At work, we’re rethinking how we represent the steps organisations take to transition away from fossil fuels powering the datacentres they use. One thing making it more complicated is that the current way of recognising people using clean energy has all kinds of issues, so there is a new, more rigourous approach being developed. These diagrams represent me thinking through how you might represent this transition from one flawed way of buying clean energy that is still worth recognising in some way, to a more rigorous one while both are in use.
Context – part of my job at the Green Web Foundation
I work at a non-profit called the Green Web Foundation, and one thing we do there is offer a “green web check” service that let’s you check any website for evidence of it running on infrastructure powered by clean energy.
You give us a domain, and where we find evidence, we’ll show it. Here’s an example of Google, who for really long time have been seen a leader in this field:
To be fair, some recent news about “Our first carbon capture and storage project” powering a future Google datacentre has taken the shine off them somewhat, but for now, the data in the public domain is better than most large organisations.
Anyway – for along time, there’s been a binary green / not green check, but in reality it’s not as simple as that. So please join me in thinking this though, as I’d appreciate input on how best to present what is in fact quite a complicated topic.
Starting out – representing how clean an electricity grid we rely on is
Right now, there is freely available open data showing us what how ‘clean’ an electricity grid is, based on how much electricity is generated by burning fossil fuels. Ember has some, and Electricity Maps used to publish data that, for the purposes of deciding where you might choose to run software was slightly more useful, because it represented some details that Ember didn’t (see the Canada example below for more details).
Here’s an example of a map from Electricity Maps, that has been set to show the percentage of energy coming from fossil-free sources around the world, over the last 12 months. In the bottom right of the image you can see a scale, listing carbon-free energy – and for the purposes of this post, you can treat carbon-free and fossil-free interchangeably. The greener the region, the cleaner the energy:
That can be a lot to take in, so, let’s look at a single place, like one part of the United Kingdom – the island of Great Britain. If we look at the carbon intensity of over the last month, it shows as 70% coming from carbon free sources.
Over the last 12 months, the figure has been fluctuating from month to month. If you averaged it out over the year, the annual figure for the UK would be around 70%.
For simplicity, let’s assume the 70% figure is for the last year, rather than showing for the last month:
Getting to around 70% fossil free power is pretty impressive for a large economy, especially when we look how much power came from coal in the UK before. The UK has been making good progress to decarbonising its grid – getting to almost 100% by 2030 might actually happen!
This would be a real achievement, as doing this on a national scale relies finding massive investment to meet all the energy demand that currently comes fossil fuels.
How does it compare to other places though? Let’s show a few others next to each other:
As we can see, different places are further along than others. However, I’ve said places because the image above is not comparing countries. Within the same country you can end up with significant differences in how “clean” the grid is. Canada is a really good example:
Why is this so different inside the same country? In Quebec there is loads of energy generation from low carbon hydropower.. By comparison, Alberta is where the Great Canadian Tar Sands are based, so this ends up being a source of fuel for generating electricity as a result the amount of fossil free energy is much lower, and it’s a much dirtier grid.
Relating this to how we buy clean energy
But none of this really represents how you and me might buy power.
I don’t have space to dive into the details, but in short, a helpful way to think about how we buy clean energy is to think about it in two main ways:
clean energy with annual matching – this is what most of us buy when we buy a green energy tariff. The total amount used at the end of a year has to be met with a matching amount of energy certified as clean to count as 100% clean. Under this scheme, as long as you have say… 200 kilowatt-hours of power generation from solar panels which only work during the day, you can still say your energy use was 100% clean even if you only used electricity at night, as long as the number of kilowatt hours used matches the number you generate.
clean energy with hourly matching – this is the newer, more rigorous approach. Here, to count as 100% clean, every hour’s worth of electricity you use has to be matched with a corresponding amount of clean generation. This means for our solar example above, we’d need to either have some kind of batteries that stored the solar power collected during the day, or some other form of clean generation. In some cases we might even need to have shaped our own use of energy to better fit when it was available from clean sources.
Let’s try representing the annual matching visually:
I’ve tried to represent this with the green “stroke” around the boxes for each region showing progress in decarbonising electricity, while keeping the fill as it was before. If we’re honest with ourselves, buying annually matched power isn’t really completing the job of cleaning up the energy we use – we’re still backed by some fossil generation, and this tries to capture that visually.
For most of us hourly matching is more in line with how we might have imagined the green energy tariff we might buy to work. But but there is a problem we need to solve now – getting to 90% hourly matched power is almost always harder than getting to 100% annually matched power, but it sounds worse because most of the time, 90% generally sounds less impressive than 100%.
In addition, hourly matched power requires data that isn’t available everywhere. We’ll likely need two systems running in parallel for a while, whether we like it or not.
How do we represent these two, somewhat conflicting approaches? Here’s how I think of it right now:
When we move to hourly matched clean electricity, a key thing is that we’re trying to move away from our generation ultimately being backed by electricity from burning fossil fuels.
I’ve tried to represent this by keeping the outer stroke like we had for annual generation, and the green fill inside each box representing progress in terms of the percentage of power met with hourly matched clean generation.
Recap: Doing nothing, doing something, doing what is needed
This has been a lot of pictures and text! Here’s the recap for something that goes beyond a overly simple binary, and tries to capture some of the key differences between the different ways of counting energy as fossil free in 2025
What did I leave out?
I’ve tried to avoid jargon like market-based and location-based carbon intensity generation here, and I haven’t touched on consequential methods for accounting for same too. This also leaves out the different ways you back up a market-based carbon intensity claim, from using a power purchase agreements, to using some less convincing unbundled certificates. These are somewhat (but not fully) addressed with a standard like EnergyTag.
My goal here is representing these ideas to people who are not domain experts, but just want to know that when it comes to using cleaner energy, they’re doing the right thing when using digital services.
Most of the time, their main levers of change they’re prepared to consider are choosing one given provider of hosted services like cloud computing over another, as they build a service from a stack of other components.
Anyway, I hadn’t seen anyone writing about how to visually represent this awkward transition period when the way we talk about clean energy is changing, so hopefully this helps serve as a starting point for how to visually communicate this change.
Does this help? Let me know, and if you want feel free to make a version of your own – here’s the excalidraw file I used for these diagrams.