Organizational Structure Part 1: How do you preserve autonomy at scale, while maintaining clarity?
A lone bee has very little chance of survival. Each hive has thousands of bees collaborating to build a sustainable colony. And the misconception of the queen bee ruling the colony. Within the colony, each bee is born with a specific role — with different jobs over the lifetime. The role is defined by the type of bee. There are three types of bees — Queen, Worker, and Drone. The role is based on sex. For instance, a worker bee(female) starts by cleaning the cell from which it is hatched, nursing the young, scouting, communicating, collecting and storing food, to removing herself from the hive sensing the end of life. The Drone bee (male) mostly feeds, waits for mating, and dies immediately after mating. The queen is the only one that produces bees decides the sex of the newly born bee. It is a good example of a decentralised system, which means there is no bee organising these activities (not even the queen bee). Most of these happen by hormones kicking it at different stages and magic of nature.
Bees have a clear definition of roles and decision frameworks. It enables autonomy and empowers the team to work efficiently without any bottlenecks.
A good organisational design is never based on a formula
A successful structure that works for 10, will not work for 50, the proven one for 50 will not work for 100. It requires ongoing attention and must adapt and evolve over time. A well designed organisational structure brings clarity to roles and the interconnections between them, enabling alignment with core values while still granting autonomy. It makes sure any change in organisational strategy is reflected across the roles and functions to be aligned with the new objectives. It provides operational flexibility and reduces bottlenecks.
Basically, it closely represents responsibilities and the way things work.
Why not hierarchy?
A common or familiar structure is hierarchal (traditional). Time and again, people report to people and there is a flow of commands from top to bottom.
In a few scenarios, it is actually a good model. It gives an easy representation of the career growth path and the phrase growing up the ladder. In a few places like very large organizations, it is like what Winston Churchill talks about democracy: “Democracy is the worst form of government, except for all those other forms that have been tried from time to time.” Also, it is easy and popular, it is highly adopted. It has become so popular that it can be found as a presentation template everywhere, you fill those boxes and voilà you have your organisation structure. Many people adopt this as they underestimate the importance of org structure or simply don’t know how to do it or many times people do not dare to drive the change.
There are a few limitations — it cannot represent a person playing multiple roles, it limits one person to report to one person only, sometimes the message is lost in translation down the chain, and relevant information can be dropped to serve self needs. In some cultures, the view represents command and control rather than enablement and support. It is a model that is not at all consistent — you play Russian Roulette and hope not to get a bad manager. So, clearly, it is not always the right model.
EverestEngineering Case Study: What worked with 15 people won’t work for 150
At Everest, when we started more than five years ago with three people and eventually grew to twenty people. We never required or represented organizational structure. A few of us wore multiple hats. But we were waiting for the organizational design to come and knock on our door one day. Everyone knew their work and who is responsible for what (even with multiple hats). We were like the bees (born with the roles). We were fast. You could spend a day with us and get a snapshot of the company.
We were scaling and somewhere between 25–50 people (90% in India and 10% in Melbourne), we started hitting a few bottlenecks and delays in the decision-making process. We realized we were more humans than bees. So, we started by defining some roles, responsibilities, workflows and communication methods. But still, the org structure looked something like this.
This structure looks simple, but still, it is complex, since you cannot get a view of responsibilities by watching this diagram. But still, it was helpful till we grew to around 70 people. You can look around, ask people and then understand (or at least get an idea). There were some micro-groups — smoking, lunch, cricket, project-based, going out for beer; that helped people to discuss and connect, which eventually helped to learn the virtual org structure. Some of our teams were in Melbourne and the remaining in India. It was still okay because it was easy to get the sense of responsibilities within a few weeks of onboarding.
Then the bat soup thing happened. This forced all of us to work remotely. This did not immediately affect us based on the org structure. Since we all had an idea of who is responsible for what. And for a few months, we were not hiring. Then, an interesting thing happened. We started expanding with most of the team members joining remotely. What we learnt soon was that it was very hard for a new person who joins remotely to get a sense of the organisation’s culture, structure, and responsibilities. Their perspective was based on the video calls that they attended, most of which were transactional.
When people had questions, it ended up coming to a few sets of people and it ended up naturally creating bottlenecks. And interestingly we were trying to double ourselves in a year, 70 to 140. It organically demanded team leadership and demanded functional specialization. We hired people who were responsible for some of the functions and some team coaches. This helped to scale the organization. This still did not help people get a view of the org structure.
When scaling, the very traits that make teams great can often work to prevent their coherence into a broader whole.
Accountability and Autonomy
Autonomy as a choice to make your own decisions. Autonomy is not a free ticket to take any decision. Unchecked autonomy can lead to ambiguity, inefficiency, and chaos. It works well with boundaries and expectations. The choice becomes efficient with better information and context.
We wanted to have two fundamentals in our org structure:
- Free choice (Autonomy)
- With the growth, maintain order (clarity)
What does that mean?
- Individuals are treated as adults.
- They have clarity on their job functions, roles, and growth.
- Teams have clarity on the expected outcomes.
- Cross-functional teams function smoothly.
So, with the scale of growth, we wanted to have a structure which is promoting free choice for teams and individuals, aligning with the organizational culture and goals. We looked into the existing structures, both popular and unique ones.
We created a quadrant of free choice, control, chaos, and order; and placed all the known models in that. We looked into these models with the dimension of scaling teams (matching the rate of our growth). We aspired to be in the free choice and order quadrant — with our scale we wanted to have a structure that gives autonomy but still brings in clarity. The free choice-chaos quadrant was right was a good model for us when we were very small, it was not working with scale.
“Old ways won’t open new doors.”
Will continue in Part 2.
