach?" on 26 May 2021 (Cybercom)

All the practical guides in the world do not have the power to change the appropriation of the subject by actors with few skills and convictions or simply with a poor understanding of the issues. So in the fog ahead how do we choose who to work with? It seems to me that the first test is the consultant's ability to question the relevance of digitisation within the framework of the project. This allows you to sort out the disguised cloud vendors but also to sort out the companies that do this for PR. Secondly, be demanding on the definition of the principles guiding the approach, if indicators are proposed without a well understood and binding framework then we will be pedalling empty. Furthermore, if the other virtuous parts of a high-quality digital environment are not integrated then some of the effort is wasted. Finally, one person is not enough to carry the weight of such a mission and the constitution of a team is generally necessary to cover all the issues (environmental science, strategy, design, development, accessibility, DevOps, etc.). So an digital sustainability mission potentially opens up the possibility of collaboration and the transmission of knowledge between peers (both internally and externally). In short, we will have to fight against the fog / confusion by continuing to open up knowledge and find the right allies to pool the energy needed to make a high-quality digital environment happen in the years to come.