Innovation

The rule of 3 and 10.

I’ve worked with companies at every stage—early startups to Fortune 50 enterprises. There’s a framework that helps explain what happens as companies scale: the rule of 3 and 10.

Hiroshi Mikitani, CEO of Rakuten, observed that as companies grow in multiples of 3 and powers of 10, systems and processes start breaking down. What works with 3 people doesn’t work at 30. What works at 30 doesn’t work at 90, 270, and so on.

I generally agree with this for most functions—finance, operations, sales, marketing. These areas need systems to scale.

But there’s one exception: innovation.

The startup phase: chaos works.

Startups have one job: find product-market fit with minimal distraction. That means failing fast, shipping MVPs, absorbing customer feedback, and iterating without overthinking it.

There’s no time for elaborate planning. Expensive tech stacks, formal processes, hierarchies—all of that can wait. You’re moving too fast to build scaffolding.

The scale trap: growing up too fast.

Once companies find product-market fit, sign initial customers, and start growing, basic systems get implemented. This is necessary—without some structure, profitability becomes unreachable.

But here’s where it goes wrong: there’s a gravitational pull to add process, planning, and technology too soon. Companies start building for future scale before they’ve proven current traction.

Scaling infrastructure prematurely overheats resources, burns capital, and distracts from revenue growth. Worse, innovation suffers because added processes create friction. The speed that got you here slows to a crawl.

The enterprise phase: the giant hairball.

Years of growth mean layers of products, businesses, technology, systems, rules, traditions, and processes. Gordon Mackenzie called this “orbiting the giant hairball”—an ecosystem that consciously or subconsciously wrings risk out of the organization.

Employees execute defined routines with little upside to innovate. Risk-taking gets punished. Mavericks get treated like viruses—corporate antibodies attack until the innovators leave or conform.

Innovation becomes incremental product improvements instead of disruptive new business models. Stage-gate processes with five layers of approval. Five-year plans that are outdated before they’re finished. Six Sigma rigor applied to exploring the unknown.

It’s all designed to reduce risk. And it works—by also eliminating the conditions that create breakthroughs.

The permanent MVP state.

Here’s my thesis: organizations should resist scaling processes and planning specific to innovation functions at all growth stages.

Minimum Viable Product, Minimum Viable Process, and Minimum Viable Planning should be the permanent state for new technology, product, and business model innovation.

Innovation culture should not grow up. It should maintain MVP vitality forever—regardless of whether the company has 30 employees or 30,000.

Why Elon gets away with it.

Love him or hate him, Elon Musk routinely wills a general concept into a billion-dollar business across industries. Ideas that would die immediately in most organizations.

Why? He thinks in first principles. He asks questions the way a child would—without baggage, without “that’s not how we do things here.” But he combines that childlike curiosity with advanced intellect and experience.

His other differentiator: bias for action. He ships. He tests. He breaks things and learns fast.

Most enterprises have the intellect and resources. What they lack is the willingness to stay in permanent startup mode for their innovation functions.

How to keep innovation wild.

Here’s how to maintain an MVP innovation culture as you scale:

1. Keep teams small.

Product development teams should stay under 10 people. Independent. Autonomous. Free to chase seemingly crazy ideas.

In my experience, five teams of five working on the same problem outperform one team of 25. Small teams move faster, communicate better, and don’t need three layers of approval to make decisions.

2. Go agile, actually.

Agile methodology is standard in startups, especially for developers. It’s less common in enterprises still clinging to waterfall processes.

Invest in real agile training across functions—not the bastardized “agile” that just means more meetings.

3. Use simple frameworks.

Business Model Canvas is visual and straightforward. It forces you to think through key assumptions on one page.

If your planning document is longer than one page in the early stages, it’s probably too much. You’re planning instead of learning.

4. Build internal VC funding mechanisms.

Many large enterprises have corporate VC arms to invest externally. Let internal innovation teams tap the same funding mechanisms.

Keep the process simple—no 47-slide decks to get $50K for an MVP.

5. Kill stage-gate processes for innovation.

This is hard for me to say—I’ve advocated for stage-gate processes most of my career. But they don’t work for early-stage innovation.

Develop a simple prioritization framework. Start working on the best ideas. Iterate based on what you learn, not what the stage-gate calendar says.

6. Stay horizontal.

Layers of hierarchy kill decision-making speed. Stick with horizontal org designs as much as possible for innovation teams.

If someone needs three approvals to run a $5K experiment, you’ve already lost.

7. Protect the mavericks.

I’ve heard the talk in large corporations about embracing risk-taking and entrepreneurial culture. I have not experienced it being real.

Risk-taking and failure lead to corporate exits—voluntary or involuntary. Mavericks get treated like threats. For every maverick pushing boundaries, there are ten employees rooting for them to fail so things can go back to normal.

In startups, everyone is a maverick—or they’d find less risky jobs that pay more. Enterprises need to create protected spaces where mavericks can operate without corporate antibodies attacking them.

8. Simple financial modeling.

I’m guilty of building complex financial models in early-stage product development. My learning: a simple model works just as well to determine if you’re on the right track.

You don’t need 20-year forecasts with sensitivity analyses when you’re testing whether anyone wants the product.

The takeaway.

Companies scale. They add systems, processes, technology, people, and planning. That’s fine—most functions need structure to operate efficiently.

But innovation is different. It’s an animal that needs to stay a little wild.

As companies grow, they try to cage the innovation function. Bring it under control. Make it predictable. Reduce the risk.

Resist.

Keep your innovation engine in permanent MVP mode. Small teams. Simple frameworks. Bias for action. Protection for mavericks.

Scale the parts of your business that benefit from systems. Keep innovation chaotic, fast, and a little bit dangerous.

That’s where breakthroughs come from.