Greening Spotify – my 2 years as a sponsor of climate neutral goal at Spotify

Setting climate neutrality as a long term goal

Suddenly the moment came. I was working as a technology director for the Spotify mobile platform department. It was fall 2017. My new manager Tyson wanted to kick off his new leadership team and as part of that we wanted to formulate north star goals, a common practice in a purpose driven company like Spotify. I was expecting the usual stuff from a platform department, around productivity, internal customers, speed and stuff. 

Spotify Sustainability report 2019 – the year climate became a company concern

And then Ramon, project manager for the ongoing Google cloud migrations, says something like: Google cloud is running on green energy, could we not have a goal around that? 

Is this the moment? No one speaks up. Are we ready? Could I finally find a situation where I can make an impact on an issue so important, to the world, and to me personally?

For me it all actually started in 2005, when I first read Mark Lynas High Tide – News From a Warming World (Oväder in Swedish translation), and later his follow up Six Degrees (Sex grader in Swedish translation), and not least Andreas Malms majestic Det är vår bestämda uppfattning att om ingenting görs nu kommer det att vara för sent (no english translation unfortunately, but Andreas Malm is nowadays a world renown climate expert).

It was a life changing moment. Up until that point I was more or less unaware about climate change and the dire situation. Since then it has been front and center, personally, and for my family. I remember quoting at length from Six degrees at the dinner table, but when the horror increased degree by degree I had to stop. The existential angst was maybe ok for me to bear, but for the kids beyond bearable.

During a couple of years I could almost not talk about anything but climate change. Eventually I gave up. People just did not want to hear and be reminded about what that trip to Italy really meant.

It was not that easy to adapt your own way of life either. We reduced private flying, started buying ecological food whenever possible, bought a diesel car which could be run partly on bio diesel. And I dreamed about writing that book, or doing that thing, that really mattered. At work I kept flying. Even when I joined the agile company Crisp (later to become very engaged in climate work), I kept flying, and I did not put up a fight to change the company policy. I was lurking. Waiting for an opportunity, waiting to get that feedback loop that ment: it’s ok to be concerned, you are not crazy, let’s do something.

And now the opportunity was there, right in front of me. I said something like: could we not be more bold, lets set a goal to be a 100% climate neutral! If we, as a platform department  – 20 % of engineering and owning the full runtime platform for all of our streaming –  could show the way, it would potentially have a huge impact. And by the way, north star goals are not time set, but bold ambitions, so we did not even have to take in if it was realistic or not. Everyone just chimed in. Later my manager felt he needed to check with his manager Gustav if we really had the autonomy and mandate to formulate a climate strategy on our own. We had. We now had a goal:

”Spotify Platform North Star Goal: 100% Climate Neutral.”

Finally there was a context, and I also got the mandate to be the sponsor for the goal. But what to do next? 

At Spotify that is actually pretty obvious: get the data. The product development process is very data driven and inspired by the scientific method: analyse data, gather insights, formulate models or hypotheses, test them by doing. But how could I get the data? How could I build a mandate to go into other peoples day to day business? Remember: this was not my day to day job.

First sustainability report 

In EU companies of a certain size have to produce a yearly sustainability report. For legal reasons Spotify needed to produce such a report for 2017, and I was included in that group. Just a few weeks after having formulated the climate neutral goal I was part of a network of people responsible for different aspects of sustainability across the company. Sustainability is much broader than climate, but I was happy to drive and influence that specific topic. For the 2017 report we managed to gather rudimentary data around our carbon impact, and also show progress thanks to our Google cloud migration. 

The first sustainability report, 2017.
Operations CO2
Google cloud (50% 2017, 100% 2018)0
CDN (2019)170 ton
Offices (2018)660 ton
On prem data centers (50% up to 2017)1,800 ton
Traveling (2017)12,000 ton (not including high altitude)

For me, three important learnings came out of this work:

First of all, being able to stand on the shoulders of the climate work Google had already done was very helpful. It created a positive narrative that it’s actually possible to reduce your carbon footprint. It also made it easier to talk about and explain what carbon neutrality actually meant.

In the 2017 report we could therefore refer to the promise made by Google:

”According to Google’s Environmental Report, Google has been carbon neutral since 2007, purchasing significant amounts of renewable energy and offsetting remaining emissions. From Spotify’s point of view, this means that the energy use in theG oogle data centers that we are responsible for can also be considered to have zero net carbon emissions.”

 In the 2018 report we could follow up:

”Over the course of 2018, the team completed our transition to the carbon neutral Google Cloud Platform (GCP) and decommissioned six of our seven data centers …. In doing so, TPM reduced Spotify’s carbon footprint by almost 1,500 tonnes and succeeded in making our streaming and computing platform … nearly 100% carbon neutral.

Learning more about Googles way to carbon neutrality also opened up more questions: how much energy does GCP actually use to run our operations, where does that energy come from, and what about energy consumption from the rest of the value chain, from GCP to someone playing music on their mobile device? More on this in a future post.

CO2 data from the 2017 report.

The second learning was pretty obvious: what really stood out was carbon emission from traveling, by a factor of 10. I was not surprised. The Spotify culture is very much centered around people collaborating, we also had a travel policy with more or less total freedom to decide on your traveling on your own. Having an office in Stockholm and an office in New York of equal size ment the SAS flights almost always included a pretty staggering number of Spotifyers. 

And from that came the third learning: we could not only rely on Google, we needed to know what to do on our own, we needed a strategy, or maybe even more so: I wanted to figure out how to approach this the ”Spotify way”.

Formulating a strategy

In many ways even the possibility to start to drive an initiative around climate ”bottom up” was well aligned with the Spotify values of autonomy and innovation. Anther Spotify value is sincerity; to me that means we had to own the problem, we had to do the work ourselves and connect it to real (and realistic) ambition: we should not buy ourselves out of the problem, and we needed to avoid the risk of shallow greenwashing.

Sustainability report 2018, reporting on the climate work.

You should steal with pride, but give due credits. To me Google’s main principles seemed to make a solid foundation for a sincere policy and strategy, based on the concept of a ladder.

  1. Most important is to reduce energy consumption, the less energy you consume, the less your total footprint. This goes for your own consumption, as well as properties of services and products you consume. Moving to a cloud like Google’s was a good step in this direction since cloud solutions are the most energy efficient way there is to get compute power and storage. As a matter of fact, moving to the cloud also made it clear you need cost control, since it becomes almost too easy to spin up new resources. To help the team keep their cost under control a cost insights tool was developed. Keeping down your cloud costs also more or less maps 1:1 to saving energy.
  2. Move all your ”remaining” energy consumption to renewable energy, and make sure all your providers do the same. At the time Google cloud was the cloud provider with highest ambitions on running their cloud on renewable energy, and when not possible to run on only green energy invest in bringing new renewable energy to the grid (compensating for the  amount of ”brown” energy consumed). By 2018 we had more or less moved 100% to Google cloud, our main offices were run on renewable energy, but what about traveling?
  3. Already accepting Google cloud as carbon neutral meant we accepted ”offsetting” as a method to achieve climate neutrality. You have a strategy to decrease your energy consumption, you have done what you can to use renewable energy, then and only then you find ways to compensate your emissions by bringing down emissions somewhere else. For most contemporary companies this is the only viable way to become climate or carbon neutral in the short run. And we were flying like crazy. We had to do something.
Summarizing what Google meant to Spotify climate work, 2018.

But I also knew ”offsetting” was potentially controversial. There’s something not quite right with you buying your way out of your emissions, often by paying poorer people to avoid their emissions. If not combined with a serious strategy to reduce emissions, it might rightfully be seen as greenwashing. What did activists and scientists say? When checking what the Swedish association for the protection of nature (Naturskyddsföreningen) said, it turned out they had stopped selling carbon offsets. Their policy was that everyone needs to cut down on their emissions, and not balance it off. The well known Swedish environmental researcher Johan Rockström was merciless: ”Simply: it does not work”.

At the same time I had recently read the book Drawdown – The Most Comprehensive Plan Ever Proposed to Reverse Global Warming where they argue that with a $130 trillion investment the next 30 years we can actually do it. But the money needs to come from somewhere. 

In 2017 Sweden adopted a climate law. I had read the preparatory work. One thing really stood out. Early in the 90s Sweden put a carbon tax in place. Today it is $100 per ton, and it has been highly successful in decreasing Swedish emissions. EU is also using carbon pricing to curb emissions through a system of emission rights. Among economists there seemed to be a consensus that carbon pricing was the way to go to decrease carbon emissions both by driving efficiency gains and substituting energy sources.

In September 2018 the report Exponential roadmap was published which influenced me a lot. The concept of a carbon law was introduced, which basically said we need to half emissions every decade until emissions are zero. As part of recommended policies carbon prices was put forward as one piece of the puzzle:

”The price we pay for goods and services rarely reflects the damage these products cause to the environment. Putting a price on carbon – making producers and consumers pay more for products with high emissions – would rebalance the playing field towards sustainable alternatives, under the ’polluter pays principle’ which underpins global environmental law.”

In the report a carbon price of $50 is suggested (expected to exceed $400 per ton by the middle of the century), and the report even writes about company internal carbon pricing schemes and suggests ”to use the revenues to fund investment in renewable energy and carbon sinks”.

Carbon emissions should have a price. And money (a lot) is needed for climate investments. There we go. Step three in the strategy became: We engage and invest in projects that reduce carbon emission. To support that we create a Green Fund funded by setting an internal price on carbon emissions.

Fix flying

So, flying stood out. As said, it was not that hard to guess, even before we had solid numbers. But how should we approach the problem? 

Firstly flying is one of the really hard problems to solve from a climate perspective, since the alternatives are more or less to not fly at all. There now seems to be small electric planes on the horizon, but they will not realistically help reduce emissions from flying in time. Airplane fleet investments are done with very long time perspectives, planes built today will (if not grounded) keep on flying the next 40 to 50 years. Flying not only emits direct CO2 emissions, it also causes global heating through affecting the atmosphere in different ways.  According to the latest estimates, the extra heating is equal to 1.9 times the CO2 emitted.

Up until corona the world population also kept flying more and more, so not only is it unsustainable as is, it’s getting worse. Flying is also unevenly distributed. For the average EU citizen, flying is an important part of a person’s total carbon footprint. 

Second, the company was more or less built on flying as a free resource. Looking at flying from an employee perspective the footprint was actually rather high. In 2018 we estimated a 4 ton per employee, but that was not including the high altitude effect. An average swede  has a total footprint of between 7 ton (domestic emissions) or 12 ton (including imported goods and international flying) per year. Working on Spotify almost doubled your personal footprint.

Travel data from 2018 report.

I was pretty convinced not flying at all was not an option. I just can’t imagine, at least then I could not imagine, how you could continue to build a high growth global company without flying. Any strategy had to include continued flying. Any strategy also had to include flying less. Looking at alternatives, the only existing one apart from decreasing flying was flying on renewable jet fuel, and that was not possible without heavy investments and would be prohibitively expensive (somewhere between $200 and $1000 per ton CO2).

And even if we could fly on 100% truly renewable, we would still have the high altitude effect. We therefore, I reasoned, had to make a case to offset our carbon footprint from flying. If we could somehow build awareness and engagement around offsetting as climate investments, maybe we could also create awareness around the need to fly less.

What could the first step be to build awareness and engagement? By finding a small set of people already engaged! I asked my department lead colleagues if they had anyone in their org that was engaged in climate issues; and so we started the first working group. Initially we talked, we read, we wrote memos. Then the perfect opportunity appeared: we were going to gather all the people at a multi-day conference in New York. We could not, and probably should not, stop the event. Instead we used it as a way to develop and test a method to do carbon offsetting and communicate around it. At the conference we could announce the first ever climate initiative at Spotify and why we were here doing it. We offset 4 whales’ weight of CO2 (650 ton), and created awareness around the issue (and growing the number of members in the working group). It was a good first step.

We had also created a skeleton for a method, that was later scaled out to the whole company, and identified what we needed to improve and learn more about.

  1. We had found a solid source of carbon reduction projects in, but we needed to find a better method to select and pay for the project. The first time around one of the members in the working group used his personal credit card. That would not scale. Engaging with Goldstandard was easy, and we could eventually provide input to a credit based system suitable for enterprise customers. 
  2. It was good to have a working group selecting projects, since it created legitimacy for the process in the organization. But I also felt we needed to base project selection on something more tangible, on a more scientific method. We did find that in the project Drawdown, where they ranked 80 different methods of carbon reduction on a scale of impact. We decided to focus on: Invest in bringing more renewable energy online, invest in protecting peat land, invest in protecting forests and reforestation, invest in clean stoves in developing countries.
  3. We needed to get a real number on the impact of our flying, regularly, with emission numbers directly from the booking system. The first time around we had used one of the available online emission calculators, that too would not scale. We engaged with IT and the travel department, and eventually were able to get total numbers regularly including high altitude effects.

During 2018 I presented the strategic elements and the climate investment framework for the Platform leadership team, as a first step toward achieving our 100% climate neutral goal. We got a go ahead, we had a department climate strategy and method. 

For me the next step was engaging the other leaders more in the effort, and proving we could scale out our method. I created a simple framework where department leads were made aware of their travel emissions, and where they could pick projects they wanted to support. In the 2018 sustainability report we could write:

”One Team’s Passion for Sustainability. The Technology Platform Mission (TPM) — the 500 Spotiers responsible for nearly all of our technology infrastructure — led the way in 2017 by initiating the company’s move to Google Cloud Platform (GCP) and helping Spotify reduce its carbon footprint, but their aspirations didn’t stop there. They developed a team-level climate strategy, and dedicated funds to offset their travel emissions through, which adheres to a rigorous global standard for carbon offset projects. In 2018 they invested in the reduction of 2,700 tonnes of CO2 to compensate for flying a total of almost 13 million kilometers.”

Creating awareness, and having shown we had a method to do it we were able to influence the rest of the company. In the report we also promised: 

”Recognizing we can do better, we have committed to offsetting travel company-wide in 2019.”

Climate conference

I was given the opportunity to lead this work. I wanted to use the same method of gathering a working group, but this time a cross the company. Through a combination of my personal network and reaching out to all Business Unit leaders, I was able to set up a cross the company working group, willing to invest time in running a climate investment fund and find new opportunities for climate work beyond focusing on traveling.

The Spotify Climate unconference from 2019 report.

Personally I felt we needed to approach climate strategically as a company. I even started crafting a strategy statement on my own. During the spring me and the author of that year’s sustainability report were invited to talk about our work in front of the Spotify leadership team.

Later that spring the working group arranged a full day, distributed climate conference, with invited speakers and open space sessions around climate topics engaging people. One of the working groups members took the initiative to ask our Chief R&D Officer if he would be willing to sponsor the grass root conference. He took almost a whole day off to attend the conference. In many ways I think that was a defining moment where our bottom up approach finally met with what you in the end always need: company leadership buy in, support and in the long run making it part of central company goals and strategies.

By 2019 we had really solid travel data – and 25 % reduction goal.

By the end of 2019 I could celebrate by leaving of my leadership role to a new formal leader: we had hired a Sustainability manager, and,

  • We had offset the full company carbon footprint from flying, including high altitude effect, in effect both setting a price on carbon emission from flying and doubling that price.
  • We had reduced flying by 9 % from 2018.
  • And we had an ambitious goal for 2020: ”will look to reduce our global CO2 emissions from air travel by 25% per employee.”
Some of the climate investments done 2019.

In the long run, no company will be able not to take climate crisis seriously. So much has happened the last couple of years too. I’m happy it was possible to initiate a change from within – through small steps build awareness and methods and eventually get central by in. Obviously you need to be in a company or environment that is supportive, but the mechanics to get it done are not that complicated. You can take direct ownership.

I have left Spotify, and other people have climate and sustainability as their profession. That makes me confident and happy. 

However, we are still so far from having solved the climate crisis. Much, much, much more needs to be done, not the least in the streaming space. We need green streaming for real. More on that in future posts.

Författare: pra

Contractor, trainer and author @, ex tribe lead at Spotify & agile coach at Crisp. Active Linux and OSS advocate and developer since 1995 and a former journalist and social scientist. Also loves outdoor activities and wine.