Scaling our business beyond 500 with distributed responsibility, free of static structures or superiors

November 16, 2018

Read time 7 min

Reaktor is finally an adult! Yes, our 18th birthday is here. Like in any coming-of-age story, there have been quite many twists and turns on our journey throughout the years.

Naturally, it hasn’t all been pleasant. We’ve triumphed through all kinds of growing pains. For a lot of the problems we’ve encountered, the solution has been far from obvious or traditional.

This post is an attempt to analyze one of the most recent mechanisms that attempts to make scaling up easier.

What does it mean to be an “untraditional” company?

I’ve been part of Reaktor from 100 to 500 employees. We’ve transformed from a local tech player with a rebellious attitude to a global cross-functional powerhouse of designers and technologists. We’ve never played by the rules and our unique organizational model is a manifestation of this.

In the digital space, having a peculiar organization is a bit of a trend these days. Traditional hierarchies and organizational matrix models are frowned upon, while terms such as “flat hierarchy”, “autonomy”, and “freedom and responsibility” are thrown around frequently.

These noble terms and phrases are rarely explained or defined. And, to be honest, some of the statements made are not always entirely accurate. A lot of the companies waving revolutionary flags still maintain superiors and managers. A lot of them still have static structures like departments or, for example, “tribes”.

Structures like that are not for Reaktor, at least not now or in the near future. So how does the world look from our perspective? How do we organize?

Self-organization is what has optimized our quality of output

Our work focuses on design and technical development of digital products for our clients, most often in the form of projects. We build teams around these projects. The teams are cross-functional, consisting of people who can handle all aspects of the work, from first meeting to the last line of code. They operate autonomously and make all necessary decisions concerning their projects. These structures are temporary, though. Teams change constantly as the projects evolve.

In a nutshell, that is how client teams are run. Or, not run. They run themselves, producing exceptional work.  With agile principles as the original foundation, we’ve evolved and perfected our own set of methodologies.

In a broader organizational sense, things are not as straightforward. People at Reaktor don’t belong to tribes or domain departments, and thus we have no managers or tribe leaders to manage people.

This ”loose” structure gives us resilience and the ability to adapt to the changes in the market extremely rapidly. Some of the characteristics of the type of organizational model we have, and what we aspire to be, can be found under ‘Teal Organizations’ in the book “Reinventing Organizations”.  The Wikipedia article for ‘Organic organisation’ also contains some relevant thoughts.

All around, ‘Reinventing Organizations’ has given us a lot to think about in our organizational growth. It’s a book I highly recommend to anyone with an open mind.

Is self-organization sufficient when business gets big?

We take our global business very seriously. So naturally, we need to have functions for all the operations that in a traditional company are run by managers. We are attempting to scale them and to distribute the responsibilities throughout the company. Making this work in an ideal way is something that requires continuous improvement.

So, at Reaktor a client project team is an individual’s closest reference group, but since there is no other static structures that people belong to (except for Reaktor itself), we try and create mechanisms that support natural organization around topics of interest. Functional programming enthusiasts have created a group where they discuss things related to that topic. Road cyclists have created a club called Stanga around their favorite activity. The company embraces these clubs, or loose organic structures, and supports them. This helps to create the “high autonomy, high alignment” environment that we thrive in.

One concrete example of high autonomy is that we don’t have approval processes for vacations. People discuss and agree on issues like this within their teams. The principle is that decisions are better when they are made close to their point of impact.

In addition to project teams, we do have semi-static teams like HR (we call it the “hug” team 🤗). In personal matters or issues related to your job, you can contact them (or the Talent Growth team that deals with recruitment and making sure growth is healthy). I like to call these teams “infrastructure”, since they are needed to create the roads for the project teams to run on.

Similarly, leadership is not a centralized function, but a semi-static construction that is distributed throughout the company.

Size creates inevitable challenges

When we grow, some of these functions don’t scale well. A great example of this is the increasing amount of recruitment interviews conducted and personal development discussions needed. The logic applies to all infrastructural work: It needs grow as we increase in numbers.

In a traditional organization, we would either 1) hire more people just for these functions and use the time required to train them or 2) pull people out from their specialist work to do “management work”. Neither of these options is actually viable for us. The first would cripple us, at least temporarily. And for the second… well, good luck telling a world class designer that we need her to “manage” other people and not design that much anymore.

So, what to do? Enter the 20 percenter model.

Distributing responsibility with purpose

The idea for the 20 percenter model sprung to life when we were short of people that do recruitment interviews. Our ideology is that only the best recognize the best. That’s why we don’t have HR or hiring managers involved in our recruitment process, but rather help developers and designers become top notch interviewers. With growth accelerating, we started needing a lot more of them.

To make this happen, and to make life easier for our interviewing experts, they allocated 20% of their time (basically one day a week) to contribute to the qualitative growth of the company, while utilizing and developing their specialists skills in the client projects 80% of the time. This worked better than we expected. Our recruitment improved holistically and people were able to organize the activity better around their everyday project work. It also gave the people an increased ownership mentality, which naturally led to new ideas and better practices.

This approach has also helped in tackling another challenge. Since people at Reaktor have only one main reference group that revolves around the day to day work (the most important part of our company, the work we do for our clients), increasing scale creates an increased demand for spreading information and providing context on what we do as a company and why. This new group of about 20 people we now call “20 percenters” gained knowledge, understanding and context related to growth, and formed a new specialized “topic of interest” group.

So as a non-direct result of the 20 percenter experiment in recruitment, we now have a whole new group of people having important discussions about the choices we make as a company.  Growth related pains are alleviated by more people having information and understanding on growth related matters. They help to spread information across the company. The potential for further scaling, in a healthy and constructive way, has improved significantly.

How impressive is that? It makes me extremely proud of our people and how they’ve stepped up to share the responsibility of evolving the entire company.

Expanding the experiment

Once we validated that our solution worked for recruitment, we noticed that there are similar scalability problems in our regular personal development discussions. Lacking in managers and superiors, these discussions were done by handful of people, and their capacity simply wasn’t enough anymore.

So, we started applying the 20 percenter model here as well, with promising results. The benefits are similar to what happened with recruitment. Now, more people are developing new understanding around a vital topic and  we don’t have to worry about the bottleneck that is often a symptom of managers taking care of this function.


The 20 percenter model is basically a mechanism to enable the scaling of an organization through decentralizing key functions that in traditional organizations is taken care of by superiors. It’s an incredibly simple idea and a concept that works, at least for us.

It’s a good example of how in our unique situation, we are constantly reinventing the wheel as our organization changes and grows. This, I believe, is one of our key strengths and creates our most significant competitive advantage.

Want to join our untraditional company? See our open positions.

Sign up for our newsletter

Get the latest from us in tech, business, design – and why not life.