IT sustainability: Why we’re still waiting for quantified emissions reductions
It is not all down to greenwashing. Genuine challenges stand in the way of many organisations implementing IT in support of their environmental, social and governance goals
That’s even though seemingly straightforward proxy metrics such as energy consumption exist – and it’s not for want of trying. Precision down to the kilowatt-hour (kWh) or megajoule (MJ) is possible, converting thereafter to CO2 in metric tonnes (mt), but things rarely play out that way in reality, he says. And, even when they do, it is often difficult to compare apples to apples.
“We have projects and examples in our portfolio,” Binkley says. “You can quibble over the numbers and calculate different ways, coming out with slightly different final results.”
Any kWh that is saved or matched with renewables must be converted to an equivalent carbon number – and every grid is different. Every country has different emission factors, varying hourly throughout the year. “That’s where it gets a little tricky, and is an issue in the US in particular,” Binkley says.
Calculation complications
Baseline mode or emissions are often taken by averaging the entire grid; residual emissions used in carbon calculations may reflect the effect of different activities. For instance, whether they have actually done some emissions-displacing action that reduced demand from a coal-fired power plant.
Energy efficiency improvements can be focused on monetisation too, diverting attention from emissions reduction as such, partly because there is no easy calculation mechanism.
“Residual or marginal emissions tend to be a little bit more reflective, generally showing less emission reduction than using the grid average,” says Binkley. “Some US utilities at best give an annual number 12 months after the year has closed, and others you get hourly data a month in arrears for almost real-time calculations if needed.”
What about green finance options? Digital Realty has issued “about half a dozen green bonds” since 2015, and as part of that reports on how the proceeds were used. “We use it for green building investments, energy efficiency improvements, and renewables investments,” Binkley says.
If 12 projects save 40,000 MWh in the next five years, that equates to an emissions amount avoided.
Yet you may not know what is going to be saved. A more efficient pump will save energy, but it will be difficult to know how often it will run over that time – and other things can happen on site with an effect that cannot be seen, he adds.
“Quantifying carbon emissions reduction is an extra layer,” Binkley confirms. “Historically, we’ve ESG-reported an aggregate number of energy savings across a conglomeration of projects – we don’t break them out individually, because there’s just a lot of little projects.”
This means smaller or run-of-the-mill projects might be siloed inside specific in-house programmes, and if a number is published, organisations need to be able to stand behind it, Binkley adds.
Daryl Elfield, partner and head of ESG tech and data at KPMG UK, notes that many organisations only began work on emissions in 2023. Questions range from “we’re doing X and we want a very specific solution to measure emissions” to “we don’t even know where to start but we should do something” and everything in between.
“A year or so ago, clients often looked at me blankly, and now loads want to talk about it,” he says.
While monitoring and management tech can slash CPU utilisation in the datacentre by up to 30% with zero impact on overall performance, monitoring sleep and wake cycles, most customers still have datacentres going flat out, 24 hours a day.
“If your CPU can mostly be in a sleep state, you can save yourself a lot, but instead they ask what they can do now they’ve got 30% more power,” Elfield says.
The issue is complex from multiple angles, with the biggest challenge acquiring a reliable baseline to start from, despite the array of tools now available, he agrees.
Strategy versus shortcuts
Elfield says some have “kind of skipped that bit”, and more or less assume reduced usage means they will “probably be fine”. Instead, they can prefer to “pull levers” such as cloud migration or usage.
On the other hand, although government customers do often work from the sustainability strategy down, they often do not have appropriate skills.
Post-baseline calculation, the next move is the digital carbon reduction plan, and from there extrapolating to possible approaches and bedding in, Elfield adds, with cost savings accruing over time.
“The idea of having a strategy that’s tied to the overall business has eluded many. Yet, if you build that ESG requirement into your overall IT strategy, you can make progress,” he says. “It takes lots of work to get down to that level of detail, but it’s the right level of detail, and people get excited about it,” Elfield says.
The idea of having a strategy that’s tied to the overall business has eluded many. Yet, if you build that ESG requirement into your overall IT strategy, you can make progress
Daryl Elfield, KPMG UK
Mattie Yeta, chief sustainability officer at consultancy CGI, adds that not all sectors of a supply chain in which an organisation has business partners as yet have appropriate standards with which to comply.
Think about the Science Based Targets Initiative (SBTI), which is only a few years old. Its methodologies and standards are new, and as yet it does not have standards for example for oil and gas – pretty crucial for energy-related calculations.
“Standards are needed to support different industries and SBTI are creating one for oil and gas at the moment. In their defence, oil and gas is complicated and they’ve got to go through a whole process, gathering insights and feedback,” Yeta notes.
That’s before approval and publication. If an organisation within the value chain cannot provide partners with a validated, net-zero compliant standard and can quantify emissions based on that, “you can be stuck”, she says.
A lack of leadership at corporate and board level, limiting pervasiveness of emissions reduction accountability can pile on top, ultimately affecting resources available to various teams. Yeta says a true willingness to want “to do the right thing beyond profit” can also be missing.
“If you’ve got a one- or even a half-person team who’s half marketing, half net-zero, it’s likely never going do the job the right way,” she adds.
Technologies such as artificial intelligence (AI) or data insights could help to reduce burdens. Think about how many organisations expense their travel with PDF or spreadsheets: data in incompatible formats can become more granular with AI.
When you’re trying to demonstrate a 1.5 linear degree reduction in relation to a specific percentage, you may need every single time reduction to count, Yeta notes.
“AI can read it all straight, then pull it all into a table. Some have not fully tapped into what technology can do for them to help them report accurately and therefore they give up because it feels quite hard.”
But Yeta says governments should incentivise more change. When organisations have tough decisions to make in the current economic climate and challenges beyond climate change, more stick or carrot can shift priorities.
“It is the right thing to do. We’ve got a UK Climate Act, organisations have to submit their carbon reduction plans, but we’re talking about a global challenge,” she says.
The sense that ongoing resource use expansion is holding greater sustainability back is partly testified to by Benjamin Brial, founder of devops platform provider Cycloid.io.
“This can be really about infrastructures. A lot of people, just like in the IT world, speak about decreasing carbon footprint,” Brial says. “But just in France, for example, you have around 20 new datacentres under construction that don’t have authorisation to be connected because of the power they’re requesting from the electricity grid.”
Another example is the footprint of burgeoning chatbot implementations, and more transparency is required from cloud providers too, beyond the “sexy” projects they often want to promote. Certainly there has been too much greenwash, which can be quickly uncovered when questions start being asked about budget and resources.
“And they say, ‘Yeah, we have an intern’ and they’re a big company,” Brial says. “Someone may wear a hat that puts some message on LinkedIn, but when you deep-dive, they don’t have any budget.”
He adds that the good news is that people are talking about carbon footprint now, which many in the industry were not when it comes to cloud or FinOps until about mid-2023; only now is the trend becoming tangible among Cycloid customers. In that sense, it is still early days for concrete emissions reduction results, although at least now some evangelisation is happening on the FinOps side and regulation is going in the right direction, he says.
“I never saw leaders coming to us saying they want to calculate their carbon footprint for their IT department,” Brial says. “One problem is it’s really vague, and there are other problems – on our side, we focus on infrastructures, but you have people who focus on software, and that’s problematic.”
The problem “for now” remains carbon accounting globally, with software for instance using “shortcuts” when working with carbon footprint and generalised approaches involving multiple criteria such as water use and so on, rather than being very specific.
Similarly, the focus on ESG reports can divert resource from actually taking action, which is surely more important, Brial agrees.
Certainly the market is developing. “We are about to integrate the mapping with the electricity map, where you are going to visualise and get inside your calculation, the real carbon footprint, depending on the kind of electricity you use,” Brial adds. “You have 50% decarbonised inside France, but it’s not always 50%, depending on if it’s nuclear or renewable energy and stuff like this.”