I was impressed that the event was live streamed, making it accessible to a wider audience. Looking at the tweets, I can see that it was viewed as far away as Philadelphia, which is really exciting.
My slides can be viewed here, and my understanding is that the video will at some point be available to watch on the WPLDN website. Below is a summary of my talk, as well as the Q&A.
I’m going to talk about climate change
In 2008, James Hansen, the head of NASA’s Goddard Institute for Space Studies, along with seven other leading climate scientists, published a report in which they stated that just 1°C of global average temperature rise would lead to practically irreversible ice sheet and species loss. They later updated their report to state that 1°C would be worse than expected.
10 years later, we’ve already reached 1°C of warming. The IPCC this month published its special report on warming of 1.5°C, in which they state that at the current rate, we will hit 1.5°C in just 12 years, and 3-4°C by the end of the century, leading to a largely unlivable planet if we don’t do something about it. It would be game over.
Is Earth F**ked?
That was the somewhat controversial title of a presentation by University of California geophysicist Brad Werner, who modeled a range of different social, political, technological and economic scenarios, and found that the only realistic chance of averting climate disaster will come not from politics or new scientific discoveries, but from active resistance by the people. His research found that all other factors are too ingrained in the existing economic system to affect the radical change required but that when the masses demand change and take personal action and responsibility, change is possible.
Okay, so let’s talk about websites
Following Brad Werner’s findings, it’s essential that we all take leadership, and there’s a lot that we can do within our industry as web and WordPress professionals.
Do websites produce carbon emissions?
The short answer is yes. I’ve written about this before, with the headline being the fact that the internet produces as much CO2 as the global aviation industry as a result of the huge amount of electricity required to power data centres, telecommunication networks, and end user devices.
As we consume more data, we consume more electricity, and our consumption is rising fast. Cisco predict that the internet will account for 3.5% of global carbon emissions by 2025, and Microsoft predict that by that time, data centres will rank among the largest users of electrical power on the planet.
So how do we green the web?
In short, we can follow a three-step process to reduce the emissions from the work that we do.
1. Measure our impact
It’s impossible to manage something that we can’t quantify, which is why we developed www.websitecarbon.com to provide a simple way for people to estimate website emissions and benchmark sites against each other.
For perspective, the average website so far tested produces 2.25 grams of CO2 per page view. The most efficient website that we’ve tested is davidandkay.me.uk, which rather beautifully is hosted on a Raspberry Pi in the back of their cupboard. It has estimated emissions of just 0.004 grams of CO2 per page view.
Contrast that with the dirtiest website that we’ve tested so far, which is www.thefamily.co and has estimated emissions of 57.323 grams of CO2 per page view. That means that one visitor to thefamily.co produces emissions roughly equivalent to 14,331 visitors to David and Kay’s website.
That, right there, is a powerful illustration of how big the opportunities are to make the web cleaner and more efficient!
2. Reduce data transfer
I used the case study of how we designed and developed the latest version of this website as an example in my presentation, and you can read about that in other posts that we’ve written about how we increased our website speed initially by 77% and then reduced our homepage to under 200kb.
In summary, the key points were:
- Set a page weight budget before designing any website, and get it agreed by management.
- Use lightweight fonts.
- Keep image file sizes small by using WebP format, compressing images with plugins such as ShortPixel, and avoid images altogether if possible, instead using CSS and SVGs to create visual interest.
- Keep code clean by avoiding bloated frameworks, removing libraries that are not required, recycling assets, minimising tracking scripts, and compressing code efficiently.
- Avoiding or minimising video and animations, in particular avoiding auto-play videos, and using efficient formats of animation such as CSS animation.
3. Use cleaner energy
As mentioned, web services use energy in data centres, transmission networks, and end user devices. However, we don’t have control over the telecoms networks or the end user devices by virtue of how the internet works.
We do have control over where we host our web services though, and we can choose data centres powered by renewable energy. This is easy to do, with a growing number of WordPress hosts using green energy, and the Green Web Foundation has an excellent database of green web hosts.
A greener web is good for everyone
To sum up, a greener web is not just good for everyone living on planet earth, but by virtue of its inherent relationship with efficiency, it’s good for all web users. More efficient web services deliver better performance, better user experience, better SEO, better accessibility for people on slow and expensive connections, and can provide reduced hosting costs.
Should server efficiency be factored into carbon estimations for websites?
In short, yes. Server efficiency is of course extremely important, and this has many aspects, including efficiency with which the software uses server resources, the way that server resources are optimised within the data centre, and the efficiency of the cooling systems in the data centre. In the case of websitecarbon.com, we have no way to measure this, and so like any assessment we have to make assumptions, but we would love to see data on this side of the equation if anyone has access to it.
Will the actions and tools required to make website a more efficient cost a lot of money?
Tools such as image compression plugins (see our WordPress image plugin review here), are relatively inexpensive and easily pay for themselves on commercial websites. In fact, our preferred comparison tool, ShortPixel, even has a free plan. On the other hand, the attention to detail required to strip out unwanted code, does take time, so if you’re paying someone to do it, then there is a cost especially if it is an afterthought. By taking the approach of building in efficiency by design, however, big gains can be made from the outset with no additional cost.
If we achieve efficiency by stripping out unnecessary code, should we use less plugins? If so, is recommending the use of image compression plugins a contradiction?
Yes, we should use less plugins as it can improve efficiency, performance, and reliability of a website. However, if a plugin’s purpose is to improve efficiency, then its existence is inherently justified.
Why have websites become less efficient over time when efficiency is so important?
Efficiency in any industry is inherently driven by scarcity, but in the web, we have been experiencing an era of abundance. Broadband has got faster and cheaper, and hosting has got more powerful and cheaper, meaning that there’s little incentive to be efficient. Instead, we can afford to be lazier, using bigger files and more off-the-shelf bloated frameworks.
Are there tangible commercial benefits that can be used to sell the idea of web efficiency to management?
In terms of direct financial benefits, higher traffic websites can reduce hosting costs by embracing efficiency and reducing their server load. It won’t make that much difference for low traffic sites where the hosting costs are relatively static. However, all websites will benefit from improved user experience, which in turn leads to higher conversion rates, not to mention improved SEO and accessibility.
Roughly what proportion of energy is used at the data centre versus the telecoms networks and user devices?
It’s roughly 48% data centre, 14% transmission, and 38% end user devices. Of course, that’s average, and it will vary depending on data centre efficiency, distance traveled from server to end user, the efficiency of the user’s device (e.g. phone or desktop) and the time that they spend on the website.
Our good friend and former Wholegrainer Ed Murfitt asked:
Should we factor in the emissions from the team that physically designed and built the website, including their office energy and travel? I estimate that it could be one third of a website’s carbon emissions on average.
Of course it’s super important, and that’s why we try to run our whole business sustainably. In terms of its impact on lifecycle emissions of a website, it depends on the website’s traffic volumes. For example, building a website for 100 visitors per month doesn’t take that much less time, and therefore not much less energy, than designing a website for 100,000 visitors, but the one with 100,000 visitors is going to use roughly 1,000 times more energy in use. Therefore, the proportional impact of the web team’s operations will be large for low traffic sites, and small for high traffic sites. As a general rule, work efficiently, and don’t fly around the world to work on projects that can be done remotely.
Can you open source your website so that we can learn how you made it so efficient?
Sure, we’ll be happy to. Watch this space.
Can you recommend any good links for further reading on greening the web?
Absolutely. Of course keep an eye on our blog, specifically the Planet category. The following are some good resources:
- Website Carbon Calculator
- The Green Web Foundation
- Sustainable UX Conference
- Designing for Sustainability (Book by Tim Frick)
If you are interested in this topic and would like to know more about it, feel free to get in touch. We are always happy to chat to like minded people.