Casumo talks: The pains of autonomy and why we love it

In 2012 we set out to create a great gaming experience. Quickly we understood that it was not only about creating the product, but also about designing the environment in which the product is created. The history of Casumo can be told in many ways. In this talk, we’re focusing on how the growth and number of talent has influenced how we need to deal with autonomy. Our agile coach Jimmy Janlén shares the story of how Casumo went from a small startup in Sweden to an agile organization of 300 employees across 4 offices and remotely, and how it’s shaped the overall working structure.

Transforming into agile

Transformation into agile can happen in different ways depending on the starting point, and the end result will look totally different. There is the traditional, top-down hierarchical structured organisations who decide to adapt to agile, and then there is the flat, startup organisations who adapt to agile principles to get shit done and gain focus. The end result looks totally different, and in this talk we are focusing on the latter. We’ve sliced it up into three phases: The startup phase, the Mission phase, and the Cluster version of our organisation. That’s where we’re at today, and a lot of things has happened during the last year.

Great people will do great things when being trusted to be great

First, let’s get a common understanding of what we mean when we’re talking about autonomous teams. Autonomy does not mean “do whatever you want”. Working in an autonomous team means that you are feeling free to act, with all your capabilities, to contribute towards a collective outcome. You think, decide, and plan together as a team, within your boundaries. You work and learn together. You choose how to conduct work, how to improve, and you are part of shaping your destiny. We always remember that the intent of others are innately good –  if you’re hired, it means you’re awesome and trusted. There are a couple of core beliefs that are guiding us in everything we do, and we believe that hiveminds towards a common goal are unbeatable. Values comes first, then process and structure, and our job as an organization is to nurture our people’s inner drive and motivation. There is more to it of course, but these are a few of the things we value the most.

We’re living in a VUCA world

Sometimes you see really strong allergic reactions in people. Change without context, nay-sayers without good reasons, acts of command and control, not wanting to collaborate or simply being a douchebag. Words like “processes”, “authority” and “manager” are gonna make people cringe. These things together kind of constitutes the mindset, and this mindset is important for the success of Casumo. The reason for that is partly the fact that we are living in a VUCA world:

Vulnerable

The environment demands us to react quickly to unexpected changes that are out of our control

Uncertain

We are required to take action without certainty

Complex

The environment is dynamic with unclear cause-effect relationships

Ambiguous 
The environment has unknown unknowns forcing us to accept that we’re most likely failing to understand what is ”real” and ”true”.

This is very much true for Casumo and the business we’re in. There are tons of fierce competition and new regulations and policies constantly that we need to adapt to. We’re dealing with ever changing technology, and the fact that we’ve grown from 200 to 300 employees in a very short time adds complexity as well.

The startup phase

When we started up as a small tight team years ago, it was easy to rally around on a common goal. There was a big sense of urgency, an exhilarating and energizing time with many sleepless nights.  The lack of structure was great as it allowed for freedom to figure out where and how you can contribute the most. Structure could emerge (and be dissolved) when needed and people worked on what they were passionate about. However, when 5 people became 30, it became stressful and chaotic, with too many people working on the same thing. At this point, total flatness and  autonomy did not scale anymore.

The Mission phase

So in 2014 we introduced the mission phase,  where we had people gathering around missions. You can think of a mission as “an idea we should run after” or “a thing that we should create or do”. Some of the teams had mission facilitators, helping the people working in the missions to move forward. Think of them as product owners, facilitators, or a combination of both.  It worked great in the beginning, and it solved the immediate problems. However, over three years time, we went from 2 missions to 33 missions and 220 people. These 220 people were depending on 5 circle leaders to give feedback and the yearly salary review. Again, this total flatness rallying around missions didn’t work. Had we brought autonomy too far again? Lack of structure lead to never ending discussions due to lack of decision making principles, several competing hidden hiercies of informal leaders occurred, a lack of purpose, ownership and an overall feeling that  “it’s always up to me” became depressing for people. There were no common rules of the game – to get things done you needed connections. This kind of chaotic environment made it impossible to get a sense of progress and capacity. We were all just assuming things were being worked on when in reality no one was taking care of it. Eventually, the pains outwaved the benefits. So in the beginning of 2018, “Clusters” were considered, designed and rolled out.

The Cluster phase

A Cluster is  a collection of teams. Now, everyone belongs to a team with a clearer team boundary, and they have a supporting structure of a Cluster.  The overall feeling today when talking to people is the feeling of finally being back in the driving seat. The ease of decision making is far greater, people have regained a sense of purpose and belonging, we’re shipping like crazy and we feel there is a shared understanding of where we’re going. Gathering people in teams is a great way of mobilizing brain power, get faster and better decisions, continuous learning and adaptation, more innovation, clear ownership and responsibility, higher quality, productivity and efficiency, autonomy, motivation and satisfaction. To mention a few things. Of course, this is a work in progress and asimplification of reality. Some Clusters have made more progress than others, and depending on who you’re asking the answers will vary as there are always different perspectives.

Casumo Pulse – strategic alignment across all teams

Our company believes are manifested in something that we refer to as Company Bets. These are large scale efforts that we need to help us build and ship together as a company. The Company Bets requires the attention, collaboration and effort from every Cluster, and they trickle down to actual work and a roadmap four times a year when the quarterly OKR (objectives and key results) season is on. The objectives describe what we should focus on, whereas the key results are smart, measurable goals. Within the current OKR:s, teams have the autonomy to figure out what to do and how to solve it.  Once the OKR:s are set derived on company bets, there is a final cross Cluster alignment to make sure we’re not creating bottlenecks and dependencies we don’t want to. We call this process Casumo Pulse, and in short, this is how we create an overview and strategic alignment across all the teams.

Decentralizing authority and information

Transparency and communication is vital for us. Decentralising control means decentralizing both authority to make decisions, and the information to make them correctly. Big news and vision is shared to help everyone connect to the bigger picture through our yearly company wide conference and our monthly breakfasts, and we have tons of communication happening daily through Github, Slack, company blog, newsletters and more. We’re training leaders in supporting their teams through active servant leadership, which includes facilitating, coaching, mentoring, teaching, feedbacking and providing  tools for decision making. To be sure our new joiners are set up for success in this challenging environment, we run a company wide onboarding programme every month. Because autonomy doesn’t come for free, we do experience many challenges every day. How can we make cross Cluster alignment smoother? How can we grow a culture that is even stronger to excel in a VUCA world? How can we give comfort and confidence in constant change, evolution in our ways of working and nurture and protect autonomy while scaling even more? These are some major challenges that comes with an autonomous structure, and something we’re working on everyday. But, despite the pains that comes with autonomy, we do love it. Why?  Because it makes us reactable, helps us attract the best talent and creates room for creativity and happiness.

 

Watch the full talk here  (disclaimer: the sound quality is not the best but also not the worst.)

Jimmy Janlén is a renowned Agile expert and teacher, most famous for his contribution to the current Spotify model that has been adapted by a number of leading tech companies.

Smart pipelining — reactive approach to computation scheduling

Many tech-eons ago (several human years now), my colleagues and I started developing an in-house big data platform. The goal was to ignite the company’s digital transformation and enable people to harvest the power of data. It was a greenfield project. There was nothing except the task, the company’s AWS account and our imagination. Now, in hindsight, I can say that that was all we needed.

Being a team of experienced people, we knew that what we start with is not going to be what we will end up with.

Read more

Web Summit means 35 km on foot

Yes, it was that big. Web Summit 2018 hosted almost 70k visitors (10k is Swieqi’s population) and 1800+ startups in Lisbon from 5th to 9th of November.

So why do 70k tech people go to an offline and kinda old fashioned event after all? I don’t think that there is only one correct answer, but I think it is a little bit of everything:

  • 1200+ speakers
  • See some celebrities on stage
  • Free coffee and some other freebies
  • To be visible to VCs
  • FOMO
  • Networking opportunities day & night
  • Pastel de Nata
  • To make some observations about tech trends
  • A trade show for some companies and countries to attract talent

I was there to get a gist of the tech ecosystem’s whereabouts and get some fresh insights. There were 24 tracks to follow on more than 12 stages, excluding the off stage events from companies like Google, AWS, etc.

So it takes a lot of walking and running in between the stages. If you plan to to get there in the following years, make sure that you go there with a group of people and make a plan weeks ago to get the most of the event. Of course, you can always watch the sessions you missed on YouTube, but going there as a team, doing a daily sense check and investigating the venue together can be really amazing.

Unfortunately we could not get organised with other sumos on time and we all flew solo, but it may be even better to share our observations here on the blog to reflect the unique perspectives and/or challenge each other’s point of view to build on an insight. Here are my two highlights to start with:

Nothing is new while everything is changing

Machine Learning, AI, Blockchain, the art of attracting investments, the cost of growth… These topics have been exploited in almost all events within these last five years. So it is still going on: Some do not expect Blockchain to impact common people’s lives not in less than 7-10 years, some already announced that they have been talking with authorities in Malta about citizenship for the robots using SingularityNet platform on blockchain.

Data conversations were everywhere. The areas that AI and Machine Learning are being used is expanding enormously. From automotive to fashion, from content creation to GDPR… everybody is after the ways to be much smarter with rules, algorithms and create lots of content accordingly. Therefore according to IBM; only in USA there will be 2.7 mio positions around data sciences by 2020. In my humble opinion, the threat and the opportunity is the same and it is pretty clear in this case: In a world where data and content is abundant, what can make your content or your product to be more meaningful or more appealing to the audience? If relevance will not be optional but the norm –  what will make your sales content stand out and be appreciated?

There may be many answers to that question but Netflix’ CPO Greg Peters shared very nice insights about ‘relevance’. He made us realise how narrowcast Hollywood’s production was -and still is- when you consider the interest areas of whole global population. He said their local productions like 3%, El Casa de Papel and Dark showed them there are much more people to cover than Hollywood’s target audience as native English speakers with a western life style which represent just 5% of the whole world. And this is not the success of localisation only, because all these productions were consumed globally as well as USA productions like House of Cards, Stranger Things or Orange is the New Black…

Trust is scarce on every level

There were three major trust issues mentioned over and over which seem to have a massive impact in tech ecosystem:

  • Respect for Privacy: Facebook was not there and people talked behind their back a lot. They were questioning their business model which is abusing data to suck traffic into Facebook and its consequences. The need for a wind of change was evident with lots of emphasis on building trustworthy platforms and ideas. Next Generation Internet is being seen as necessary next step to create a sustainable future internet.

 

  • Mainstream Manipulation: We heard lots of “Fake news” conversations. It is not only about a country or any elections, but people are well aware how technology can be abused to create a public opinion and/or feed polarisation. Media representatives and journalists -as expected- claim to be very responsible about it, they talk about content safety a lot. On the other hand, the amounts of content that they have been publishing every minute is getting more and more. It is already mind blowing to hear that even mid size content creation companies can manage to test thousands of contents for their clients every week. So with such a scale, media need to trust the ‘machine’ to choose well for them among all that ‘vomit’ at the end of the day. And we need to trust them to do so?

 

  • Authenticity: This is more of an introverted topic compared to the first two. Most of the speakers who represent giant tech companies were putting a very honest auto-critic about the wrong decisions they made for the sake of growth. Not very openly, but they were referring to the unbearable pressure coming from the investors for growth and its unpleasant consequences. They were advising the new startups to bootstrap as much as they can, and choose a slow growth to be able to chase their dreams. So sustainable growth, sticking with the values, focusing on learning first to be able to replicate success, building stronger brands… These messages were very popular. An interesting case about authenticity was coming from Slack. I didn’t know that Slack was a by-product of a games company. So the people who created Slack were working together to design and deploy games, but not an enterprise communication software. Their story is very inspirational to see how entrepreneurship functions, and to understand how creativity and fluidity can help a culture to focus on solving real problems rather than just delivering what they are supposed to deliver.

And a quote from Slack’s Cal Henderson also summarises the whole event real well. He said -more or less-: “It is very easy and it is a trap to find yourself being really busy. You’re probably too stuck in tactical, day to day work, and you’re not focusing enough on the long term… Even if long term means just three months from now.”. And remember, this comes from a leader who was supposed to deliver a game not Slack.

So some of the tech people are too busy to deliver what they need to deliver to please the Investors. And some of them have the luxury to choose to focus on real problems of real people to get humanity to the next level…

 

Characteristics of an awesome Data Crew Culture (or reasons why I love working remotely)

I am often asked what it’s like to work remotely, mainly by young people working in corporate environments. The efforts those organizations make to appear flat or very agile do not matter once you join, you start understanding how things really work. If you’re cool with that, all good. But if you don’t like it, there is an alternative. This is why I want to share my experience and insights from my first 100% remote position and why I can’t imagine myself working differently in any close future.

I am a data engineer at Casumo, and I’ve been enjoying my time here since I joined, almost two years ago. Back then, data team had just been created. Today, we are ten people with seven different nationalities, working in five different countries, which possibly makes us the most distributed team in the organization. Our way of working together offers us a certain level of freedom and flexibility, but also demands discipline and professionalism, something that all the team members are aware of.

Mutual respect

Given the different nationalities and cultures, our team could have faced serious communication issues. But the reason we haven’t is trivial: mutual respect. We grew our team around individual cultures. We all know the typical time of Spanish lunch, the dates of Orthodox Christmas or Estoril Open, and we plan and organize accordingly.

Communication

People in this organization have true, human and humble core values and communication is something everyone constantly strives to improve. Every interaction starts as an asynchronous one – if the person on the other side of the channel can answer immediately, good; if not, there is bound to be a good reason for it and he or she will get back to you.This is very important to note because it underlies what comes next: we don’t waste each other’s time in meetings.

Effective meetings

When it comes to meetings, we are trying to stay lean. A weekly meeting is enough to understand who is doing what, and why. If more discussion is needed, it is taken outside the group meeting. Direct communication between fewest people works the best.

Knowledge sharing

When mutual understanding was strong enough, team culture that fosters knowledge sharing has grown naturally. Now there are weekly Show & Tell sessions in which team members present… well, whatever they feel should be presented. It could be a pet project, their latest work on a product feature, an introduction into a topic unknown to others or a concern-raising session. These S&Ts are great opportunities for receiving and giving feedback — a good chance for the team to mature. In addition, our culture is also knowledge caring, if I can put it like that.We motivate and challenge each other to work on new interesting stuff, to broaden horizons and to dig into more details or raise the quality. Considering the tech stack we are using, the challenges we face everyday, increasing data volumes and opportunities those bring in combined, we shouldn’t be doing anything less.

Team Mantras

In regards to our approach to software development, there are two mantras we live by:

  • “No code is good code.” — This is what makes us tidy. If we can get rid of a service, component or a piece of code, we go for it. It makes our systems less complex and our lives easier.
  • “If there is a chance it will be run more than once, automate it.” — Speaks for itself. Automate it to “forget” about it.

Culture cannot be enforced. It has to be built, developed and nurtured. Changes should be made over time, with care and with a clear understanding why. That’s what we strive for and why I enjoy where I am. The data crew culture resembles the organization as a whole, with a few special characteristics, and it would not work this way if Casumo wasn’t already built on solid principles and values.