A Startup Is Defined by Uncertainty

Garage Duo, Fortune 500 Team

Picture two founders in a garage. Now picture a new-product team inside a large company.

Eric Ries would call both of them startups.

The settings look nothing alike, yet both teams face the same open question about whether customers want their new product.

His definition: "A startup is a human institution designed to create something new under conditions of extreme uncertainty."[[3]](https://orthogonal.io/insights/lean-startup/a-conversation-with-eric-ries-at-the-chicago-lean-startup-circle/)

What the two teams share is simple. Neither one knows yet whether customers want what it is making.

Why care? "Most startups fail. But many of those failures are preventable."[[2]](https://www.penguinrandomhouse.com/books/210088/the-lean-startup-by-eric-ries/)

What the Definition Leaves Out

Look at what is missing from Ries's definition: company size, industry, and sector.[[4]](https://www.strategy-business.com/article/00224)[[3]](https://orthogonal.io/insights/lean-startup/a-conversation-with-eric-ries-at-the-chicago-lean-startup-circle/)

That omission is the point. A startup can sit in a garage, a large company, a government agency, or a nonprofit.[[4]](https://www.strategy-business.com/article/00224)[[3]](https://orthogonal.io/insights/lean-startup/a-conversation-with-eric-ries-at-the-chicago-lean-startup-circle/)

Take a hospital system testing a new home-care service for patients it has never served.

It is large, and it is not a tech company. It still qualifies, because it faces extreme uncertainty about whether this service will work for these patients.

Small and New Is Not Enough

A bakery opens a second shop with a proven menu. It is small and new. It is not a startup.

Its customers and business model are already known, so little is uncertain.

Ries draws this line himself: some of his advice is "not at all true of many profitable small businesses, but they are not what I mean by startups."[[5]](http://www.startuplessonslearned.com/2009/04/validated-learning-about-customers.html)

Apply the same test at scale. A large company extending a proven product line faces little uncertainty, so by this definition it would not count either.

So ignore size. Ask how much you do not know about customers and the model.

A Startup Is an Institution

Why does Ries use the word institution?

Because a startup "is not just the creation of a heroic individual." It is "a system of individuals working together toward a common goal."[[4]](https://www.strategy-business.com/article/00224)

That leads to his bold claim: "entrepreneurship is management."[[4]](https://www.strategy-business.com/article/00224)

If a startup is a system of people rather than one hero, it has to be managed. A garage duo needs that as much as a corporate team does.

But which kind of management? The usual tools carry a hidden requirement.

Why Forecasts Need History

A search-engine manager launches a feature the day before the Super Bowl. Sports queries spike.

Can the manager claim credit? No. Years of history show the spike would have happened anyway.[[3]](https://orthogonal.io/insights/lean-startup/a-conversation-with-eric-ries-at-the-chicago-lean-startup-circle/)

History lets an established business tell cause from coincidence. A startup has no such history to separate its own effect from noise.

Ries puts it bluntly: "It is impossible in a startup situation to make an accurate forecast."[[4]](https://www.strategy-business.com/article/00224)

Neither Chaos Nor a Stricter Plan

When a business plan misses, a tempting response is to write a more detailed one.

The real problem is that the plan's inputs cannot be known yet. Ries likes to ask founders whether their plan had a spreadsheet predicting year-five revenue; nearly every hand goes up, and few stay up when he asks who actually hit it. "We're really good at making plans but not necessarily making any money."[[3]](https://orthogonal.io/insights/lean-startup/a-conversation-with-eric-ries-at-the-chicago-lean-startup-circle/)

Skipping process fails too. A startup is still an institution that needs management.

Ries offers a third path: management built for uncertainty. His book lays it out in three parts, Vision, Steer, and Accelerate.[[1]](https://theleanstartup.com/book)

Spot It in Your Own Work

Look at a project on your desk. Do you know who the customer is and how the model makes money?

If yes, a plan and forecast can serve you. You have history to judge results against.

If no, you are running a startup, whatever your company's size or sector. Your forecast will mislead you.

That is the moment to switch to startup management and start testing what you do not yet know.

A new-product team inside a large corporation can be a startup when it faces extreme uncertainty about whether customers want its product.

• True
• False

Answer: True

Startup status depends on extreme uncertainty, not on being small, young, independent, or in the technology sector. A team inside a large company can qualify if its customers and business model are still unknown.

A startup is a human institution designed to create something new under conditions of _____; this definition does not depend on _____.

Answer: A startup is a human institution designed to create something new under conditions of extreme uncertainty; this definition does not depend on company size.

Ries's definition focuses on the uncertainty surrounding a new product, customers, and business model. It deliberately leaves company size out, so both a garage team and a large-company team can qualify.

Which project is most clearly a startup under Eric Ries's definition?

• A bakery opening a second shop with its proven menu
• A large company's team testing an unproven product for unknown customers
• A profitable retailer adding another location using its established model
• A manufacturer expanding production of its best-selling product

Answer: A large company's team testing an unproven product for unknown customers

The corporate team is a startup because it is creating something new while lacking reliable knowledge about whether customers want it. A bakery opening a second shop with a proven menu is small and new, but its customers and business model are already known.