We often encounter leaders who talk about “flattening the hierarchy”. The assumption is that, if they can smash the rigid pyramid, allow anyone to use whatever tools they like, and tell everyone to move fast, a productive, innovative, happy organization will spontaneously emerge.
Then reality inevitably steps in and punctures that assumption. This happens because the ‘flat = best’ policy ignores an essential idea: cybernetics.
Before the word ‘cybernetics’ was hijacked by sci-fi authors to describe lasers and glowing retinas, it belonged to the physicist and mathematician Norbert Wiener. Derived from the Greek word for steersman, cybernetics is the study of control and communication in complex systems. It’s about feedback loops and how a system senses its environment, calculates the gap between where it is and where it needs to be, and corrects its course.
The best argument for flat organizations is that they should be good cybernetic organisms: highly responsive to ever-changing conditions. They distribute authority to the ‘edge’ because they want to avoid the centralized analysis and decision-making that creates too much lag. They favor ‘edge computing’.
But edge computing only works if you are careful about how each bit of the system is scoped. Should it favor exploration (requiring loose boundaries and high flexibility) or exploitation (requiring tight constraints and hyper-focus)? The fatal flaw of the modern flat org is treating flatness as a permanent vibe rather than as a dial that needs to be tuned constantly as the world changes.
A team optimizing a mature, core product requires completely different structures and norms compared with a skunkworks unit trying to discover the future. The job of leadership is to detect quickly whenever a part of the system has its tuning wrong and help drive it to a better setting.
Detection
So how do you actually tell when a team’s dial is pointed at the wrong setting?
When a team is tuned too loosely, the dominant effect is high entropy. You’ll notice an explosion of meta-work. Decisions get dragged into sprawling, multi-party loops where everyone has an opinion and it’s often unclear who has the final call. The team produces plenty of heat (vision docs, strategy workshops, and Slack threads with fifty-plus replies) but very little light. The telltale signal here is decision latency (sluggishness over how to execute).
On the other hand, when a team is tuned too tightly, the symptoms are quieter, but arguably more dangerous. The key effect here is brittle efficiency. The team hits 100% of their quarterly targets, yet their actual impact on the business seems to plateau. Give them a clear specification and they’ll execute it with almost robotic precision and (usually) a decent level of quality. However, hand them an ambiguous problem, and you’ll see them sputter into paralysis. Such a team will give the sense that novel ideas or unexpected customer feedback are annoyances that threaten their throughput. They are running a remarkably tight ship, straight into an iceberg. What you’re looking for here is contextual blindness (paralysis over what to build next).
Spiking latency tells you the team’s boundaries are too porous and need tightening. Contextual blindness tells you their boundaries are too rigid and need room to breathe.
Action
Once you’ve identified the direction of tuning needed for a team, resist the urge to deliver a pep talk or drop another inspirational slide deck into their lives. Cultural vibes don’t fix structural misalignment.
If a team is bogged down in decision latency, your job is to shrink their world and muffle the background noise…
Define who actually holds a veto versus who is merely being kept in the loop.
Assign a single decision owner.
Shut down non-essential, cross-functional check-ins.
Insulate them from broader organizational concerns.
“For the next month, your universe consists solely of 6 people trying to achieve this specific outcome. Ignore everything else.”
Unsurprisingly, if a team is suffering from contextual blindness, you do the exact opposite…
Deliberately throw a shoe into their neatly ordered machinery.
Loosen their deadlines and ask them to plan over a longer timescale.
Replace hyper-specific targets with broader, exploratory goals.
Put team members directly on calls with customers (the more disgruntled, the better).
Ask the team to defend their assumptions to outside groups that don’t share their jargon or experience.
Encourage them to spend a chunk of their bandwidth on experiments that stand a good chance of failing.
“For the next few months, run out and explore the forest with as much curiosity and as you can. Drink everything in.”
We had this exact situation arise with a team we supported in 2024. Their leader was obsessed with quality and predictability. Everything had to be just-so, and it worked until the world stopped asking for the same thing. Then, in their market sector (almost overnight) AI features went from ‘interesting toys’ to ‘absolute dealbreakers’. They didn’t even know if they liked or approved of AI in general and certainly couldn’t imagine tearing up most of their roadmap to accommodate it. So we worked with their stakeholders to remove almost every near-term goal the team had and instead give it a singular mission to invent a new product based around the features customers were clamouring for. They were also given strong encouragement to create an internal “AI University” where other teams could come and exchange ideas without the pressure to deliver something immediate. Within a month, the team had shipped a working prototype and other departments were asking to shadow them.
Whatever the situation (too loose or too tight), the way you support such teams is fundamentally the same. You tune the strictness of their inputs, the horizon of their metrics, and the permeability of their walls. Once that’s done, you can get out of the way and go back to monitoring the signals again.
Steering True
Asana’s Anatomy of Work Index revealed that knowledge workers spend 60% of their time on ‘work about work’ - all the coordination and collaboration stuff needed to get a network of people to deliver real value. Ultimately it’s this reality that forces us to be smarter about the flatness of organizations.
An unstructured organisation will find that coordination activities expand uncontrollably until they occupy most of its energy. And yet a rigid organization has great difficulty knowing what to do with the energy that it has.
So take a look at your organization’s telemetry this week. Look past the vanity metrics on the project boards to the actual patterns of friction where your people are burning their calories. If you see a team suffocating under decision fatigue, build them a wall. If you see a team executing flawlessly toward irrelevance, cut them a door.
And if your teams could use a better read on their own dial, take a look at our leadership program.