The Five Tenets of Autonomy

In a world where companies seem to copy one another's mistakes, I feel the need to argue for what you need to build successful autonomous product teams at scale. You might agree or disagree, and you might even be provoked regardless of your degree of agreement.

Certain things need to be present in people's lives to thrive; Purpose, psychological safety, and autonomy are three. There are several books on those first two words, but I struggle to find anyone who can define what is needed to succeed with the third word; Autonomy.

To make my point, I need you and me to try and agree on the actual meaning of two words, so that we have a shared understanding of their meaning. Let's begin with the word autonomy (/ɔːˈtɒnəmi/) "the right or condition of self-government", often misunderstood with the word anarchy (/ˈanəki/) which means "the absence of government and absolute freedom of the individual with no surrounding controlling systems."

Then let's look at the word competent (/ˈkɒmpɪt(ə)nt/), which mean "having the necessary ability, knowledge, or skill to do something successfully." often misused by people with a high degree of perceptive knowledge, who should be classified under the contronym incompetent (/ɪnˈkɒmpɪt(ə)nt/) "not having or showing the necessary skills to do something successfully."

Humour me, and let's put these words together in a matrix; Competent autonomy vs. incompetent anarchy and incompetent autonomy vs. competent anarchy.

Now, I know you would think that this is a hypothetical question, but which one of these would you say is an ideal combination? Suppose you're rooting for the competent anarchy. In that case, you should consider joining or starting a political career with a manifesto built upon anarcho-syndicalism and rejoice with the likes of Errico Malatesta. (I'll leave it up to you to stop reading this article and instead read up on his work in your own time). Those that landed on competent autonomy keep reading. If you landed on any of the other alternatives, I can't help you.

Before we learn, we are all incompetent. It's why we learn. There is no shame in admitting one's own incompetence; it's actually a necessity to start learning. A well-known theory in psychology is the “Dunning-Kruger effect”. It talks of the lack of ability to admit to one's own incompetence.

This theory describes a cognitive bias whereby people with limited knowledge or competence in a given intellectual or social domain vastly overestimate their expertise or competence in that domain relative to objective criteria or the performance of their peers or of people in general.

Let's call this the first tenet

#1 Autonomy requires competence

What I'm saying is simply that to be truly autonomous, you also need to be competent. This includes being aware that autonomy requires a high level of skin in the game, also known as accountability — if a pig and a chicken join forces and starts a restaurant that serves only bacon and eggs, the pig has quite literally "skin-in-the-game", which is the fact or condition of carrying the responsibility of your actions. In this case, I would dare to claim that the pig is far more committed to the restaurant's success than the chicken.

When shit hits the fan, and someone needs to be held accountable for a mistake, we tend to feel the need to point the finger at someone or something—either an individual or a system. Now, if we juxtapose autonomy and anarchy, you will see that there is no defined role of leadership or the presence of a surrounding controlling system. So there is no one to blame or to point fingers at; hence there is no accountability. In autonomous teams, where accountability is spread across its members, everyone is on board with the concept of communal responsibility. Successful autonomous teams create space for mistakes; it lends room for the success of failure and supports its members grasps for the unknown, and allow room for error. People feel comfortable messing up when they know they won't be punished for it. If one fails, we all fail. In other words, autonomy requires a very high degree of skin in the game.

Autonomy is a situation in which the learner is totally responsible for all the decisions concerned with his [or her] learning and the implementation of those decisions’.

Leslie Dickinson

Let's write that down as our second tenet.

#2 Autonomy requires a high degree of communal accountability that lends room to the success of failure.

Let's start with creating a direct link to the second tenet above to get to number three. Accountability is a necessary component to achieve trust, and trust is paramount to succeed with autonomy. I could stop there and state that as the third tenet, but I won't.

A startup usually requires fresh capital from investors, and these investors now represent a controlling system. If you don't want a controlling system, all members of the team need to equally carry the cost and risk on their own. This entails that everyone trusts one another and that the people you choose to share this risk with has the necessary ability, knowledge, or skill to pull this off successfully. That is a direct link to tenet number one.

Controlling systems can be many things, but they all share an attribute, a certain degree of control and a balance of trust. Control and trust are two words that can either be seen as direct opposites or as I like to see it, dependencies. Trust helps to reduce the need for control, and the lack of trust does the opposite. Where there is a lack of competence, the lack of trust quickly ensues, we tend to build strong controlling surrounding systems or frameworks to compensate for the lack of trust. These frameworks usually reduce a teams ability to turn quickly and react to change and room to fail.

And that leads to my third tenet.

#3 Autonomy requires a balance between control and trust, scaffolds, not frameworks

Instead of frameworks, we need scaffolds. A scaffold filled with methods, tools, good practice, teachers, and coaches will help teams achieve a necessary degree of competence to become autonomous. Frameworks are permanent structures designed to control and prevent mistakes and do not adhere to levels of competence. Highly competent people trapped in an encumbered framework with no room for error quickly becomes frustrated, and such frameworks don't induce learning. In contrast, when we build scaffolds around a building, we partially remove the scaffolding as the building takes shape — as we do when we try to build competence, and we lay the foundations for a learning organisation. This theory is the basis of all educational institutions all around the world, scaffold learners to be autonomous learners. When the necessary degree of competence is achieved, an equal degree of trust and autonomy should follow; hence we remove the supporting structure that surrounds it. Companies that succeed with this is what we like to call "Learning organisations". These are organisations where people continually expand their capacity to create the results they truly desire, where new and expansive patterns of thinking are nurtured, where collective aspiration is set free, and where people are continually learning to see the whole together.

But, be warned, scaffolds also hold a certain degree of control that usually comes with power. And once you have power, letting it go can be a painful and challenging exercise.

Give a man a fish, and you feed him for a day. Teach a man to fish, and you feed him for a lifetime.

— Chinese proverb

#4 Autonomy requires a well-defined problem and system-oriented thinking.

A solution starts with the writing of the problem. Charles Kettering, the famed inventor and head of research for GM, said, "a problem well-stated is half-solved."

A well-defined problem is not the same as having a common direction. A common direction points to the desired end result, a description of a change in human behaviour that is required in order to succeed. And this is important as well, but the concept of a common direction needs to be extended to include longitude, latitude, depth, height and width, it doesn't help just describing your end goal.

On Youtube, I found this clip from a car accident in Poland where a car flies straight over a roundabout at high speed. Now, this clip works as a perfect analogy for the most commonly stated argument for autonomy, that the main benefit of autonomy is speed. But speed is second to direction.

There are many things you miss when you drive fast or run as fast as you can, you seldom look for gold nuggets on the ground whilst running intervals. Teams act like they want to reach the end goal as fast as possible, and we applaud ourselves when we manage to fail fast in the process. It doesn't help to fail fast if the problem you're trying to solve is the wrong one in the first place. I've never seen anyone being able to iterate themselves towards the right solution based on the wrong problem. In other words, make sure you're building the right thing before you focus on building things right.

Being able to be thinking in systems is paramount to understand why autonomy without these tenets won’t work at scale. I recommend reading “Thinking In Systems” by Donella H. Meadows to understand more on the subject matter.

#5 Autonomy requires leadership

Spotify is often used as a successful example of fully autonomous teams. The less known story is that Spotify actually tried and failed miserably initially, leading to a disastrous situation from which it took them years to recover. Yet companies worldwide have copied the "Spotify model" straight out of the box, despite that Spotify themselves tried to warn others to do so.

Full autonomy in a startup team without any clearly defined leadership present can be a great solution, as long as everyone clearly understands the rules of the second tenet; Autonomy requires a high degree of communal accountability that lends room to the success of failure. However, in larger companies and complex organisations with multiple interdependent teams at scale, it's not.

As they grew in size and complexity, Spotify failed to identify that fully self-leading autonomy, combined with higher levels of complexity and scale, leads to fragmentation, both for the business and its users. Instead, Spotify was left with small independent cogwheels spinning fast, non-aware of one another, solving fragments of a problem with no common direction.

So, when I state that autonomy requires leadership, I'm talking of what has become known as "servant leadership". In autonomous teams, leaders serve the team, not the other way around. As leaders are seen as those who make the decisions, they need to be listening. They should strive to understand and empathise with others. They should be healers who bring about transformation and integration— holding personal traits such as awareness, persuasion, conceptualisation, foresight and stewardship.

Great servant leaders are those who see to it that all the other four tenets are present.