I’ve been thinking about something that came up in a project I’m delivering at the moment (it’s funny how these little moments sit with me for days afterwards).

I’m working with a tech SaaS client on building the “practices” component of their bespoke business operating system. I was leading a workshop exploring how decision-making needs to change as they make a fairly rapid push from 35 to 95 people over the next year.

I gave the participants the opportunity to share their experiences of decision-making in the first exercise of the workshop. I was surprised by the level of frustration expressed by the two founders. They felt they’d given their leadership team complete autonomy but things kept drifting (projects, decisions etc).

That lack of impact of what they perceived as having given autonomy was making them question it as a concept. As a bit of sidebar, I would always challenge the notion we can give autonomy. The way I see it we can only create the conditions for others to take autonomy – it’s subtle but important nuance.

Anyway, back to that moment sticking with me….the reason it kept zinging around inside my head is because autonomy without boundaries isn’t autonomy. It’s just chaos. I know this from personal experience because as an MD back in 2015 I created that exact situation.

But fouders keep making this mistake. In tech circles especially there’s endless talk about autonomous teams as if they’re a platonic ideal. But no team is ever fully autonomous and no team lives in a vacuum.

Let’s think about a Premier League football team for a moment. These are some of the highest-performing teams on the planet. They absolutely need to be empowered to self-organise during a match – to adapt to what the opposition is doing, to spot opportunities or to make split-second decisions that no manager from the touchline could possibly dictate in advance.

But their autonomy exists within super clear boundaries like:

• Eleven players on a team.
• A defined pitch.
• Offside rules.
• No handling the ball (unless you’re the goalkeeper) etc

Without these constraints, you’d basically have the Royal Shrovetide Football game – an early form of football dating back to the 12th century which is essentially just a massive fight. The main rules are essentially no killing and no motorised transport (for my many non-UK readers, check out the Royal Shrovetide Football on Youtube – it is truly one of those eccentric British customs that we have).

Let’s get back to it…the boundaries in football are what gives the players that autonomy on the pitch.

And this isn’t just a football thing. Teams are complex adaptive systems and just about every successful complex adaptive system in nature is bound by rules.

A flock of starlings creating a murmuration is just following three simple rules (separation, alignment and cohesion). These rules allow them to self-organise and avoid crashing into each other.

But without those rules, it’d just be a load birds flying about randomly.

So why do we struggle with this in business?

I think it’s probably because defining boundaries feels uncomfortable. It can feel like control or like you’re somehow undermining the very autonomy you’re trying to create.

But the opposite is true and that’s the key point here.

What I’ve noticed working with rapidly scaling businesses is that the teams that thrive with autonomy aren’t the ones with the most freedom. They’re the ones with the clearest boundaries around their decision-making authority:

• They know what they’re responsible for.
• They know what decisions they can make without asking permission.
• They know where their remit ends and someone else’s begins.

And crucially they know what success looks like within those boundaries.

Here’s what this might look like in practice:

Clear ownership. Not just “the product team” but specifically: “this team owns the checkout experience end-to-end, including performance, conversion rates and customer satisfaction scores.”

Decision rights. Explicitly stating what the team can decide autonomously versus what needs broader input. “You can change the UI, rewrite the code and adjust the user flow. You need to consult others on anything that affects pricing or payment integrations.”

Constraints that enable. Setting the non-negotiables that actually free people up. “All changes must be backwards compatible. Load time performance can’t be worse than before. How you achieve that is entirely up to you.”

Success metrics. Defining what good looks like so the team can orient themselves. “We need checkout completion to stay above 85% and page load under two seconds. How you get there is your call.”

The tricky bit for founders is getting the boundaries right.

If you go too rigid you inadvertently limit the very autonomy you’re trying to foster. Too loose means ending up with misalignment and inefficiency.

You’re after enough structure to provide clarity whilst preserving space for creativity and adaptation.

Think of yourself as a conductor with an orchestra. As conductor you set the tempo and you signal when to bring in the strings or damp down the brass section. But you don’t tell a violinist how to move her bow.

You create the conditions for the musicians to shine within a coherent whole.

I worked with a scale-up recently where the engineering team was really paralysed by their supposed autonomy. They had freedom to choose their tech stack, processes and for the most part their priorities. But they spent more time in meetings debating these choices than actually building anything.

Decision fatigue was killing them.

We introduced what felt like constraints but turned out to be liberating for them: a mutually consented-to defined tech stack (with a clear process for proposing additions), bi-weekly planning cycles aligned with company goals and explicit decision-making authority at different engineer levels.

Pretty quickly velocity increased significantly and people were happier.

That didn’t happen because we’d restricted them, but because we’d removed the cognitive load of unlimited choice and given them a clear playing field to operate on. They were playing Premier League football not Royal Shrovetide Football.

So here’s what I’d encourage you to think about this month?

• What decisions do your teams actually have authority to make autonomously?
• What are the non-negotiable constraints they need to operate within?
• Where are the boundaries of their ownership clear and where are they fuzzy? (look for delays or regular friction “hotspots” as indicators for fuzzy)
• Are you trying to give them autonomy without giving them the boundaries that make autonomy effective?

Autonomy isn’t about unlimited freedom. It’s about clear freedom within defined constraints.

The starlings know this instinctively. Perhaps it’s time we became a bit more starling.