Space Rockets, Mushrooms, and Button Factories

Choose the Right Business Growth Model

What we kill makes us stronger. How fast a company can develop depends not just on its innovative ideas but on how much chaos and disruption its leaders are willing to embrace.

Before crafting your strategy, decide what to destroy, when, how quickly, and who will make the final call.

You can choose from five growth models.

1. Button Factory

2. Conveyor of Button Factories

3. Space Rocket

4. Conveyor + Lab

5. Mycelium

In four, you decide the pace and nature of the necessary destruction in advance. In the fifth, you’ll focus on creation; others will destroy.

The fifth one is new, and few leaders have even heard about it. But I’m convinced the future belongs to it – and here’s why.

Button factories don’t fly into space

In 2009, I revamped my company’s strategy with a shiny new Big Audacious Plan. By April 2010, we were embarrassingly behind schedule. I blamed my executives. They blamed the market and overwork.

This is an exclusive additional monthly newsletter for paid subscribers. Upgrade to paid to read the full article.

The real reason was that I wanted my company to be both a champion of operational efficiency and a rocket-fueled startup. As you’ll see soon, I was trying to build a perfect “button factory” and launch it into space at the same time. I wanted to build but was unwilling to destroy.

But in the real world, building is impossible without destruction.

A forest has two purposes – to survive and grow. To fulfill them, it must change – and change demands death. Trees fall, and new ones rise to take their place. Years go by, and it’s technically a different forest from what it was 10 or 20 years ago. The trees, animals, birds – they’re different now. And yet, it’s still the same forest.

Business follows the same rules. Profit, market share, and market cap are just small markers along its longer path. Its purpose is the same – to survive and grow. Products “die,” people leave. Over a decade, the entire staff, including the leader, might change. Technically, it’s a different company – and yet, it’s still the same.

A business has to change, and change demands “death.” As a leader, you must be the driving force behind it. You decide what to preserve and what to sacrifice. To make room for new products, processes, or ideas, you have to ‘kill’ the existing ones. And the speed and direction of change determine your growth model.

A business must both live and die to grow, and its leader must master the art of balancing life and death.

Four growth models

The balance between what a CEO “kills” and what they “keep alive,” along with the direction of growth – qualitative or quantitative – defines the growth model.

Take a look at the image:

1. Button Factory (Low destruction, incremental change, focus on efficiency)

Examples: Toyota, McDonald’s, Bic Cristal
Approach: Fine-tune operations for reliability and efficiency
What gets “killed”: Very little—change is slow and careful
Works best for: Established markets where consistency trumps innovation

The Button Factory is a well-oiled machine that values stability. Products evolve gradually, prices remain fair, and customers know exactly what to expect. The CEO rarely needs to eliminate existing elements to make room for new ones.

2. Conveyor + Lab (Moderate destruction, focus on qualitative growth)

Examples: Apple
Approach: Keep manufacturing/operations stable while innovation happens separately
What gets “killed”: Products and processes periodically, but in a controlled fashion
Works best for: Product-based companies that don’t need physical presence at points of sale

A Conveyor + Lab (bottom right) is both a “button factory” and a startup. But instead of mixing the two (like I did in 2009), it keeps them strictly separate. Apple is a conveyor in manufacturing and shipping, a mad scientist’s Lab in design and new products.

3. Conveyor of Button Factories (Moderate destruction, focus on quantitative growth)

Examples: Early Uber, chain retailers
Approach: Perfect a single operational model, then replicate it globally
What gets “killed”: Many ideas during the development phase, very little during scaling
Works best for: Businesses requiring physical presence at points of sale (services, retail)

Uber tested its business model in San Francisco and then began scaling in 2011. After perfecting its reliable ‘button factory’ model, it replicated it globally. This model prioritizes replication over innovation. The initial design phase involves extensive testing and killing off weak approaches, but once the model is perfected, the focus shifts to faithful reproduction.

4. Space Rocket (High destruction, focus on both quantitative and qualitative growth)

Examples: BYD, Xiaomi, early-stage startups
Approach: Rapid simultaneous growth in multiple dimensions
What gets “killed”: Processes, projects, and products constantly
Works best for: Markets where first-mover advantage is critical

The Space Rocket requires constant reinvention. Nothing lasts long; the leader must “kill” and replace elements daily. Businesses rarely stick to this model for long — they morph into another type or spin out of orbit.

All models have their flaws and are challenging to run. The more you “kill off,” the more complex the model becomes.

You can choose any model, but only one. I tried to smash together a button factory and a space rocket — and failed. If I knew then what I know now, I’d have gone with the Conveyor + Lab model.

However, there is the fifth model, and it differs significantly from the four. Companies that choose it also have to “kill” old ideas or products to grow. But they don’t focus only on qualitative or quantitative growth – they respond to the situation as it unfolds – and decentralize their “killing” decisions.

It’s difficult to build and run, but I believe the future belongs to it. I call it “mycelium.”

Want to develop a unique product or business model in just 5 weeks using the ideas from this article? Take a look here.

Mycelium – the art of decentralized adaptation

If you’ve ever hunted mushrooms, you know they’re hard to spot. They rarely cluster, often appearing as random patches on the forest floor. Even distant mushrooms can be part of the same mycelium – a vast, thread-like network functioning like a neural system.

This network has no central control; each fragment adapts to local conditions. It’s constantly growing and evolving, with sections that look nothing alike. Some companies grow the same way – quietly, adaptively, without a central blueprint.

In the four models above, leadership decides what to ‘kill’ once and sticks to that playbook. In the ‘mycelium’ model, CEOs let local leaders choose their paths and share lessons across the network. If a team in Hong Kong, NYC, or Madrid has a great idea, the CEO ensures others learn from it.

Each local leader decides whether to focus on quantitative or qualitative growth – and what to ‘kill,’ when, and how.

Examples: Mondragon Corporation, Spotify, and some others
Approach: decentralized decision-making, opportunistic growth, adapting to the environment
What gets “killed”: Decided locally, then spread as best practices across the company
Works best for: Distributed companies

In such companies, the headquarters is more of a knowledge center than a decision-making hub.

Here are a few examples:

1. Mondragon Corporation (Spain)

Mondragon is a federation of worker cooperatives. Each unit operates independently and is guided by cooperative principles but still benefits from the collective resources and support of the larger network. Mondragon’s structure allows it to remain adaptable and resilient without heavy-handed centralized control.

2. Spotify (Sweden)

Spotify’s organizational structure is built around squads, tribes, and guilds — small, semi-autonomous teams handling different product aspects. Each squad functions almost like a mini-startup, with the freedom to make decisions quickly without waiting for top-down approval. This setup lets Spotify stay nimble while expanding globally.

3. Valve Corporation (USA)

Valve’s flat organizational structure is famous for having no formal managers. Employees self-organize, choosing projects and forming teams based on interest and expertise. This model encourages creativity and rapid iteration, making Valve more of a decentralized network than a traditional company.

4. Haier Group (China)

Haier includes hundreds of micro-enterprises, each handling specific products or markets. These micro-enterprises operate autonomously, making decisions aligned with Haier’s strategic vision. This approach helps Haier remain flexible and customer-focused in a rapidly changing market.

This model is hard to build because you need to trust the people you have and have people you can trust.

But the future belongs to such models because:

They are resilient. They don’t rely on just one decision-making center. Even when some parts of it fail, others remain stable.
You and your executives aren’t the only ones looking for new solutions.
These models can operate in various countries and under different conditions because they are highly adaptive.
They are flexible and can quickly respond to change.

Conclusion

What’s Your Model Right Now?

Most of my clients think they’re following one of three models (excluding the Button Factory), but in reality, many leaders repeat the same mistake I made in 2009 – demanding both operational excellence and Space Rocket-style growth at the same time.

To figure out what model you’re actually running, try this:

Grab a sheet of paper and list up to 20 key processes in your company – production, procurement, marketing, logistics, etc.
Mark the ones that have significantly changed over the past year. If over 10 processes have changed, you’re likely running a Space Rocket model, even if you don’t realize it. If fewer than four have changed, you’re probably building a Button Factory.
If most of the changes clearly align with either quantitative or qualitative growth, you might be running a Conveyor of Button Factories or Conveyor + Lab. But look at your business honestly – are you fooling yourself?
If the changes can’t be clearly categorized as qualitative or quantitative growth – and you’re still demanding operational perfection – chances are you’re on the same path I was in 2009.

Here are a few thoughts on choosing a growth model:

If you have a proven concept that’s thriving in one market and you’re confident its value will resonate elsewhere – consider the Conveyor of Button Factories model.
If you have a broad product lineup for different countries and markets, the Mycelium model could be a solid fit.
If your company must maintain a physical presence in every market it operates in, the Mycelium model will help you adapt. But if you want every division to be a carbon copy of the other, the Conveyor of Button Factories might be more suitable.
The Button Factory model is stable and reliable but risky – it grows cautiously and can get outpaced by more innovative competitors.
The Space Rocket model is the hardest to pull off because it demands constant reinvention of the entire company. It’s only worth pursuing if you’ve got killer value propositions and need to grab market share fast – before your competitors do.

If you need help finding innovative ideas, drop me a message.

Svyatoslav Biryulin

––

Me, me, me

The second audio chapter of my book Red and Yellow Strategies is now available as a podcast on Substack. Interested? Subscribe to that section and start listening.
Now accepting keynote bookings for 25Q3-25Q4.
The early bird discount for Uniquation program is ending soon — learn more here.
Check out my book Red and Yellow Strategies: Flip Your Strategic Thinking and Overcome Short-termism.